Oral argument — United States Sugar Corp v. EPA - Part B: Environmental

I'd like to address the first three issues that the Court scheduled in these arguments. EPA's decision to set scores for the statutory minimum stringency at a 99% upper prediction limit instead of the average, EPA's decision to exclude the best-performing sources from its floor analysis because they were burning cleaner fuels, and EPA's use, again, of carbon monoxide as a surrogate. On the use of the upper prediction limit, the Clean Air Act unambiguously requires floors to reflect the average emission limitation achieved by the best-performing sources. And EPA has agreed consistently in notice and comment rulemaking that the meaning of the term average is its ordinary meaning, that is, that it refers to a central tendency such as the arithmetic mean or median. EPA has taken that position consistently through the years, for example, at page 46 of the Joint Amendments. The upper prediction limit is not a measure of central tendency. It is a level that, by EPA's own description, any future test by any source in the top 12% will fall below. EPA says it is 99% confident that any future test by any source in the top 12% will fall below the upper prediction limit. That's at page 911 of the Joint Amendments. What's more, EPA defends the upper prediction limit approach with a description of its goal as ensuring that all the sources in the top 12% will be below the upper prediction limit at all times. And a quote on that that I think is especially worthwhile is at page 1861 of the Joint Appendix, where EPA rejects another floor approach because, quote, it wouldn't guarantee that the floor can be met by all the existing sources in the floor at all times. The data in EPA's records confirms that the upper prediction limit does exactly what EPA designed it to do, that is, virtually every source and virtually every test result is below the upper prediction limit for all the standards. In fact, EPA, to set a level that is lenient enough to guarantee that every source in the top 12% will be below it at all times, EPA actually set standards that are so lenient that in the case of one group that we looked at for one data set, 85% of all the sources for which EPA had emissions data tested below the upper prediction limit. A level that is higher than any of the data it purports to represent cannot be the average of those data. It certainly isn't a central tendency for those data by any stretch of the imagination. And by insisting that the upper prediction limit is an average, EPA drains the meaning – drains the meaning from the word average and completely defeats the purpose of a textual limit on its discretion that Congress put in the Clean Air Act before language. In fact, going back to EPA's notice and comment interpretation of the term, even EPA agrees, at least in its prior interpretation, the only one that has gone through notice and comment rulemaking, that this language simply cannot be read to mean all the sources in the top 12%. So at page 37 in the joint appendix, EPA says, Section 112b3a simply cannot – simply does not say the emission limitation achieved by all sources – all the best sources – all the sources within the best performing 12%. The other point I'd like to make on the upper prediction limit, apart from the fact that it is clearly not an average and therefore violates the law, is that EPA still has not explained how it is an average. This court gave EPA the opportunity to do that in an ACWA case where EPA used the identical approach, and the court couldn't tell what EPA's interpretation was. It thought there was an interpretation that EPA might be advancing, but it wasn't sure, and it couldn't tell how that was an average. And so EPA gave EPA an opportunity to explain how this number is an average. EPA came out with a litigation memo that didn't go through notice and comment, and which had actually disavowed the interpretation that it had – the court thought it was in the rule in the sewer sludge incinerator case in ACWA. But it still hasn't explained how this is an average. The idea that this is an upper limit, as EPA describes it, that all future tests will fall below, and an average, cannot both be true. And EPA has never, in the rulemaking record or in its post-doc litigation memo, explained how those two ideas can be reconciled. In fact, if you look in their brief, EPA's lawyers are reduced to making the same claims that EPA made in the ACWA case, which is that the formula speaks for itself, or that it's built into the methodology. Well, the same formula was before the court in the NACA case, and it didn't speak for itself there, and it still doesn't speak for itself. If I could move on, unless there are any questions on the effort prediction limit, I'd like to address EPA's exclusion of the best-performing sources. And really, I think there are only a couple of points to make on this. The – EPA doesn't disagree that the agency had a policy of excluding best-performing sources if they were burning cleaner fuels. And it also – I don't believe it disagrees that that policy would violate the existing source for language in the Clean Air Act, which requires for us to reflect the average emissions achieved by the best-performing sources for which EPA has emissions information. EPA's two arguments are that as to existing sources, the issue is moot, because EPA only excluded a few times, and those standards have either been renamed already or changed. And with respect to new sources, EPA finds it's reasonable. So I'd like to go to the mootness argument first. And this is a – sort of a threshold point where a party is claiming mootness, that party has a heavy burden of proving mootness. But I don't know of EPA's kind of precedent argument. I think it's actually based on just a mistaken reading of a memo in the record. Because the record makes entirely clear that EPA did have a blanket policy of excluding any source that burned over a certain threshold of clean fuel. And I think it's probably easiest to explain this as an example. So for the solid fuel subcategory, for example, EPA defined that subcategory, solid fuel boilers, as including co-fired sources that burn up to 90 percent natural gas. That's EPA's own definition, and it's in page 270 of the joint appendix. But EPA also said that it excluded from its floor analysis any boiler that was burning more than 10 percent natural gas. That's at page, I think, 1008 of the joint appendix, which is the response to comments document. It's also on page 908, which is a memo about how EPA set up the floors. So there's no question that this was EPA's policy. And by the way, EPA also excluded any source that was burning any natural gas from its new source floor approach. And so for existing sources, the cutoff was 10 percent, and for the new source floors, the cutoff was any. EPA's lawyer's argument relies entirely on a memo that describes exceptions to that policy. And the word exceptions, the fact that that memo is describing exceptions, is at page 1454 of the joint appendix. And the particular exception it describes is that in a few cases where EPA was looking at existing source floors, the agency changed the cutoff to any natural gas. Normally it would have thrown out any, excuse me, any boiler that was burning more than 10 percent natural gas. But for those few exceptions, it threw out any boiler that was burning any natural gas. And the reason it did that is that those were for categories that were so small that they were subcategories that were so small that the best performing 12 percent was actually one unit. And EPA didn't want that one unit to be burning natural gas. So EPA's entire mootness argument is premised on this one memo, and we think it's clearly – the memo doesn't say that. The memo doesn't purport to enumerate all the boilers that EPA threw out under its normal policy. It just enumerates a few exceptions. It doesn't purport to change that policy. But EPA hasn't met, or can't meet, its burden of proving mootness by relying on that memo, which is all it decides. The other point, and just very quickly on the existing source floors, EPA doesn't really dispute that that's unlawful, but clearly these are best performing sources for which EPA has emissions information. So if EPA throws them out of the floor analysis, it's violating the law. And EPA's rationale for throwing them out, which is that they're burning cleaner fuel, is also at odds with this court's precedent, where the court has decided over and over again that EPA can't do that. This is the Arcliffe case, Montgomery Cyclone case, Carleton case, National Island case. As for new sources, EPA agrees that this is a live issue, but claims it's reasonable to exclude the best performing new sources because the new source language refers to the best controlled, quote, similar source. And it says, well, if they're burning cleaner fuels, they're not similar to the sources that are burning dirtier fuels. There are a couple problems with that. The first one is textual. The Cleaner Act actually provides for EPA to put similar sources together and separate dissimilar sources. That's the language that says that EPA can set separate standards for different classes, types, and sizes of source. And that's what EPA did in this rulemaking. EPA created an unprecedented 36 different subcategories in this rule, some so small that they contained only a few units, precisely so that it would make sure that all the similar sources were alike, or all the similar sources were regulated together, and all the dissimilar sources were separated. But what EPA is saying is that similar serves exactly the same function as that class, type, and size provision, so that EPA can divide sources once, and then even within the category that it's created, it can choose not to base floors on the best source just by saying that it's not similar. First of all, that makes language in the Cleaner Act redundant. Second, it really makes the whole new source floor provision completely discretionary with EPA, because EPA gets to pick the source that it thinks is similar. And finally, it fails even on its own terms, because it makes the question similar to what? If EPA can pick the source on which it's going to base floors just because it's similar to what standard applies to those sources if the standard is based on a dirtier source? There's simply no rationale for saying that the standard has to be based on one similar source if EPA itself is contending that the sources within the category, the subcategory by the way, that it carefully tailored to contain only similar sources, actually contains dissimilar sources than the best source, and it represents some of them. That's arbitrary as well as unreasonable. The last issue I'd like to address is EPA using carbon monoxide as a surrogate for all of the hazardous air pollutants, all of the organic hazardous air pollutants that boilers emit except dioxide. There are a number of problems with this, but I'd like to start with the statutory requirement. The fundamental statutory principle is that EPA has to set standards for every hazardous air pollutant that the source category emits, and each of those standards has to reflect the maximum degree of reduction that is achievable at Section 112.2, and has to reflect the emission levels achieved by the best performing 12% of sources at Section 112.3. And that's why in the copper smelters case, this court held that for a surrogate to be reasonable under Section 112, for Section 112 purposes, it has to identify the sources that are best performing with respect to the hazardous air pollutants that it's supposed to control, as well as what those sources can achieve with respect to the hazardous air pollutants that it's supposed to control. And that holding just tracks the statutory language and the statutory intent. If that weren't true, then EPA could simply dodge the statutory... Have you shown us anywhere in the record where there's control technology that's targeting carbon monoxide, and is it also controllable via PAHs? I know there's a study that shows that they diverge at certain temperatures, but don't you need to show... I mean, the EPA has also shown there are control technologies that address both, and have you been able to show, or have you shown that there are control technologies that address both, but that they diverge somehow? Well, yeah. I mean, the argument that you have is a theoretical one, but we need to have something in the record to show control technologies where they don't correlate. I think we can bring it down to a very practical point, although I think it's the flip side of what you're saying, because what the record shows... In fact, these aren't our comments, by the way. These are comments from the manufacturers of emissions controls, is that there are controls such as activating carbon injection and selective phthalate reduction that reduce polycyclic organic matter and PAHs without having any effect on carbon monoxide. Also, the study in the record shows that sources can significantly reduce their emissions of higher, and again, that has virtually no effect on their emissions of carbon monoxide. Now, if EPA had looked at the sources that are doing this, that are either using these control technologies or burning cleaner fuels, it would have found that there are sources that are achieving better reductions in polycyclic organic matter, and other than those that are simply meeting the carbon monoxide level, and that's why the copper smelters test. EPA's carbon monoxide does not identify the best performers in respect to polycyclic organic matter, because there are sources out there that are doing things, taking practical measures to reduce those pollutants that are not necessarily the ones that are meeting the carbon monoxide limit. There's a completely additional reason that carbon monoxide is not a valid surrogate, and that is that EPA itself says that the relationship between carbon monoxide and the only actual organic hazardous air pollutant that it looked at, which was formaldehyde, breaks down below 130 parts per million. The record shows that a number of boilers are operating in that breakdown zone. That's where they operate is with their carbon monoxide levels below 130 parts per million, and by EPA's own conclusion, carbon monoxide isn't a valid surrogate there. So if carbon monoxide isn't a valid surrogate for at least some of the sources in the category, it's not a valid surrogate. The other point is that the reason there's a breakdown is that formaldehyde emissions start going back up as carbon monoxide emissions go below 130 parts per million. So the only way that sources could actually minimize their emissions of formaldehyde is to keep their carbon monoxide emissions right at 130 parts per million. They couldn't go over, and they couldn't go under, because either of those things makes formaldehyde emissions go up, according to EPA. Therefore, you have a standard where sources are in compliance with a carbon monoxide standard if their emissions are below 130 parts per million. All the sources that are operating below 130 parts per million may very well have significantly higher emissions of formaldehyde. And by EPA's calculations, other organic hazards are pollutants as well, even though they're in full compliance with the carbon monoxide standard. If there are no further questions, I guess I'll reserve the rest of the time. All right. Thank you. Mr. Gormley? Good morning again, and may it please the Court. My name is Neil Gormley for the Environmental Petitioners in Cases 11-1125 and 11-1141. I'd like to begin with Case 11-1125, which concerns challenges to EPA's Standards for Commercial and Industrial Solid Waste Incinerators, or CISWI. And I'd like to focus on two issues in that case. First, EPA's decision to exclude the vast majority of CISWI from the pollution limits. And second, EPA's failure to set beyond-the-floor limits requiring the maximum achievable reduction in pollution. Out of many thousands of CISWI that are polluting the air all across the country, only 106 are following environmental rules. EPA claims that pollution limits are not required for the remaining units. Burn-off ovens, cyclonic burn barrels, foundry sand reclamation units, and space heaters. EPA's decision to exclude those units from pollution limits is contrary to the plain language of the Clean Air Act. As interpreted by this court in its 2007 decision, NRDC, Section 129A1D required EPA to set standards for CISWI by November 15, 1994. And this court held in NRDC that pollution limits must apply across the board to all CISWI. EPA's decision to exclude the vast majority of CISWI from these standards is unlawful under those authorities. EPA claims that standards are not required for these units because they're not needed to satisfy a different statutory provision, 112C6. Now, we don't have a 112C6 challenge in this case, but it's not relevant because EPA has an independent obligation to satisfy Section 129. EPA also claims that it's excused from this obligation because it lacks data on the remaining CISWI. But EPA's mandatory duty to set standards for all CISWI by 1994 is not conditioned on whether EPA chooses to collect data. If the lack of data in the record is relevant to this case, it's relevant only as to remedy. And I'd like to emphasize, with respect to remedy on this point, just how long EPA's evasions have gone on with respect to these units. Even if we focus just on the period of time during which EPA has been in violation of the congressional deadline, EPA has had over 20 years to collect the necessary data. EPA has also received plenty of prompting from stakeholders and from this court to finally set these standards. A 1997 consent decree required EPA to set standards by 2000. A second lawsuit, leading to a voluntary remand in 2001, required EPA to address the narrow coverage of those initial standards. And then, in 2007, this court... So what's the remedy you seek? Are you seeking a backer or remand with instructions to get it done? We ask the court to remand with instructions to set standards for all CISWI. For all? For all. For all commercial and industrial solid waste incinerators. That's right. If I could say something briefly about jurisdiction. This court has jurisdiction because EPA took final agency action with respect to these units. It did so in at least two ways. First, it adopted definitions for these units that exclude them from the pollution limits. And that's final agency action with respect to these units. Second, it reached a conclusion, a legal conclusion, that pollution limits are not required for these units. Both of those are final agency actions that this court is empowered to review under the Clean Air Act. If I can turn to the second issue, the beyond the floor standards. EPA failed to set standards that require the maximum achievable reduction in pollution, considering the statutory factors as required by the plain language of the Act. In considering beyond the floor standards, EPA claimed authority to set whatever level of protection it deems reasonable. And it then exercised that purported discretion to reject every single beyond the floor standard for every single pollutant from every single subcategory of CISWI. The statute simply does not give EPA authority to choose the reasonable level of reduction. The statute says EPA shall require the maximum achievable reduction. And EPA is bound by the statute. Reasonable and maximum achievable simply are not synonyms. And EPA applied the wrong standard here. When – in the 1990 amendments to the Clean Air Act, when Congress intended to grant the regulatory authority discretion to adopt a reasonable level of control, Congress said so expressly, as it did in Section 172C, which calls for, in certain nonattainment plans, standards based on reasonably available control technology. The contrast with Section 129, which requires not reasonable reduction, but the maximum achievable reduction, is significant. But throughout its decision, EPA repeatedly rejected stronger standards on the basis that they're allegedly not reasonable, without even claiming, let alone demonstrating, that they're not achievable. To take an example, for carbon monoxide emissions from coal-fired energy recovery units, EPA predicts that those units will only have to conduct a tune-up under these standards. It rejected control technology that could achieve low-cost reductions on the ground that they're allegedly not reasonable. It rejected thermal oxidizers, a standard based on thermal oxidizers, as allegedly not reasonable because they require natural gas, without claiming, let alone explaining why using natural gas is not achievable for these units. By failing to apply the statutory standard, EPA violated the statute at Chevron Step 1. If there are no further questions on the environmental petitioners in the CSRI case, I'll turn to the area source boiler case. Even before EPA authorized them to burn materials like scrap tires, used oil, and demolition debris, area source boilers already emitted very large quantities of hazardous air pollution. But EPA's final rules set pollution limits for far fewer than 1% of the area source boilers, and by EPA's own analysis, will achieve virtually no reduction in hazardous air pollution. EPA violated several legal requirements on route to that result, and I'd like to address three of them. I'll start with 112C6, and then address the GAC standards, and finally, Title 5. With the court's permission, I'd like to also say something briefly about the work practice standards. While they're not, strictly speaking, a GAC standard, they're another route by which EPA seeks to justify establishing just two notes as the control. First of all, 112C6. Even though oil-fired area source boilers and wood-fired area source boilers are currently listed for regulation under 112C6, EPA refuses to set 112C6 compliance standards for those units. EPA's brief defends an action that's entirely different from the one the agency actually took here. EPA says that standards are not required because these units weren't listed. The actual agency decision concedes that coal combustion, oil combustion, and wood combustion are listed, that the 112C6 list currently includes oil combustion, coal combustion, and wood combustion, and makes no distinction between major sources and area sources. EPA's concession that coal-fired area source boilers are on the list compels the conclusion that oil-fired and wood-fired boilers are also on the list because they're treated exactly the same in the decision. That conclusion is confirmed by the actual 1998 listing action, which stated that area source components of the listed categories would be subject to standards, and it also said that in the event that the emissions from area sources turned out to be trivial, the area sources would be removed from the list. EPA couldn't, it would be impossible to remove sources from the list unless they were on the list in the first place. EPA's brief also claims that the agency has authority to remove sources from the 112C6 list at any time. That is also not the situation presented by this case. EPA could not have been clearer, and this is at J53, that area source boilers were not removed from the 112C6 list here. The actual situation presented by this case is that these sources are sitting there on the 112C6 list, and yet EPA refuses to set 112C6 compliance standards for them. EPA has not explained how that's consistent with the statute, and at the very least, that requires a remand. With respect to the actual language of 112C6, which again EPA's brief doesn't address, and EPA has not provided a reasoned interpretation of that language, and that's why we think a remand is required on this issue, I would say that if the court wants to examine that language, a lawful interpretation of 112C6 must give effect to, must give the effect that Congress intended to the listing action. Congress was clear that the listing action must, quote, assure. The listing action must assure that sufficient sources will be regulated under what appears to be EPA's interpretation, which is that the listing action is just an estimate of what sources may ultimately come to be regulated. With no binding effect, the listing action can't assure anything, and that interpretation would not be consistent with the language. It's also important to note that the 90% requirement in 112C6 is a minimum requirement. 112C6, in fact, contemplates that EPA can and may list sources above 90% of the emissions of these very dangerous pollutants. But under, again, what appears to be EPA's interpretation of the statute, if EPA takes Congress's invitation and lists additional sources above 90%, it's essentially a meaningless act because EPA has complete discretion, even after that listing, to either set or not set 112C6 compliance standards for those sources. One of the counsel for the industry petitioners described 112, Section 112, as elegant, and we agreed in the respect that it is meant to work in a very orderly way. It begins with listing under 112C. That triggers the scheduling provisions of 112E. Now, all of this was supposed to have happened much faster than it actually did happen, but the way that Congress contemplated it to work, listing under 112C would lead to scheduling under 112E, which the legislative history makes very clear, EPA does not have authority to modify after the fact. And then finally, standard setting is determined by the manner in which sources were listed. EPA has not given any reasoned explanation of why 112C6 standards are not required by how failing to set 112C6 compliance standards for sources on the 112C6 list is consistent with the statute. Turning to the GAC standards. One of the GAC standards for one of the subcategories of sources that are subject to GAC standards is based on the emissions from an uncontrolled coal-fired boiler. And that's even though EPA concedes that fabric filters are a generally available and cost-effective control technology for that subcategory. That's setting the emissions for that subcategory on the basis equal to the emissions from an uncontrolled boiler violates the plain language of D5, which requires that EPA provide for the use of generally available control technology or management practices. Neither EPA nor industry intervenors have defended that standard in their briefs. The GAC standards are also arbitrary in general because EPA failed to consider setting MAC standards instead of GAC standards as a matter of discretion. Discretion that the agency does not dispute that it has. Under basic administrative law principles, EPA has an obligation to rationally exercise that discretion and to give reasons for why it didn't set MAC standards in lieu of GAC standards, especially after receiving comment on that issue as it did. EPA completely failed to consider setting MAC standards as a matter of discretion. The GAC standards are also arbitrary because EPA completely failed to consider the congressionally defined objectives of the area source program. These are given in 112K, namely a substantial reduction in area sources hazardous air pollutant emissions, a substantial reduction in associated health risk, and a reduction of not less than 75% in associated cancer risk. EPA adopted standards here pursuant to GAC that will achieve virtually no reduction in hazardous air pollution and that it doesn't claim will yield any health benefits. And it did so while completely failing to consider the congressional objectives. And in this respect, on this point, I'd like to direct the court to a 2001 decision by Judge Henderson in Association of American Railroads, that's 237S3-676, in which the court held in action arbitrary and capricious for failure to consider the congressional objectives in the preamble to the statute. I'd like to say something briefly about the work practice standards. The work practice standards that EPA set for coal-fired boilers, namely tune-ups and following manufacturer-recommended procedures, are contrary to the plain language of 112H1 because EPA failed to determine whether they're consistent with the stringency requirements of 112D or F, as 112H1 clearly requires. These are sources that are subject to MAC standards. These are coal-fired boilers, and EPA agrees that it is obligated to set MAC standards for these sources. But it set just a tune-up, and it did so without considering whether that's consistent with the stringency provisions of 112D or F. This is a statutory requirement that all of the parties to this case concede, but EPA failed to make the determination it concedes it should have made. EPA is wrong when it argues that simply by identifying these work practice standards as work practice standards, it somehow determines that they're consistent with the stringency provisions of the statute. First, that's wrong because 112H1 expressly directs the EPA administrator to make a judgment on this point, and EPA hasn't done that. EPA is also wrong because that argument doesn't make any sense. A work practice standard is essentially any non-numeric requirement to reduce emissions. It could be anything from a requirement to run a sophisticated and highly effective control technology all the way down to just a tune-up every five years, which is what EPA adopted here. The mere fact that a standard is non-numeric tells us nothing about its stringency, and in particular, it tells us nothing about whether it's consistent with the stringency provisions of 112D or F as required by 112H1. Finally, I'll turn to the Title V issue. Out of about 92,000 sources across the country with area source boilers, an estimated 48 are large enough to be major sources, but promise to bring their emissions below the major source threshold using control technology. EPA's decision to reverse course and exempt those estimated 48 controlled synthetic area source boilers from Title V was arbitrary and capricious. It was arbitrary and capricious, first, because EPA has no evidence to support extending its rationale for the remaining sources, the natural area source boilers, to the controlled synthetic area source boilers. In fact, EPA initially claimed that its complete lack of information about the synthetic area source boilers itself justified exempting them from Title V. That's incorrect because no evidence is not substantial evidence, and agency action is arbitrary and capricious without substantial evidence. EPA also claims that the natural area source boilers and the controlled synthetic area source boilers are equivalent, and that's arbitrary because there's no evidence to indicate that they're equivalent, and in fact, all of the evidence in the record suggests that they're not equivalent. Recall that by definition, the controlled synthetic area source boilers are large enough to be major sources. In stark contrast, the vast majority of area source boilers, the natural area source boilers, are very small, below 10 million BTUs per hour in size. There's a memo in the record, and I regret that it's not in the joint appendix, but it can be found at docket ID EPA HQ OAR 2006-0790-0035 at page 3, where EPA explains that the synthetic area source boilers are much larger than that. For synthetic area source boilers burning biomass, they're at least 25 times larger than 10 million BTUs per hour, and for synthetic area sources burning coal, they're at least five times larger than that. The Title V exemption is also arbitrary because EPA failed to consider and address contrary evidence in the record. Comments submitted by the Partnership for Policy Integrity show that of 32 recently permitted biomass boilers, and these boilers were recently permitted as area sources, many are large standalone power plants, and that they overlap in size substantially with the major sources. Many are larger than major sources. EPA gave no reason for rejecting that contrary evidence and didn't consider that contrary evidence, evidence that contradicted its determination that they are equivalent, that the synthetic area source boilers are equivalent to natural area source boilers. EPA's attorneys offer a post hoc rationalization for rejecting those comments. This court can't uphold the Title V exemption on the basis of a post hoc rationalization, but even if the court were to consider it, it lacks merit. What EPA's attorneys say is that EPA isn't sure it considers those sources to be area sources. That doesn't just not provide any help. It actually underscores the problem here, which is that the state authorities are permitting as area sources, large power plants that even EPA isn't sure are truly area sources, and yet without the data that's available only under Title V, citizens have no ability to challenge the claimed area source status of those units as mistaken or fraudulent. So EPA's attorneys' post hoc rationale for rejecting that memo actually shows why Title V is so important and needed for these estimated 48 sources. EPA completely failed to reconcile its recognition that Title V is important and needed for these sources with its simultaneous claim that the benefits of Title V are minimal or nonexistent. And for that reason, too, the exemption is arbitrary. If there are no further questions, I'll save my time for rebuttal. All right. Thank you. Well, it's now good afternoon, Your Honors. Norman Ray from the Department of Justice just presented Respondent EPA. I'll start with the major boiler case and the use of carbon monoxide as a surrogate. This is a technical judgment that the agency has made that carbon monoxide is an appropriate substitute, and therefore the fishers, of course, have a high burden to show that it's arbitrary and capricious, and they failed to do so. The use of carbon monoxide as a surrogate is based on the chemistry of combustion and is supported by evidence in the record. In a combustion device, such as a— Isn't there some post-combustion problem? Don't you have to deal with—isn't there evidence that post-combustion is showing that chemical reactions post-combustion show that carbon monoxide is not a good surrogate? That's not for these units, Your Honor. There was no—the only compound for which there was any evidence in the record that there was any formation of half post-combustion is for dioxins and furan, and the agency established a separate work practice standard for those. For none of the other organic halves was there any evidence of any post-combustion determination, and the agency determined that there's no evidence, there's no reason to believe that there would be any post-combustion creation of organic halves other than the dioxins and furans, which is addressed separately. So the carbon monoxide standard is based on knowledge of the basic chemistry of combustion. In a combustion device, such as a boiler, where complex organic compounds are broken down by reactions with oxygen, ultimately form carbon dioxide and water. The reaction goes to completion, and it goes to— these larger compounds are broken down, they go through a number of steps, including the PAHs. The PAHs are created and then destroyed in the process. But the second last step in the process is the conversion of carbon monoxide to carbon dioxide. So that, if you're measuring the levels of carbon monoxide as the combustion reaction proceeds, or on a continuous basis, it's telling you how well the process is moving towards complete combustion. In other words, it's telling you how far it's moving towards the complete destruction of all the organic material, or at least all that's possible to combust in the process. And that's why carbon monoxide is the surrogate, because by measuring the carbon monoxide— because you know the carbon monoxide— Isn't there evidence that there's actually an inverse correlation between PAH and carbon monoxide at high burns? No, not certainly under the conditions that are found in these sorts of boilers. There was some evidence that at very low— one of the issues we have here, Leonard, is that as the amount of organic material decreases to a level where carbon monoxide levels are at about 130 parts per million, you hit a threshold where there's essentially no other organic material left in the combustion process, maybe just tiny amounts. They refer to the fact that formaldehyde levels seem to fluctuate around that level. EPA explained that by talking about the fact that there's uncertainty in the measurement of formaldehyde at these extremely low levels. I thought there was evidence that at the high temperature, the high burning temperature, you had the phenomenon where the PAH drops off the chart, but the carbon monoxide spikes, shows that they have a different reaction. Is that not right? That's not right, Your Honor. What one of the petitioners tried to argue, pointing to one of the studies in the record, based on one table in that study, was that differences in temperature and differences in concentration might change the relationship, but that table doesn't talk about emission. It's talking about the amount per amount of fuel, and the authors of that study themselves, and we cite this in R3, specifically say that they interpret the results of their study to show that carbon monoxide is a good surrogate, is a good indicator of PAH production. That's at pages 497 to 498 of the joint appendix. The petitioners are trying to take one figure out of the study, but the overwhelming evidence in the record is that there is a strong relation, and carbon monoxide is a common surrogate. It's used as an indicator of combustion efficiency, because it tells you how well, and it also gives you a continuous method for controlling combustion. If the carbon monoxide level comes up, you have to adjust your combustion conditions to improve, so you are destroying the hazardous pollutants, organic pollutants that are in the process. It's based on the chemistry. There is this threshold, and yes, you can get funny effects at the threshold, but it's because there's so little of the organics left that you have measurement difficulties measuring those well. The whole issue of the surrogate was removed from this case. It was subject to a petition for reconsideration that EPA has just recently addressed. The question of whether 100... EPA has set a lower limit here, 130 parts per million of carbon monoxide, because they found that at that level, you don't see any further reduction for the organics below that, because it has essentially all been destroyed. That particular issue is not an issue in this case, because EPA is undergoing a reconsideration process, which is now being completed. That's why the standards are set at that one level, and that's where you get them. They cite some of the industry comments. The industry commenters were talking about these low levels and saying that the CO, the carbon monoxide standard, was too conservative, because they contend that this relationship, this threshold is achieved at higher levels. There's nothing in the record that demonstrates that carbon monoxide is not an adequate surrogate, and it follows directly from the chemistry of the combustion process. Let me turn to the exclusion of certain units from the MAC analysis. In exclusions, EPA... First of all, it's important to recognize that this is a very heterogeneous category of sources. They burn large numbers of different kinds of fuels. That's why we have so many different subcategories. They burn biomass, whether it's agricultural waste, whether it's wood products. They burn coal. They burn oil. They burn a whole range. In addition, many of them are set up to use different kinds of fuels. Some of them are set up to use gas. Some of them are not used gas. And the types of fuels that the particular boiler will use will vary over time, depending upon the price of different fuels, the availability of different fuels, seasonal effects, et cetera. EPA has a very heterogeneous group of boilers that it has to set standards for and has to set it based on the MAC floors. That's why it broke down and created so many categories with these different fuel types, which also reflect different types of units, physical types, feed structures, et cetera. But in setting the standards and determining what's the appropriate data to determine what is achievable by standards in the category on a practical basis, the agency had to utilize the fuel that these units may use up to 90% or up to 100%. And therefore, it excluded units that are tests that are based on natural gas because many of the units don't burn natural gas and many of them don't burn it all the time. So to get a MAC standard that is representative and that really reflects what these units can achieve on an operational basis, the agency had to look at what reflects kind of the core fuel that they're using and set the standards based on that fuel. And really, the statute gives the agency sufficient latitude to do that. In terms of the mootness argument, and I believe Mr. Pugh said this backwards, but in their opening brief, or at least as we understood their opening brief, they were only talking about new source standards and they identified four specific source categories. Two of those source categories were already remanded because the standards were based on small data sets and two of them were set, were carbon monoxide standards that were set at the threshold level and therefore, it didn't matter which units we looked at or not. In their reply brief, they seem to have switched to a whole different set of standards that they're arguing about, which are the existing source standards, which obviously, we haven't had a chance. We believe that our industry waives. It wasn't waived in the reply. But even if they are, as I just explained, we believe it's reasonable and okay to make a technical judgment, just as it would make a judgment if it determined that a particular test was faulty because equipment was used or because there was some technical issue with it. We believe it has the discretion to determine that the test conditions that set the standards that must be achievable by these sources should reflect the fuels that these sources generally use. In that case, would not include natural gas. Let me talk now about the upper prediction limit or UPL. I'll make two points. UPL is consistent with the statutory requirements that standards be no less stringent than the average performance achieved by the top 12%. And it's not arbitrary. The UPL... I'm going to address what the UPL is. And it is not the standard level at which we believe any test from any source was lied below. The UPL is what EPA predicts using standard statistical measures that the average of the best performing sources will achieve under a broad range of operating conditions, under all reasonably foreseeable conditions. And it is the same sort of analysis one might do if you were a manufacturing plant, manufacturing parts on a continuous basis, and you would pull samples off the line, a small number of them, measuring parameters, and try to get an estimate of how well your process was looking. And you would recognize that the overall production process, making thousands of units, would be more variable than the 5 or 10 or whatever you pull off. So there are years you develop these standard statistical techniques that's going from a small sample size and predicting what the larger sample size would look like. And that's what EPA has used the UPL for here. They've taken the test that they have, the average values, and said, okay, let's say we took those same facilities and they ran tests thousands of times under a wide range of conditions, sort of normal operating conditions, and given the variability that we see already in the test, what would we predict that all those tests would look like? What would the body of tests of all these tests and any average would be? And that's how they calculate the UPL. And that's what we've always maintained the UPL. I understand there's been some confusion. I understand there's some confusion in the NACLA decision. We certainly have not claimed that it's the, we're not saying that the average is the average of a freeze test run. It's the average of the best performers. That's what we've said, and that's what we've laid out very clearly in our response to the remand. And that is what is inherent in the formula. The formula is on page 1866 of the Joint Appendix, and there are two terms. There's first an X with a bar over it, which is the average of the values that you're using. It's a mean. That's the standard notation for an X, a variable with a bar over it, plus a set of terms, a T with some subscripts, and a T is a statistical method. It's a student T test. It's a method that's used for predicting variability, and then a square root of the standard deviation, which is a characteristic of the sample set itself, and then two values that reflect the number of tests. It is the average value plus the statistical variability. So it does, in fact, reflect the average. I predict you're not going to get any questions about that. Okay. That's fine. I don't know. Maybe I shouldn't speak for my colleagues. But that's, you know, in terms of saying that's what the formula has always been. That's the formula we've used, and that's what it is intended to do. It's intended to predict the average of the best-performing sources over time, which the statute requires, because the statute requires tests to be achievable, standards to be achievable, and this court has recognized that standards should be achievable. In Sierra Club 167F3665, the court said that EPA could reasonably set the standard at the level the top-performing units can achieve under the worst foreseeable circumstances. The agency has discretion to decide what's the average, what is achievable, and what's being achieved by sources. And that's the purpose of the UPL. That's what it's doing. So it does achieve the average. It is reasonable. Petitioners rely a lot on their argument that for some sources the UPL is above the measured values. They have tables in their briefs showing that. But that's not unreasonable, because the variability, and UPL is looking at the variability in a large sample based on the variability found in the small sample, and by statistics it will always be larger, and therefore it will be higher value than most or perhaps even all of it. But it's also important when you look at those tables to look at the scales, because what we're looking at on the differences here are thousands and millions of parts of pounds per million BTU. We're talking about very small numbers and very small differences in numbers, not enormous amounts. And it just reflects standard operation of statistics. The variability of a large sample is going to be larger than the variability of a small sample, and if you want to set a standard that encompasses that entire universe of variability, it may well be higher than the individual test results. If there's no further questions on those major source issues, I will turn to the SISLI issues. And first I will talk about that EPA did not set any above-the-floor standards. Under Section 7429, standards must reflect the maximum degree of reduction the administrative terms is achievable, taking into account costs and non-air quality, health and environmental impacts, and energy. As used in the statute, achievable does not mean achievable with infinite resources. Any facility that can be achieved in some theoretical way, it means achievable in the real world. And particularly when going beyond the statutorily required floor. That's what we're talking about here. We're talking about either EPA's decision not to go beyond the floor, which is calculated, you know, on the basis of the best performing standards. But say the EPA should go beyond that, the statute clearly requires EPA to consider costs. And that's exactly what it did here. It examined the available control measures for SISLI and determined that there are no practical beyond-the-floor measures because of either excessive costs or energy use. You know, for example, petitioners have argued for the use of fabric filters for particulate matter control. The agency found those costs were excessive. In one case, to the tune of $350 million a ton for additional mercury removal. Petitioners don't really address the reality of these costs. Rather, they try to say that, and look at the terminology we use, and say that we talk about reasonableness, and reasonableness does not achieve ability. But when looking beyond the floor, the statute required, in order to decide what's achievable, you have to look at costs and you have to look at other impacts. How are you going to consider costs unless you are looking at what is reasonable, what's reasonable to spend, what are reasonable costs, given the availability of the controls and given the reductions that you're likely to achieve. And that's exactly what it did here. The agency determined that the amount of additional pollutant removal that would be achieved by these various technologies was minimal and that the costs would be excessive. Petitioners also seem to argue that if a technology is used in one subcategory, it must be considered achievable for another, but that's simply not the case. The whole purpose of subcategorization is the fact that these units are different. They have different fuels, they have different setups, and it's not necessarily true. So you can't simply assume that because some units in one category are using a controlled technology, others are not. And so the bottom line is EPA, reasonableness is the test in determining what is achievable when looking beyond the floor because you must consider costs. The only way you consider costs is by looking at how the costs line up, but not only with the ability of sources to achieve it, to prepare the costs, but also with the amount of reductions. And EPA did that analysis, it's documented in the record, and made a rational decision that there were no achievable beyond-the-floor standards. The second argument, the second issue under the SIFWG... Are there no other questions on that? The second issue on the SIFWG is the units that EPA determined were not subject to regulation. Burn-off ovens, cyclonic burn barrels, boundary sand reclamation units, soil treatment units in space. Are there plans to promulgate standards for those units? It's still under consideration. Is that a deadline? Well, this issue really came up during this rulemaking. The agency did not appreciate that there was a large universe of these units. These units are largely... Many of them are essentially parts of manufacturing processes, like burn-off ovens, boundary sand reclamation units. A burn-off oven is an oven that's used in the manufacturing process to remove contaminants from metal parts. It's a part to put in those units. It's heated up to drive off the contaminants. So you're saying the agency really wasn't aware of the extent of... They were not aware of the extent. They weren't aware of how they were used. Are there current plans to gather that data? I'm not aware that there are any specific current plans. It's certainly something the agency recognizes that needs to look into. Am I right? My notes are showing that, at least with regards to cyclonic burn barrels, that the agency has made a determination that it combusts solid waste? No, we haven't. Well, I don't know that we have. I don't know that we would argue about it. They're fairly... That is a waste of smoke for units. A 55-gallon bomb, it'll pop in a pan. How one would... If that's the case, don't you have a duty to set standards for that? We may well have... I think that's an issue. To the extent we have a duty... If we have a duty, and we may well have a duty to set standards, at least consider standards. I'm not sure how you would set standards for such a unit. But if you could, then we would have a duty. But if the petitioners believe that there's a mandatory duty that we failed to perform, the remedy is to go to district court and bring in action to enforce that mandatory duty. Because we have not made any determinations. We've not excluded them. Petitioners have said that we've made a determination that they're not going to be subject to regulation. And we have not. We have not made that determination with regard to any of these units. The agency simply decided, determined that it did not have the available data. It said they had made proposals with regards to some of these units, and then they got flooded with comments saying, wait a minute, you don't know what you're talking about. We don't... There's not 10 of these units, there's not 30 of these units, there's thousands of these units, and they're very different than what you were considering. So, we simply haven't made a determination on it, and we don't have the data. For some of these units, you know, there's probably... I mean, there is an open question, like the burn-off units, a soil... Foundry sand reclamation units are used in the automobile manufacturing industry, perhaps other places. They're... Sand is used to make temporary molds for metal parts, and the sand is reclaimed through a heat reclamation process. Again, these are parts of manufacturing. This is not what we were thinking of as SISWI. When we think of SISWI, we think of something else. Is there anything in the record that we could cite to, to support the argument you're making now, which is you haven't made a final decision on this? Well, for instance, JA-1245, in response to the comments, we specifically said we weren't addressing... We weren't addressing this issue at this time. And we... And we're supposed to take from at this time that it will be addressed at some future time? It's... The agency has a mandatory duty to regulate SISWI units. The agency will have to determine whether some of these units are SISWI units. It's still an open issue. I can't stand up here and tell you that there's a schedule or there are any sort of ongoing rulemaking, but it is certainly still something the agency has to address. Finally, let me turn to the area boiler diverted issues. And first, let me address the issue of the 7412C6 list and petitioner's argument that every category that ever was placed on this list must have, subject to MAC regulations, even if it has negligible emissions of the pollutants regulated by the section. Now, this section of the EPA is required to list sources of seven specific pollutants and assure that sources accounting for 90% of the aggregate emissions of each pollutant are subject to MAC standards. And with regard to area sources, what that means is that the agency would not have its normal discretion to utilize GAC sources. But it does not... But the substantive requirement here is not the listing. The substantive requirement here is that you achieve the 90% level. The list is basically an accounting tool so the agency knows which sources it needs to cover, can kind of keep track of whether it's met this, and ultimately make a decision. And the agency has made a determination. The initial determination that it had made the 90%, that was sent back for notice and comment issues, and has recently made another, and there is now a case pending in this court. I frankly don't have the site in front of me, challenging EPA's determination that it has reached that 90% threshold. But the agency... The list was never intended to be set in stone. It was never intended that once a source was set on this list, EPA would have to set MAC standards for it if it wasn't required by some other provisions. And the agency, in fact, specifically stated when it promulgated the list in 1998 that it would evolve as more information was achieved. As more, excuse me, as the agency acquired more information about the sources. And in fact, the list has been modified a number of times as some sources became subject to regulations, as the agency determined other sources were no longer, were to change their operations, were no longer major sources, or discovered other ones. And with area sources, EPA... First of all, the list never said area sources. It simply said coal-fired boilers, it said biomass-fired boilers, and it said oil-fired boilers. And the agency always said that the list was subject to modification. And in general, it said... And so with regard to the area sources, EPA determined, when it came to do the standards for these boilers, it determined that while coal-fired boilers are a significant source of mercury and so in the organic caps on the list, the biomass-fired boilers and the oil-fired boilers had minimal half emissions and were not needed on that list. And in fact, the coal-fired boilers, area source boilers, have many times higher emissions than the other two. And therefore, it removed them from the standards. Nothing in the statute requires that simply because EPA at one time said it thought it might need these units to make its 90% goal, that it's locked in forever to setting a max standard. The statute simply doesn't say that. It simply says, go out, figure out what sources you need to regulate, and make sure by the time you're done promulgating regulations that you've covered 90%. And that's exactly what EPA has done. It's covered the 90% and therefore it has not decided that for the area source biomass and oil-fired boilers, the max standards aren't required. Now, it did set GAC standards for those because those are covered under another provision for the urban air toxics. So it identified those as also required standards for those. So there are standards. But this is not a case where these units are being removed from Section 112 regulation entirely. EPA is simply being removed from this requirement which is contrary to EPA's normal expression that it has to apply max instead of having the option to apply GAC instead of applying a GAC standard. And that's why the court's decision in New Jersey is adopted. Because in New Jersey, the issue was whether EPA could remove a category entirely from regulation under Section 112 after having listed it without going through the C9 process. And frankly, C9, which is the provision for removing a source entirely from the list, requires certain health-based standards that really aren't relevant to the issue of whether it's needed for 90% of these particular pollutants or not. Turning to Title V, the exemption for Title V. The Act allows EPA to exclude... Title V is an operating permit requirement that was added in the 1990 amendment. It's not intended to impose substantive requirements on sources, but certain sources are required to get an operating permit. Partly, one of the rationales was for large sources that are subject to numerous requirements to pull them all into one place and it's for administrative purposes, but it allows EPA to exclude area sources. And EPA, in the rulemaking here, EPA determined area source spoilers should not be subject to Title V. And the factors EPA considers are whether permitting would significantly improve compliance requirements, whether it would impose significant burdens, whether the costs are justified, and whether the implementation enforcement programs are the relevant requirements are sufficient. And EPA determined that for area source spoilers, those were not necessary. And the petitioners don't challenge that for the vast majority, about 92,000 area source spoilers that are natural area source spoilers. But they do challenge it for what are called synthetic micro-sources, synthetic area sources. And these are sources that have an... What makes a source major or minor, in this case, is not the sort of absolute size of the unit or the power output, it's how much hazardous air pollutants it's capable of producing. And if it's above the threshold in the statute for it, that is a major source of hazardous air pollutants. And if it's below that threshold, it's an area source. In these cases, these are units that, if uncontrolled and run at their full capacity, would have the potential to emit above the major source threshold, but have gotten permits, either from EPA or the state, that impose specific requirements that limit their emissions of hazardous air pollutants to below the area source level. And in the proposed rule, EPA, based on experience with other rulemaking, had thought that these units would be similar to major sources and had proposed to include them. But then it received comments indicating, well, we know that these area source boilers are different. And what's really key here is, when you're considering, when you're looking at determination of whether a Title V permit is appropriate for an area source, what you look at and what's important is not the size of the boiler. It's what sort of institution or facility the boiler is located in. Because what you're concerned is, are the people who are operating this boiler, what are the costs for them? What are the burdens, the administrative burdens? And how likely or unlikely is it that adding Title V will contribute to compliance? For instance, area source boilers are located at a wide range of facilities, including schools, hospitals. Some of them are located at large industrial facilities. By and large, they are, the vast majority, are located at facilities that don't have a lot of tech support expertise, don't have a lot. And so in considering these issues of whether Title V is appropriate, you consider the cost, and you consider what sort of entity. And what EPA learned from comments is that the entities that run these natural, these synthetic area source boilers are not that different, and they're not different from the ones that run natural source boilers, area, excuse me, talking to the audience, natural area boilers, and therefore the same rationale that applied with the natural ones applies here. And the petitioners haven't addressed that. They keep talking about the size of the boiler. The size of the boiler isn't relevant. And they also talk about, you know, and there is, and this is, you know, cost of obtaining a permit is not necessarily trivial. You know, J384 in Dow Chemicals promised that it can cost as much as $50,000. Okay, I want to stop you there because I'm not sure that I'm following your argument here. You said the size of the boiler is not relevant for this purpose. But when you're talking about these synthetic boilers, it would seem to me that size would be part of what you're looking at because they're synthetic only because they have agreed to controls which makes them emit less. And so they can fall under the area boiler definition. Correct? That is a factor, but looking at the factors EPA considers for the Title V permitting is appropriate, which are the cost and the burden, you have to look at the entity that's offering the boiler, whether it's a school or a church or a chemical plant or an electric utility. I guess what I'm trying to get at is I don't see in this record, and perhaps you can tell me where it is, where EPA actually did that analysis. I mean, I understand that there are certain factors that EPA looks at to decide whether it's useful to have them under this Title V requirement. But I don't see that here. And what I do see is that at least in 2010, EPA went through something like this same exercise and said, well, they should be under the Title V requirement. So what I'm trying to find out, and this is a record that has a lot of material in it, but where does EPA decide that we're going to make a change here and then explain why we're doing it? Well, in both the 2013 rule and the reconsideration, and I confess it's not documented in as much detail as you might like looking back on it, but what EPA said, and this is at JA-124, was that its original proposal was based on its belief that these units were founded, that the types of facilities that were typical of nature was based on its experience with other rulemaking, but that based on comments, it had determined that they were not. In fact, they're just the same types of facilities as at the natural source boilers. But why other rulemakings? I mean, I'm just trying to understand what's going on here. And on the one hand, you have boilers that would be major, but for the fact that they have voluntarily agreed to do something, use some kind of controlled technology that reduces their emissions to a level where they can be treated as area boilers, correct? That's correct, John. Okay. So then they have a history with EPA, do they not? I mean, they had to reach this agreement, or do they do it with the state? It's more likely it was done through the state. Okay, it was done through the state. So what you're saying is that the fact that they became area boilers was not necessarily the result of some kind of interaction with EPA. No, not at all, Your Honor. So EPA simply accepts that they are area boilers because of their emissions level? Because they have these enforceable controls. That's why they are. Okay. And it's done by the states. I'm sure EPA just pulled the information and leaves it off the table. Okay. But EPA made some decision about whether they should come under this Title V requirement in one rulemaking. They didn't change their mind. And all I'm trying to understand is... Well, the first was a proposal, Your Honor. Okay. And that was a proposal. They proposed that. They proposed it. And then they got comments. And after they got comments and they thought about it, they decide maybe we're not, we shouldn't do it that way. That's correct, Your Honor. But it looks like the factual findings that, you know, the data that was the basis of that proposal, I don't see in the record where that was changed. Well, again, Your Honor, I confess it's not terribly clear. But they basically received comments. And based on the comments, they determined that their initial assumption in the proposal was that these would be located at large, sophisticated sources for which a Title V permit would be appropriate. They received comments from users saying, no, that's not true. These are more standalone. It's not, at least in my, I don't have anything handy. Certainly, we can look through and try to send you something else if you'd like, Your Honor. But that is, you know, that's what they stated in the rule, that they received comments. The comments demonstrated that these boilers, the types of facilities were essentially the same and, therefore, the same rationality. They didn't find all the same other facilities. And, you know, the fact that they're synthetic boilers is one factor. But it has to be balanced against the cost and the burden on the institution of obtaining it. And in terms of even looking at the factors, Your Honor, while, yes, these are large, could be larger, and could have higher emissions, they are really subject to essentially two levels of control here. They are subject to whatever the state control is, presumably state or federal control, that makes them a minor source boiler. And they're also, of course, subject to the area source standards themselves. So they are subject to two levels of control. And there's nothing that shows that putting that information into an operating permit would make any difference to their ability to comply or the ability of the state or EPA or anyone else to enforce it. Okay, just checking where I am. Okay. I think I'm down to one last issue, which is the EPA's determination of what constitutes the GAP standard for metals. And it's really not clear to me what EPA thinks. You know, the petitioners have argued that EPA should have exercised discretion in deciding whether to use MAC standards. But I really don't see what it is that they believe that EPA should have done differently here. EPA, in analyzing the available controls for metals for area sources, first of all, it decided that these sources, and we're talking about the ones that it decided did not need MAC controls for, so it's the biomass and oil fire sources. They looked at and reviewed what area sources are using and determined that most of them have no controls on them at all. And those they do have a multi-club, which is designed to form a dust collection unit, anything else that's not particularly applicable. And that there simply just wasn't anything else that was practical or usable. And they determined that certainly for existing source standards, that retrofitting with a multi-club is impractical for both size and cost. And that there's simply no basis to impose any higher level of control because there were no additional controls that could be placed on existing sources. And with regard to new sources, again, this is the only source that was found for multi-club. And so it's not at all clear what EPA was supposed to do. They looked at all the controls. Maybe they did not say, oh, you know, we could have done that, but we're not going to. But I think it's very clear that there was nothing, no reason to. Particularly for the metals or the oil and biomass, they've already determined that the levels of, have emissions from those is quite low, which is why they didn't include them on the 90% list. They determined that there was no additional practical control unit. And therefore there was no, clearly no reason to impose a different level of standard other than what they determined to be that. All right. Petitioners talked about the work practice standards. My understanding was that was not on the list of issues to be discussed today. As I think our brief makes our argument well, that we believe that the work standards were appropriate and met the ethical statutory standards. The court has no further questions. All right. Thank you. We'll hear from the interveners. Thank you, Mr. Friedland. Just one point I'd like to make, and that is to respond, Judge Griffith, to your question. On EPA's decision to defer regulating burn-off ovens, et cetera, is there any record evidence that supports that? Indeed there is in the intervener brief at pages six and seven, the intervener brief at pages six or seven, we set forth the argument that there is the threshold question of whether those units are solid waste incineration units. And we cite to the ACC comments, the American Chemistry Council comments, which is JA601 to 602, that, for example, for burn-off ovens, the purpose of a burn-off oven is to clean parts for reuse, to clean the gunk off the parts for reuse. It's not to incinerate. Indeed, we want the part to remain because the part is going to be reusable. And so, therefore, we use pyrolysis or those kinds of methods to clean off the gunk so that the residual metal can be reused. And, therefore, we argue to EPA that that's not an incinerator. It's not a CISWE. It's not incinerating. It's designed not to incinerate. And, therefore, it's not a CISWE. And, therefore, EPA's conclusion to defer regulation of those units and certain others as CISWE sources is well supported in the record. And, again, that's our brief at 6 and 7, and then the ACC comments at 601 to 602 of the joint appendix. Thank you. Mr. Wuerl. Thank you, Your Honor. I'd like to make four brief points in support of EPA. The first is on the use of the UPL, so-called UPL Upper Predictive Limit. There's a key issue that has not been raised in this argument that significantly undermines the environmental petitioner's argument here. And this, in their brief, appears on pages 37 to 38. And there, what the environmental petitioners, they acknowledge, and, in fact, do not dispute their words, that well-established case law in this court allows EPA to consider variability in determining what appropriate so-called floors are when establishing standards under Section 112. But the environmental petitioners argue those cases have no relevance to this question as to whether the UPL is appropriate. We disagree with that and believe that the government made exactly the right argument here, which is the emissions test data that are taken and that EPA were required to be taken, that, you know, represents most of the data they relied on to set the standard, are snapshots in time. And, you know, EPA's test methods require you to take the test under optimal conditions. So the purpose of the UPL is to say, if you took that snapshot at another time when conditions are not optimal, what would you expect the average emissions to be? And so that's a method that entirely comports with this court's case law that says EPA can and should consider variability in determining standards under this part of the Act. So that's point number one, Your Honors. Point number two, the plaintiffs or the environmental petitioners, and I'm shifting to CO as a surrogate, made a couple of arguments, both in their brief and in their remarks today. One is that levels of carbon monoxide can actually increase while levels of organic pollutants decrease. This is in Petitioner's Brief, page 24. And Judge Griffiths gets to a question you asked, Mr. Rabe. So that argument was raised, and what the environmental petitioners cite to in support of that argument is comments in the record submitted by Southern Company at JA 798-800. So in our response brief, we addressed this issue that's on page 18 of our response – I'm sorry, it's on page 19 of our response brief. And, in fact, what Southern Company says in those comments is, as temperature goes up, the amount of organic half, you know, the amount that's destroyed continues to increase. So the levels of organic half go down as temperature goes up. But at a point, you see a blip with CO. And inexplicably, CO goes up for a period of time as temperature goes up and then comes back down. And that's what Southern Company pointed that out in their comments. But they pointed that out to make the opposite argument the environmental petitioners are here. Southern Company said, as a result, CO is too conservative as a surrogate because it will show higher levels at some times when you don't have higher levels of organic half. So Southern Company said, you know, it's too conservative and EPA shouldn't use it. From EPA's perspective, that's okay. Being too conservative is all right. You know, the opposite would be a problem. That's not what Southern Company was complaining about in their comments. Thirdly, the environmental petitioners also on page 24 of their brief and in remarks today said, organic hazardous pollutants are reduced by means other than those that reduce carbon monoxide. And, again, Judge Griffith, you asked questions on this before. And in our response brief, we look very, very carefully at that, and that shows up in our response brief at page 18. Those assertions are supported by comments submitted by a trade association called ACAC, the Association of Clean Air Companies. And when you look at what they said, they said two things. They talked about catalyst, which is one form of pollution control, and they talked about carbon injection, which is another form of pollution control. The environmentalist petitioners cite that report for the proposition that, you know, these comments support the idea that there's not a – these types of air pollution controls do not equally control carbon monoxide and organic half. When you look closely at what they say, that's not what they say at all. The ACAC report for catalyst actually says catalyst is effective in controlling both CO and organic half, so it says the opposite of what the environmentalist petitioners are arguing here. And for carbon, all they say is carbon is not effective in controlling CO, period. There's no data in their comments to support that. It's just a bald assertion, and as such, it should be given no weight by the court. Lastly, the environmentalist petitioners cite on page 6 of their brief, and this again came up in the argument, that one of the problems they had with carbon monoxide as a surrogate is there's evidence in the record showing that as carbon monoxide goes down, below a certain point, formaldehyde appears to go back up. That's on page 6 of their brief. And they say that's the reason this surrogacy relationship is no good. EPA addressed that express issue in the record. It shows up in JA-377, and in JA-377, EPA said, hey, we realize that's what those data seem to show. We don't think those data are right. We think those higher formaldehyde concentrations are a result of measurement error, and therefore we don't think that upsets the relationship. So EPA took it on directly. Thank you, Your Honors. I appreciate the opportunity. Thank you. Mr. Pugh has how much time left? Okay. EPA's lawyer represented that the upper prediction limit is not a level that any test by any source will fall below. I'd just like to read from the record on page 911 where EPA says, it is the level EPA, I'm sorry, let me actually get the quote for this. The upper prediction limit represents the value we can expect the mean of three future observations of three run average to fall below. In other words, if we were to randomly select a future test condition from any of these sources, we could be 99% confident that the reported level will fall below the max floor emission limit calculated using EPA. EPA also says at 1859 that it's an upper limit, and at 1861 that it rejects another approach, as I said, because it can be met. Another approach it's saying is unaffectable because it couldn't be met by all the existing floors at all times. So that, I think that representation is simply wrong. EPA also says that it took all the facilities and predicted their test as for what an average would be. Well, it did look at the facilities and it predicted what the range of their emissions would be, but then it didn't set an average. It set the standard at the far upper limit of what they would achieve. It didn't, there's not, that limit was not an average. What EPA looked at was the full range, and you can see on page 1864 it shows bell curves. On 1865 it gives an example, and the actual limit EPA set is off the right-hand side of the chart because it's an extreme upper limit. It doesn't purport to be an average. Well, EPA calls it an average in the memo, but what EPA actually says when it's describing it is that it's not an average. EPA's lawyers once again said that it's inherent in the formula. Before I address that point, I would just like to say that that's really their claim. The sport should reach the same result as it did in NAFTA, where EPA also claims it's inherent in the formula. But what I would like to point out in that formula is the T value, which Mr. Rigg referred to, and I don't want to go into the details, but that's the value. That's the way that EPA ends up setting the standard at the far right-hand side of the curve. EPA also says that EPA can take account of sources' emissions during their worst foreseeable circumstances, and this court held that in NAFTA. We're not disputing that point. EPA could look at what the sources did under the worst foreseeable circumstances, but it would still have to take an average, and that's what it didn't do here. It set a level that any source would meet under any circumstances. Finally, EPA's lawyers are suggesting that this is sort of an unimportant problem because the numbers are small. The numbers actually are very significant because they allow boilers to emit many times more than they would if EPA set valid standards. That's why if you look at the record, it shows that boilers are emitting six or seven tons of mercury every year. These standards reduce them by less than 2%. So these emissions, the absolute numbers may be small, but setting standards at the right level would reduce them by many times, and when you look at the entire category, the significance would be very big. If I can just take a couple points to talk about the waiver claim that EPA's lawyers made on the exclusion of best sources. If you look at our brief, it's not waived at all. We say very clearly that EPA's exclusion of both the sources burning more than 10% and sources burning any violates the law, and in our facts description, we describe both the existing source floors and the new source floors. There's absolutely nothing in our brief that suggests we were only talking about new source floors. I'm not sure why EPA read it that way, but that wasn't our point, and we didn't waive the argument. What EPA didn't say, and I don't think, is that excluding best-performing sources is lawful. There's simply no statutory basis for EPA to make this claim, because these are undisputedly best-performing sources. They're undisputedly sources for which EPA had emissions information, and the statute clearly requires floors to be based on the best-performing sources for which EPA has emissions information. The last point, with the court's permission, if I could just address the carbon monoxide point. No one is – I didn't hear any dispute that sources can reduce their carbon monoxide emissions by using – I'm sorry, excuse me. I didn't hear any dispute that sources can reduce their emissions of polycyclic organic matter by using activated carbon injection or selective catalytic conversion. The makers of these control technologies are the ones who said they reduce those pollutants and not carbon monoxide. Industry intervenors may have a different view, but EPA's job is to look at the information in the record, and the information in the record came from the makers of the control technology, and that's what they said. Your Honor asked a couple of questions. One was about temperature. If you look at a study in the record at page 491 of the Joint Appendix – or actually, I think it's slightly later in the Joint Appendix – the conclusion in that study is that increasing temperatures reduces PAHs dramatically and has no effect on – or relatively little effect – or actually, excuse me, it causes CO emissions to go up, so the answer to your question on that point is yes, there is evidence in the record. It's in that study, and the conclusion at the very end of that study makes the point. The other study, I think, which goes directly to EPA's representation that keeping carbon monoxide levels to a certain – to the 130 parts per million standard destroys all organic hazardous air pollutants is actually in the – that's in the 70 to 80 hour temperature decrease. It's not in the 10 to 80 hour decrease. If you look on page – on table 812 in that study, you'll see what the PAH levels are, according to EPA itself, at various carbon monoxide levels. And at the level of sources that are meeting the 130 parts per million carbon monoxide level, this table shows that PAH levels are still 157. That's not zero. All the PAHs are not being destroyed by keeping carbon monoxide down to the standard – to the 130 parts per million standard. And more importantly, for the purposes of surrogate arguments, that doesn't even report to show what the best sources can achieve and are achieving with respect to PLM. All right. How much time does Mr. Wolbein have? One minute, Mr. Wolbein. Okay. Quickly, with respect to the beyond-the-floor standards in the SESWI case, EPA says that it has to consider cost in setting beyond-the-floor standards. We don't dispute that EPA has to consider cost in setting beyond-the-floor standards. But it must relate its findings about cost to the correct statutory standard. It hasn't done that here. EPA has discretion in how it does that, in how it considers the statutory factors, and in what it determines is achievable. But EPA can't just jettison the word achievable and replace it with the word reasonable, because those are different terms with different meanings under the statute. EPA denies that it took final agency action with respect to the exempted units. That's incorrect, because EPA defined them out of the pollution limits, and because EPA claimed that this rulemaking is its final response to the previous remands, and EPA also claimed that no additional standards are required. Even if all EPA had done was make a decision not to act, and, again, that's not the case, but even if it were a considered decision not to take action, especially after taking comment and especially when the deadline for taking action was long ago, is final agency action that's reviewable under this Court's precedence. And I would point to Fox Television Stations, 280 F. 3rd, 1027, and also Hercules v. EPA, 938 F. 2nd, 276. And finally, with respect to Title V, EPA defends the exemption on the basis of its balancing of various factors, but Judge Brown, as you pointed out, they never did that balancing for the synthetic area source spoilers. You didn't find it because it's not in the record anywhere. EPA simply claimed after reversing course that they were equivalent to the natural area source spoilers, and there's no evidence in the record to support that, and EPA still hasn't pointed us to any. Thank you. Let me just say on behalf of my colleagues, I hope we agree on the cases, but we certainly agree that we have listened for the last three-and-a-half hours to really top-flight professionals. You all have been very helpful to us, and we thank you.