Valley Citizens for a Safe Environment v. Edward C. Aldridge, Etc.

BREYER, Circuit Judge. The United States Air Force has transferred 16 C-5A airplanes from Dover, Delaware to Westover Air Force Base in western Massachusetts. Valley Citizens for a Safe Environment, an association of local residents, opposes the transfer, primarily because its members believe the airplanes are too noisy. They claim that the Air Force did not prepare a proper Environ *459 mental Impact Statement (“EIS”) before deciding to send the planes to Westover. After examining the record, we conclude that the Air Force’s Final Environmental Impact Statement adequately sets forth the likely “environmental impact of the proposed” transfer, including any unavoidable “adverse environmental effects” and reasonable “alternatives to the proposed action.” 42 U.S.C. § 4332(2)(C)(i-iii). The Air Force, in promulgating that Statement, acted lawfully. We affirm the district court’s similar determination. I Background When the Air Force received new cargo airplanes in 1982, it decided to transfer 16 older C-5As from Dover Air Force Base in Delaware to a new location. The Air Force wanted to assign the C-5As to the Air Force Reserve’s 439th Tactical Airlift wing, located at Westover Air Force Base, Massachusetts. It recognized that the C-5As made more noise than the 16 C-130s currently at Westover. And, on September 26, 1985, the Air Force met with local citizens to discuss the problem. After the meeting, the Air Force prepared a Draft Environmental Impact Statement; it released the Draft for public comment on December 5, 1985; and it closed the comment period fourteen months later, on February 11, 1987. On April 10, 1987, after revising the draft Environmental Impact Statement in response to the comments received, the Air Force published a 173-page final Environmental Impact Statement with five technical appendices. Subsequently, after considering likely environmental effects, as revealed in the Statement, the Air Force decided to transfer the C-5As to Westover. On June 30,1987, after the Air Force had transferred some, but not all, of the planes, Valley Citizens brought this legal action. It asked the federal district court to enjoin transfer of the planes on the ground that the Air Force had not prepared an adequate Environmental Impact Statement before making its transfer decision. See 42 U.S.C. § 4332(2)(C); Robertson v. Methow Valley Citizens Council, — U.S. -, 109 S.Ct. 1835, 1846, 104 L.Ed.2d 351 (1989) (agency must adequately identify and evaluate adverse environmental effects of proposed action). Valley Citizens said that the Environmental Impact Statement was not adequate in that it failed to take account of alternatives to the Westover transfer; it failed properly to describe potential adverse air pollution effects, and it failed to predict just how irritating airplane noise would prove to be. See, e.g., Vermont Yankee Nuclear Power Corp. v. NRDC, 435 U.S. 519, 98 S.Ct. 1197, 55 L.Ed.2d 460 (1978) (NEPA requires discussion of feasible alternatives to the proposed action); Sierra Club v. United States Department of Transportation, 753 F.2d 120 (D.C.Cir.1985) (noise relevant environmental concern of proposed action). The district court, after reviewing the record before the administrative agency (the Air Force) and examining the parties’ additional evidentiary submissions, granted summary judgment for the Air Force. Valley Citizens appeals. After examining the record, we conclude that the district court was legally correct. II Standards The National Environmental Policy Act (“NEPA”) requires an agency to prepare, in respect to proposed “major Federal actions significantly affecting the quality of the human environment,” a “detailed statement,” describing, among other things, the “environmental impact of the proposed action ... any adverse environmental effects which cannot be avoided, [and] ... alternatives to the proposed action.” 42 U.S.C. § 4332(C). In a typical challenge to the adequacy of such a statement, a reviewing court will apply “a reasonableness standard ... aimed at insuring a good faith effort by the Agency.” Conservation Law Foundation v. Andrus, 623 F.2d 712, 719 (1st Cir.1979); Silva v. Lynn, 482 F.2d 1282 (1st Cir.1973) (review of EIS governed by APA “arbitrary and capricious” standard, 5 U.S.C. § 706(2)(A)). In applying *460 this standard we recognize NEPA’s basic objective, namely to inform the agency and the public, before the agency makes a final decision, about what adverse environmental effects might occur and whether less harmful alternatives are likely to be available. Robertson v. Methow Valley Citizens Council, 109 S.Ct. at 1845-46. We must ask whether, in light of this objective, the agency has carried out NEPA’s mandate in a reasonable way. In a typical case, a reviewing court, in answering this legal question, looks first and foremost at the record before the agency. That is because one cannot ordinarily expect an agency to do more than make reasonable efforts to gather relevant information and then to evaluate that information in light of the comments interested parties have made. The relevant legal question therefore is normally whether the Statement is “adequate” in light of the information and comments before the agency at the time it produced the Statement. Commonwealth of Massachusetts v. Watt, 716 F.2d 946 (1st Cir.1983); Roosevelt Campobello International Park v. United States Environmental Protection Agency, 684 F.2d 1041, 1046 (1st Cir.1982). And, the record compiled by the agency will often contain sufficient information to permit the court to make this judgment. 0Changes in circumstance are relevant to the different legal question of whether the agency must prepare a supplement. See Commonwealth of Massachusetts v. Watt, supra; 40 C.F.R. § 1502.9(c).) A court’s tendency to review the legal adequacy of an Environmental Impact Statement on the basis of the record before the agency also may reflect the fact that such review often takes place in a court of appeals, a court that is not well equipped to try factual disputes de novo. E.g., 28 U.S.C. § 2342 (granting court of appeals exclusive jurisdiction to review certain final orders from certain commissions, including the Nuclear Regulatory Commission). See Sierra Club v. Marsh, 769 F.2d 868, 872 (1st Cir.1985). And the degree to which a party is, or is not, legally free to build a new record ought not to reflect simply the happenstance of which level of reviewing court a particular jurisdictional statute selects. Thus, it is not surprising that the Supreme Court pointed out in Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 420, 91 S.Ct. 814, 825, 28 L.Ed.2d 136 (1971), and explicitly stated in Camp v. Pitts, 411 U.S. 138, 142, 93 S.Ct. 1241, 1244, 36 L.Ed.2d 106 (1973) that “the focal point for judicial review should be the administrative record already in existence, not some new record made initially in the reviewing court.” See also Florida Power & Light Co. v. Lorion, 470 U.S. 729, 743, 105 S.Ct. 1598, 1606, 84 L.Ed.2d 643 (1985). The fact that review sometimes or often focuses on the initial administrative record does not mean it must, or always, will do so. It could happen that a particular instance of judicial review of an EIS raises a “genuine” and “material” dispute of facts that requires a trial: Did the agency know, for example, about some important matter that the EIS ignored (and which the commenting parties did not know about and could not have pointed out?). See Friends of the Earth v. Hintz, 800 F.2d 822, 829 (9th Cir.1986); County of Suffolk v. Secretary of Interior, 562 F.2d 1368, 1384-85 (2d Cir.1977), cert, denied, 434 U.S. 1064, 98 S.Ct. 1238, 55 L.Ed.2d 764 (1978). Or, did the agency improperly rely upon some other important, but secret, information not part of the record? Friends of the Earth v. Hintz, supra. Moreover, a reviewing court might want additional testimony by experts, simply to help it understand matters in the agency record; indeed, it might ask for additional factual evidence as an aid to understanding. However desirable this kind of evidentiary supplementation as an aid to understanding highly technical, environmental matters, its use is discretionary with the reviewing court. Love v. Thomas, 858 F.2d 1347 (9th Cir.1988), cert. denied, — U.S. -, 109 S.Ct. 1932, 104 L.Ed.2d 403 (1989); Asarco, Inc. v. U.S.E.P.A., 616 F.2d 1153 (9th Cir.1980). We mention some of the circumstances in which an EIS review might lead to a trial to distinguish those circumstances from the present case. Here, the district court de- *461 cided not to exercise any of its discretionary power to hear oral expert testimony. Moreover, here there is little “genuine” dispute about the information that was before the Air Force when it promulgated the EIS. The questions before us, rather, are legal: (1) Given the information before the agency, should the Air Force, in its EIS, have analyzed alternative locations for the C-5As more closely? (2) Did the EIS make a significant error in its estimate of the likely C-5A-eaused increase in nitrous oxide emissions? (3) Did the EIS properly analyze likely increases in noise levels? Although the parties’ briefs make factual claims that diverge, so widely that one is tempted to think some “material” factual issue must be in dispute, a close reading of their briefs, and an effort to track those divergent claims to their sources in the record reveals that there are no “material” disputes of fact. The parties argue about how many people the C-5A noise would irritate and how many pounds of nitrous oxide the plane will emit, but the answers to those factual questions do not determine the answer to the relevant legal question, whether the EIS reasonably, hence adequately, discussed these issues in light of the information the Air Force had, or should have had, at the time. Sierra Club v. Marsh, 872 F.2d 497 (1st Cir.1989). To put the same point differently, even if we assume the answers to “genuine” factual disputes in Valley Citizens’ favor, we must find the Air Force’s discussion of “alternatives” adequate, its mistake about nitrous oxide inconsequential, and its noise analysis permissible. Our reasons for reaching these conclusions follow. Ill The EIS’s Discussion of Alternatives Valley Citizens first argues that the EIS did not adequately describe “alternatives to the proposed action.” 42 U.S.C. §§ 4332(2)(C)(iii), 4332(2)(E). In particular, Valley Citizens says the EIS should have discussed further the possibility of sending the C-5As to other Air Force bases. The legal question, as we have said, is whether the EIS’s discussion of the alternatives was a reasonable one. Vermont Yankee Nuclear Power Corp. v. NRDC, 435 U.S. at 551, 98 S.Ct. at 1215 (NEPA does not require discussion of alternatives that are remote and speculative possibilities); Roosevelt Campobello International Park Commission v. United States Environmental Protection Agency, supra, (agency should study alternatives that are reasonable and appropriate); Grazing Fields Farm v. Goldschmidt, 626 F.2d 1068 (1st Cir.1980) (courts cannot require agencies to include in an EIS alternatives that are too fanciful or hypothetical). And, we think it was. A. The EIS Discussion. The EIS says that the Air Force, in deciding where to locate the planes, applied several critical now-environmental criteria, including (1) the adequacy of physical facilities, such as runways, ramps, etc.; (2) the recruiting potential for reservists in the base area; (3) the costs of additional needed construction; (4) the relationship to existing base uses; and (5) the adequacy of fuel systems. The EIS then states that the Air Force has examined five other possible locations: Orlando International Airport, Florida; Patrick Air Force Base, Florida; Cape Canaveral Air Force Station, Florida; Charleston Air Force Base, South Carolina, and Hunter Airfield, Georgia. The EIS goes on to describe how each of these locations fails to meet critical non -environmental criteria. In about two single spaced pages, it says that moving the planes (a) to Orlando would require spending $83.4 million for construction and would require buying additional land (which is not available); (b) to Patrick would require spending $83.4 million for construction plus the cost of filling in 70 to 150 acres of the Banana River because there is no land available to buy; (c) to Cape Canaveral would require spending about $138.6 million for construction and would also risk interference from missile launches; (d) to Charleston, would require spending only $23.6 million in construction, but it would also require moving 29 C-141s at a cost of double the construction estimates, and “the ability to recruit additional personnel necessary to operate a C-5A unit is doubtful;” (e) to Hunter Air *462 field would require spending $79.1 million for construction and “the demographics of the Savannah area are not considered adequate to recruit personnel with the skills necessary for a C-5A operation.” Finally, by way of comparison, the EIS says that moving the C-5A to Westover would impose added construction costs of about $46.9 million, significantly less than every other base but Charleston (where recruiting the 1,000 or more additional personnel would prove difficult). B. The reasonableness of the discussion. 1. The environmental discussion. The lack of discussion of the likely environmental effects of stationing the C-5As at Orlando, Patrick, Cape Canaveral, Charleston, or Hunter was perfectly reasonable. The EIS makes clear that the Air Force will not send the C-5As to the other bases because of significant added construction costs or recruitment problems. It will not send them irrespective of environmental effects at those other bases; it will not send them even if there are no harmful environmental effects, even if no one in those areas thinks the planes are too noisy. What purpose, then, could a discussion of environmental effects at those other bases serve, at least as long as the Air Force makes clear it is prepared to evaluate those alternatives on the assumption that their “adverse environmental effects” are zero? See 40 C.F.R. § 1502.14(a) (if alternatives are eliminated from detailed study and consideration, the agency shall briefly discuss the reasons why the alternatives have been eliminated). 2. The non-environmental discussion. The EIS’s discussion of the wow-environmental aspects of the other bases was brief but adequate. For one thing, the Air Force’s reasons for not sending its C-5As to those other bases, on their face, seem reasonable. For another thing, none of the comments that any person made about the draft EIS suggested that the Air Force go into the matter more deeply. In Roosevelt Campobello International Park v. U.S.E. P.A., supra, we made clear that the agency’s duty under NEPA is to study all alternatives that “appear reasonable and appropriate for study at the time” of drafting the EIS, as well as “significant alternatives” suggested by other agencies or the public during the comment period. In order to preserve an alternatives issue for review, it is not enough simply to make a facially plausible suggestion; rather, an intervenor must offer tangible evidence that an alternative site might offer a “substantial measure of superiority” as a site. Id. at 1047. (Emphasis added.) The only relevant comments we have found consist of requests to send the C-5As to “a remote location” or to a base where takeoffs and landings would be “over water.” The Air Force in its EIS responded adequately to those comments, pointing out that the only practical “remote” location was Cape Canaveral; that the only “over water” locations were Cape Canaveral and Patrick; and that cost and other wow-environmental objections made those alternatives impractical. Nothing in the agency record, or in the comments made to the draft, or in the record before us, as expanded in this judicial proceeding, points out any inaccuracy in the Air Force’s cost descriptions, or in its discussions of other non-environmental considerations. And nothing whatsoever suggests that any of the other bases enjoys (in the words of Roosevelt Campobello) a “substantial measure of superiority” to Westover. Without some good reason to doubt the accuracy of the Air Force’s wow-environmental conclusions (hence to believe that adverse environmental consequences might have played a role in ruling out these other bases), one cannot fairly criticize the EIS’s discussion of them, at least not when one reads that discussion in light of the comments made on the EIS draft. Cf. Vermont Yankee Nuclear Power Corp. v. NRDC, 435 U.S. at 553, 98 S.Ct. at 1216 (“while it is true that NEPA places upon an agency the obligation to consider every significant aspect of the environmental impact of a proposed action, it is still incumbent upon intervenors who wish to participate to structure their participation so that it is *463 meaningful, so that it alerts the agency to the intervenors’ position and contentions.”) Valley Citizens responds that the EIS discussion of “alternatives” should have seemed obviously far too abbreviated irrespective of the comments, and this fact itself would be obvious were the agency action at issue the building of a large nuclear power plant or the building of a double deck bridge, or the dumping of millions of tons of refuse. But the problem with this argument is that the agency action at issue is not the building of a power plant or a bridge or ocean dumping; the action consists of moving some Air Force planes from one base to another. And, what counts as a “reasonable” or “adequate” discussion in an EIS depends upon the circumstances, including the nature of the action at issue. Manygoats v. Kleppe, 558 F.2d 556 (10th Cir.1977) (kind of impact statement depends on kind of federal action); Sierra Club v. Lynn, 502 F.2d 43 (5th Cir.1974) (adequacy of EIS must be determined in light of nature of federal action), cert. denied, 421 U.S. 994, 95 S.Ct. 2001, 44 L.Ed.2d 484 (1975). What counts as a reasonable discussion in the context of a proposed Defense Department decision to move a battle group east to make it more readily available should trouble arise in the Middle East might seem not at all reasonable or adequate in the context of a plan to build a major oil refinery at the gateway to a national park. To insist that, in both sorts of instances, an EIS must reflect the same number of hours of work, or embody the same time-consuming and deeply probing investigation is to misuse the words “reasonable” and “adequate.” And, to force every EIS to fit any such procrustean bed would risk a procedure that, in some instances, might unnecessarily prevent an agency from taking timely and critical action, and in other instances, might fail to produce the detailed explanation of environmental effects needed for protection of the environment. NEPA’s success in large part arises from the use of legal concepts such as “reasonableness” and “adequacy” that permit courts to adapt it successfully to so many different kinds of circumstances surrounding so many different kinds of governmental decisions. Valley Citizens adds that the Air Force, in considering alternatives, should have taken account of potential adverse environmental effects and added costs at Westover arising from the cold winter weather. We have found only one comment on this matter, however, a letter from the Air Force’s regional civil engineer, stating that, if the Air Force intends more de-icing operations than at present, it should amend its spill control plan to take account of the risk of spilling antifreeze. The Air Force responded adequately to this count by saying it would do so. We have read and reread the Freedom of Information Act letter that Valley Citizens refers to in its reply brief without finding anything that one could fairly interpret as calling the Air Force’s attention to a need to say something about some special “cold weather” environmental (or other) effects. Nor was this need somehow obvious. Consequently, in this respect, too, the EIS, viewed as of the time it was made and in light of the comments, is reasonable and adequate. IV Nitrous Oxide Valley Citizens claims that the EIS failed accurately to portray the extent to which the transfer would increase air pollution, specifically by increasing concentrations of nitrous oxide. They have shown (and the Air Force concedes) that the EIS did not take account of the extra (i.e., compared to the C-130s) nitrous oxide that the C5-As would emit during some of the airborne portion of their flight “sorties” and during engine testing. The legal question before us is whether these emissions were “significant.” Conservation Law Foundation v. Andrus, 623 F.2d 712, 719 (1st Cir.1979) (a minor deficiency in an EIS does not entitle the court to disregard the deference the agency is entitled to); Commonwealth of Massachusetts v. Andrus, 594 F.2d 872, 884 (1st Cir.1979) (a court “may not use minor lapses in the statement as an excuse to thwart actions that it believes to be unwise ... or require of the discussion a *464 degree of detail too exacting to be realized”). We conclude that the EIS’s inaccuracies about nitrous oxide were not significant. A. What were the inaccuracies? We have used the parties’ briefs and related record references to determine the precise nature of the factual dispute between them and therefore the size of the inaccuracy we must assume for present purposes. We end up assuming that, for summary judgment purposes, a court could at most find that the EIS failed to take into account about 75 tons of nitrous oxide per year. We shall first explain how we reached this conclusion and then discuss its significance. 1. The EIS calculations. The EIS calculated the extent to which flying C-5As might add to air pollution by taking the following four steps. Step One. The EIS described what would occur on a typical C-5A training “sortie,” as follows: A typical C-5A training sortie would provide training for four flight crews. Initially, two crews would board the aircraft, start the engines, taxi to the end of the runway, perform an engine runup, and then take off. The aircraft would fly for approximately 2lh hr, with flight crews changing places at approximately the midpoint of the flight. At the end of the 2V4-hr period, the aircraft would land and then taxi to the operations area where two additional flight crews would board the aircraft and the original crews would deplane. The engines would remain running during the crew change. Following the crew change, the aircraft would taxi to the end of the runway, and take off and fly for approximately 2xh additional hours, with the aircrews changing places at approximately the midpoint of the flight. At the conclusion of the flight, the aircraft would land, taxi to the ramp area, and the engines would be shut down. Including the time required for startup and engine check, engine running, crew change, and shutdown, a typical training sortie would last approximately 5 hr_ During the sortie, the aircraft would make two takeoffs from a full stop, two landings resulting in full stops, and approximately 18 touch- and-go landings or low approaches. (In a touch-and-go landing, the aircraft touches down on the runway, rolls for several hundred feet, and then takes off again without stopping. In a low approach, the aircraft descends to approximately 50 ft above the runway and then departs without touching down on the runway....) Step Two. The EIS determined how many pounds of pollutants (carbon monoxide, hydrocarbons, nitrous oxides, sulfur dioxide, and particulates) each C-5A plane emitted during each take-off-and-landing (which it defined to include idling time, taxiing, take-off run, ascending to 3600 feet, descending from 3600 feet, and touchdown) as well as during each touch-and-go operation. The Air Force made the EIS calculation through use of Table A-4 contained in a document called “Aircraft Engine Emissions Estimator, 1985.” Exhibit 27 in the court record. That table sets forth typical emissions for a C-5A. It shows, for example, that during a typical take-off-and-landing cycle, a C-5A emits 3.0 X 10-2 metric tons of nitrous oxide {i.e., 30 kilograms) and during a touch-and-go the same plane emits 1.3 X 10-2 metric tons {i.e., 13 kilograms). The EIS translates this into pounds (66.2, 28.7) and places the result in EIS Table 4-43. Step Three. The EIS determines “emissions per sortie” by multiplying the “takeoff-and-landing” figure by two (each plane takes off and lands twice during each sortie); by multiplying the “touch-and-go” figure by 18 (each plane “touches-and-goes” 18 times during each sortie); and then adding the results together. It multiplies the “emissions per sortie” by the number of sorties each plane will make per week (4); it multiplies again by the number of weeks per year (52); and it multiplies a final time by the number of C-5A planes (16). The result is total C-5A emissions per year, which (converted into metric tons), says the EIS, in the case of nitrous oxide, amounts to 73.1 tons per year. *465 Step Four. The EIS makes a similar calculation for the C-130s that the C-5As will replace. The C-5As emit more pollutants, but the Air Force will fly them fewer hours. The result is a decrease in all pollutants except for nitrous oxide emissions which will increase from 34.6 tons to 73.1 tons per year. 2. The EIS’s mistakes. Valley Citizens says that the EIS made two important mistakes. First, the EIS did not take account of emissions created during engine testing. Second, the EIS did not take account of pollutants emitted during sorties while the C-5As are flying, but not taking off, landing or engaging in “touch-and-goes.” Such flying obviously would likely take place at quite a distance off the ground, but still at altitudes of less than 3600 feet, a height that the Air Force concedes is significant in respect to pollution. (The record suggests that the EIS also omitted to count pollutants emitted during certain “cargo flights”.) 3. The amount of emissions omitted. Before we can decide whether these mistakes are significant, we must determine how much pollutant is at issue. And, since the parties seem to dispute this matter, we must determine the outside limit of what (the record here indicates) Valley Citizens might prove. a. Engine testing. The Air Force has recalculated likely emissions amounts. Its recalculation (embodied in Exhibit 29) finds that engine testing of C-5As will produce 41.4 tons of nitrous oxides per year; the C-130s produced 4.0 tons; thus “engine testing,” omitted from the EIS, will increase nitrous oxide emissions by an extra 37.4 tons per year. (Exhibit 29, p. 2). Valley Citizens disputes this figure. It claims that “engine testing” C-5As would produce 144.5 extra tons of nitrous oxides per year. (Reply Brief note 3.) Valley Citizens bases its estimate, however, on Air Force documents provided during discovery, contained in Exhibit 37. That document, an Air Force study done before the study that the Air Force used to make its 41.4 ton estimate (Exhibit 29), based its finding on an assumption that is not true (which is why the Air Force rejected it). The amount of “engine testing” emissions obviously depends upon how many times each week the Air Force will test each engine. The early study of Exhibit 37 assumed that the Air Force would test its planes 24 times per week. The earlier study based this assumption upon two sets of phone conversations with Air Force officials, one of which led the earlier study’s authors to believe the Air Force would test its planes more than 30 times per week and the other of which led them to believe the Air Force would test its planes 16 times per week; the authors then simply averaged these two numbers, pointing out in a footnote that their assumption about how many times each week the Air Force would test its engines was somewhat arbitrary. The second, later study, Exhibit 29 (the study the Air Force used to obtain the 41.4 ton estimate) is considerably more detailed; it measures emissions at different engine speeds over sets of different time periods. It bases its assumption about how often the Air Force will test its engines upon the “actual maintenance history" of C-5As “Air Force-wide” and the C-130’s “actual history (at Westover, AFB).” (Emphasis added.) It adds that this “actual history” shows that “engine runups” will be “much less frequent and of briefer, duration than previously assumed.” Having examined the two studies in light of the explanation and the record, which contains nothing that casts doubt on the explanation, we do not see how any reasonable person could accept the results of the first, rather than the second, study as showing what is, in fact, likely to happen at Westover Air Force Base. The record therefore requires us to accept the Air Force estimate that the amount of extra nitrous oxide due to “engine testing” omitted from the EIS amounted to 37.4 tons. b. Cargo flights. The Air Force also modified its estimate of the nitrous oxide emitted during operations to take account of five “non-local cargo missions” per week. For this reason, it increased its estimate from 73.1 to 84.3 tons per year, thereby increasing total nitrous oxide emis *466 sions by 11.2 tons. Valley Citizens does not dispute this amount. c. Sortie emissions. Valley Citizens argues that the EIS estimate of nitrous oxide emitted from the C5-As during sorties underestimated the amount by 321.3 tons. We have searched the record, using the references Valley Citizens provides in its footnotes, to try to understand how it arrived at its estimate and what accounts for the difference. We have found that both the Air Force’s estimate and Valley Citizens’ estimates come from using the same reference, the Aircraft Engine Emissions Estimator, 1985. This Estimator provides two different methods for estimating air pollution. One of those methods (Table 3) uses time periods that different airplanes, with different engines, will fly at different engine modes, and, using different calculation fractions for different pollutants, it determines an amount of pollutant the plane will emit during virtually any kind of flight. The other method (Table A-4) uses what seems to be a finer set of calculations that describes just how much of each pollutant a C-5A will emit during a take-off-and-landing and during a touch- and-go operation. The Estimator makes clear that a user can use either calculation method. Valley Citizens relies upon a set of figures that a Mr. Weister at Robbins Air Force Base developed (for the Air Force) using the first of these methods. It applies Table 3 to an estimate of just how much time during each 5-hour sortie a C-5A will fly (or taxi, or take off, or land, or “touch-and-go”) with its engine set at each of four different possible engine modes. (Exhibit 26.) The Air Force relies upon the second of these methods. It applies Table A-4 to the 2 take-off-and-landing operations and the 18 touch-and-go operations conducted during each 5-hour sortie. (Exhibit 29.) Since the second of these methodologies is not “obviously incorrect,” Commonwealth v. Andrus, 594 F.2d at 886, and since Valley Citizens concedes that it will not argue “that the methodology used by the Air Force to calculate air pollution is wrong” (Reply Brief at 8), one might think that this is the end of the matter, that a court must accept the Air Force’s estimate, which, in essence, is that of the EIS. Valley Citizens, however, goes on to make an argument that the Air Force does not fully refute. It says that the reason these estimates differ is that the first method took account of the entire 5-hour sortie, while the second method (the EIS method) took account only of the time the plane was taking off, landing, or engaging in “touch and go.” It omitted (says Valley Citizens) the time the plane spent circling around (though Valley Citizens does not tell us how much of the 5 hours the planes spent just circling around). The Air Force, of course, might claim that there is no such thing as “circling around time;” it might claim that the two take offs, the two landings, and the 18 touch-and-go operations (accounted for in the second methodology) take up the entire five hours. But, even if the Air Force might say this, it has not said it. And, we must take the facts on a summary judgment record as a reasonable trier of fact might find them. Still, no one could reasonably conclude that the entire 321 tons represents nitrous oxides that the C5-As will emit while they are “circling around.” For one thing, Exhibit 26 itself, which embodies the first methodology and which produces the 321 ton figure, makes clear that the 321 ton figure represents emissions over the entire 5-hour sortie, including nitrous oxide emitted during taking off, landing, and touch- and-go operations, (as well, possibly, as “circling”). Two take offs, two landings and 18 touch-and-go operations simply must consume nearly all, if not all, of a 5-hour sortie and therefore must account for nearly all, if not all, of the 321 tons of nitrous oxide emissions. No one could reasonably conclude that much of the difference between the 321 ton “first method” estimate that Valley Citizens puts forward and the 73.1 ton “second method” estimate that the Air Force uses represents the second method’s omission of circling time. (Certainly, Valley Citizens has not, on this appeal, pointed to anything in the record that would refute the common sense view *467 that 2 take-offs, 2 landings, and 18 touch- and-go operations would consume almost all of a 5-hour flight.) Rather, the differences in the estimates, if they reflect the second method’s failure to count circling must to a far greater degree reflect the use of different methodologies. (The Estimator’s differing numbers in respect to other pollutants confirm the fact that different results reflect the use of different methodologies, and not simply the omission of circling time.) Moreover, one must adjust Valley Citizens’ claim downwards to reflect the fact that the C-130s, too, would have circled around during their flights; if the C-5As emit nitrous oxide during such circling so must the C-130s have done; and the relevant question is how much additional nitrous oxide emissions replacement by C-5As will bring about. The upshot is that, in our view, a reasonable trier of fact might find that the original EIS understated the nitrous oxides the C-5As will emit during flight by an amount that is much less than the 248 ton difference between the first method’s 321 ton estimate and the second method’s 73.1 ton estimate. To put an outer bound on the total “misestimates,” we add the 37.4 ton “engine testing” omission to the 11.2 ton “cargo operation” omission, and then add some small extra number for any “circling around time” omission. Simply to put an outer bound number on this extra amount (which the record does not quantify but which must be small), we shall assume for summary judgment purposes that Valley Citizens could show that the EIS underestimated nitrous oxide emissions by a figure in the range of, say, 50 to 75 tons per year. 4. The significance of the EIS’ omission. We now must determine whether, in context, the EIS’s 50 ton (more or less) nitrous oxide omission is significant. If it is, the Air Force must redo the EIS. Sierra Club v. Marsh, 872 F.2d at 499-501. On the one hand, the error is large in terms of absolute numbers of tons and large in relative terms comparing a revised, to the initial, nitrous oxide emission estimate. On the other hand, undisputed figures in Exhibit 29 indicate that the region’s total yearly emissions of nitrous oxide amount to over 34,000 tons. And, the region is in attainment for nitrous oxide. Further, the C-5A transfer, while increasing nitrous oxide emissions, would, to a small degree, reduce emissions of hydrocarbons as well as emissions of carbon monoxide and particulates, pollutants in respect to which the area is not in attainment. Additionally, the Massachusetts Department of Environmental Quality raised no objection to the fact that a different, but related proposal in the same EIS (a proposal to extend operating hours at the base) would have increased nitrous oxide emissions by 152 tons per year. Also, some of the omitted nitrous oxide emissions would take place when airplanes circle several thousand feet above ground — lower than the EPA’s 3,600 feet ceiling, but still at a considerable distance from the earth where emissions most seriously add to pollution. Finally, and perhaps most important, the entire record makes clear that the basic environmental problem with the transfer of the C-5A concerns noise pollution, not air pollution. Ultimately, though we find the issue closer than the district court suggested, we agree with its conclusion. We would state the matter narrowly by noting that, assuming fact findings appropriately favorable to Valley Citizens, Valley Citizens still has not convinced us that the EIS’s failure to consider engine testing, or other factors that account for differences in nitrous oxide emission estimates is significant. Conservation Law Foundation v. Andrus, 623 F.2d at 719; Commonwealth of Massachusetts v. Andrus, 594 F.2d at 886. V Noise Transfer of the C-5As to Westover undoubtedly means more noise. The EIS analyzes the noise question in detail. It draws map contours showing the places and the number of people that the sorties will expose to noise levels above 65 decibels (a level that the National Academy of Sciences (“NAS”) Guidelines, Exhibit 49, says will “significantly” disturb outdoor speech) and above 75 decibels (a level that the NAS *468 Guidelines says will produce a “very significant disturbance.”) It specifies, for example, that a typical 5-hour sortie operation from runway 05 will expose 25,649 people to noise levels between 65 and 70 decibels, 14,958 to levels between 70 and 75 decibels, 5,933 to levels between 75 and 80 decibels, and 937 to sounds above 80 decibels. It discusses the resulting interference with speech, with sleep, and with land use. It also describes what it calls “annoyance,” using an average, or cumulative noise analysis, in accordance with the National Academy of Sciences Guidelines. It says that, over the course of a year, it expects 3,550 persons to be exposed to an average, or cumulative noise level of more than 65 decibels, and it predicts that, of these, 22 percent will be highly annoyed. Valley Citizens claims that the EIS should not have used the NAS Guidelines to determine the number of people whom the flights would “highly annoy.” It correctly notes that the irritation that a loud noise can cause will depend on (a) how loud it is, (b) how long the noise lasts, and (c) how often it occurs. A very loud noise that occurs once may irritate only briefly; a fairly loud noise that lasts for 50 minutes, 5 days per year may irritate more; a slightly loud noise that recurs every day may prove still more annoying. The relevant questions are: how loud? how long? how often? The NAS Guidelines try to take these matters into account by measuring average noise levels over a 24-hour period, weighing nighttime noise more heavily; and then averaging these averages over the course of a year, producing what it calls a yearly average “day/night noise level.” The Air Force, after calculating that level for the C-5As, checked it against a chart prepared by Professor T.J. Shultz in 1978 (based on studies done at Heathrow Airport in London and elsewhere). The chart shows what percent of those exposed to a particular cumulative yearly average level will likely prove to be “highly annoyed.” This chart yielded the Air Force’s estimate of 771 “highly annoyed” individuals. To show that this NAS Guidelines cumulative approach is unacceptable, Valley Citizens introduced into the record two affidavits by Professor Richard Freyman and affidavits from 1535 citizens of the area. The former stated that the Air Force should not have applied the NAS cumulative average methodology to Westover, in light of the fact that it would only fly sorties approximately every other, rather than every day. The latter affidavits stated that Valley Citizens found the noise (from the C-5A flights that the Air Force had begun to make) extremely bothersome. Professor Freyman’s affidavits (whether or not taken together with the 1535 citizen affidavits) are not sufficient, however, to warrant a legal conclusion that the Air Force used an improper methodology for several reasons. First, Jerry Speakman, an Air Force research physicist, states (by affidavit) without contradiction that the “Air Force’s methodology ... is also the fundamental methodology of all other federal agencies, such as EPA, HUD, and FAA, ... widely endorsed in one form or another since the early 1970s.” See Sierra Club v. United States Department of Transportation, 753 F.2d at 128 (within FAA’s discretion to use cumulative noise impacts to assess environmental effects). Second, Professor Freyman’s affidavits offer a fairly obvious criticism of the standard methodology (that it does not work as well when a loud noise occurs every so often as when it occurs every day, for it “averages out” the occasional noisy day with quiet days, thereby equating the disruption to that caused by a less loud noise that occurs more frequently); but those same affidavits do not offer any alternative measure that seems both practical and superior. Indeed, Professor Freyman says that the Air Force used its methodology “mostly for lack of a better alternative.” Third, NAS methodology, in fact, takes special account of the sudden, occasional loud noise. As Robert Miller, an acoustical expert, points out in his affidavit, average DNL “weights very significantly the loudest days.[in that] the average of two [decibel average] values of, say, 55 and 75 *469 decibels is 72 decibels and not 65 decibels as might be expected.” Fourth, the comments on the draft EIS do not take issue with the standard methodology. It should have been clear from the start that since the methodology is standard, the Air Force would likely use it. Valley Citizens’ comments indicate awareness that the Air Force relied upon Professor Schultz’s work to determine annoyance levels. Thus, the time to complain, and to complain clearly, about methodology was at the comment stage, not two years later after the EIS was complete. Roosevelt Campobello, 684 F.2d at 1047, see pp. 462-63, supra. Fifth, the EIS, taken as a whole, shows a clear awareness that the individual sorties, in and of themselves, create a noise problem leaving aside any matter of yearly averages. The EIS describes the number of sorties, their frequency, and their noise levels. It thereby fully discloses the problem. Its use of the figure 771 tied to the words “highly annoyed,” and to a particularly historical methodology, does not, in context, suggest the EIS believes that the number represents the full extent of the noise problem. Sixth, the law makes clear that the agency has discretion “to determine proper testing methods.” Sierra Club v. United States Department of Transportation, 753 F.2d at 128. See also Suburban O’Hare Commission v. Dole, 787 F.2d 186, 197 (7th Cir.1986) (same), cert. denied, 479 U.S. 847, 107 S.Ct. 169, 93 L.Ed.2d 106 (1986); Webb v. Gorsuch, 699 F.2d 157, 160 (4th Cir.1983) (“When there is conflicting expert opinion, it is for the administrative agency and not the courts to resolve the conflict”). See p. 466, supra. And, it also makes clear that we are normally to judge the reasonableness of the agency’s action in light of the circumstances (normally as revealed in the agency record) as of the time it made the decision. See pp. 459-60, supra. Given these seven sets of circumstances, we cannot say that the Air Force was unreasonable in using the NAS methodology. Although we approve its use here, as did the District of Columbia Circuit in Sierra Club v. United States Department of Transportation, supra, we do not imply that it is immune from criticism or legal attack. But, the place to attack standard methodology, at least in the first instance, is before the agency, not before a reviewing court. Given the commentators’ failure to launch any such attack in their comments, the fact that the methodology is well accepted, and at least a very rough fit between methodology and problem, we find its use in the final EIS reasonable. And, we do not believe the later fact, that 1535 persons signed affidavits saying the noise is very bothersome, can change the reasonableness of the agency’s having previously used the methodology at hand. For these reasons, the judgment of the district court is Affirmed.