California and Nevada each petition for review of the Environmental Protection Agency’s approval of the other’s state implementation plan (SIP) for attaining air quality in the Lake Tahoe Basin under the Clean Air Act. 42 U.S.C. §§ 7401-7642 (1982). The League to Save Lake Tahoe also challenges the Nevada Plan. 1

The bone of contention which precipitated this feud is the proposed construction of two casino parking garages in the town of South Lake Tahoe. If completed, they would increase the legal number of parking spaces available in the town by twenty to twenty-five percent. Such structures can be indirect sources of carbon monoxide, one of the pollutants for which the Clean Air Act sets attainment standards.

In 1977 the Lake Tahoe Basin was designated a nonattainment area for carbon monoxide. A nonattainment area is an area that has been shown to exceed the national ambient air quality standard for a particular pollutant. 42 U.S.C. § 7501 (2). Once an area has been designated a nonattainment area for carbon monoxide, the state is required to promulgate a plan to achieve attainment of the national ambient air quality standard by December 31, 1982. 42 U.S.C. § 7502 (a)(1), (2). If attainment by that date is impossible despite the implementation of all reasonably available measures, attainment must be secured as expeditiously as practicable but not later than December 31, 1987. 42 U.S.C. § 7502 (a)(2).

The Clean Air Act at one time required indirect source review (ISR) as part of a state implementation plan. In 1977, however, Congress allowed states to amend their SIPs and to repeal the requirement for indirect source review so long as the plan provided for sufficient air quality control. 42 U.S.C. § 7410 (a)(5)(A)(iii); see also Manchester Environmental Coalition v. *1440 EPA, 612 F.2d 56, 58 (2d Cir.1979). Since final approval of the Nevada SIP means that indirect source review will no longer be required, repeal of Nevada’s ISR has been stayed pending resolution of this litigation. The practical effect has been that construction of the garages has been halted since 1979. See California Tahoe Regional Planning Agency v. Sahara Tahoe Corp., 504 F.Supp. 753, 767-69 (D.Nev.1980) (effectively enjoining construction of garages until EPA approves Nevada’s SIP).

The parties raise various objections to the EPA’s actions and we deal with each in turn, bearing in mind that we may overturn an action of the EPA only if it is arbitrary or capricious. See 42 U.S.C. § 7607 (d)(9)(A). We may not substitute our judgment for that of the EPA, see Columbia Basin Land Protection Association v. Schlesinger, 643 F.2d 585, 592 (9th Cir.1981), but we must satisfy ourselves that the EPA examined the relevant data and articulated a satisfactory explanation for its action, in light of the policy behind the Clean Air Act. Motor Vehicle Manufacturers Association v. State Farm Mutual Automobile Insurance Co., 463 U.S. 29, 43 , 103 S.Ct. 2856, 2866 , 77 L.Ed.2d 443 (1983); see also Sierra Club v. Costle, 657 F.2d 298, 322-23 (D.C.Cir.1981). We conclude that the EPA’s approval of both the Nevada and California Plans must be upheld.

I. The Challenges to Approval of Nevada’s State Implementation Plan.

A. Repeal of indirect source review.

California tells us that its “chief concern” is that to approve Nevada’s Plan is to approve the repeal of indirect source review. It correctly points out that repeal is permissible only if the state implementation plan, evaluated without indirect source review, is sufficient to attain and maintain national ambient air quality standards as required by section 7410. Manchester Environmental Coalition v. EPA, 612 F.2d at 59-60 . California argues that the EPA failed adequately to take into account the future absence of indirect source review when it evaluated the sufficiency of the Plan. Specifically, California criticizes the failure of EPA in approving the Plan to take into account the effects of the two South Lake Tahoe parking garages, which California asserts would likely have serious air quality consequences.

California’s premise is that before determining that a Plan will adequately attain and maintain air quality standards in the absence of indirect source review, the state in its Plan, and the EPA in approving that Plan, should take into account known sources of pollution. With that premise we have no quarrel. We cannot, however, accept the conclusion that the failure of Nevada and EPA to consider the effect of the garages on the environment constitutes a failure to consider a source of pollution that could significantly affect attainment and maintenance of air quality standards. The record does not bear out California’s assertion that construction of the garages would have adverse effects on air quality.

California cites to four items in the record to support its position. They are letters to EPA officials stating that garage construction will have an adverse effect on air quality. Two are from California state officials and a third is from a lawyer for petitioner League to Save Lake Tahoe. Two of these letters refer to modeling which purportedly shows that there will be a doubling of traffic. The letters’ authors do not furnish that modeling analysis, however; nor has California in its briefs provided any indication of where, in the record, such supporting material may be found. We have independently conducted a thorough search of the record in vain pursuit of any modeling or other data that appears to support the proposition that the garages will have the effect of jeopardizing attainment or maintenance of air quality standards. We therefore are constrained to hold that the EPA was not arbitrary or capricious in approving a Plan which did not treat the garages as a significant potential source of carbon monoxide.

*1441 California additionally argues that even if the record contains no modeling analysis supporting the assertions in the letters, the agency failed in its duty to respond to “significant comments” in the proceeding. 42 U.S.C. § 7607 (d)(6)(B) requires the Administrator to respond to all “significant comments” made during the comment period. The statute does not require the agency to respond to every comment, but only to those which can reasonably be deemed significant. The question is whether the passages in the letters should be considered “significant.”

The District of Columbia Circuit has held that where the only discussion of a problem in the record was a two-page discussion in the comment of one of the petitioners, which did not directly reject the administrator’s decision in question, there was no “significant comment” requiring response. The court held that the “unsupported claim simply did not rise to the level of a comment which required a response from the Administrator.” Lead Industries Association, Inc. v. EPA, 647 F.2d 1130, 1167 (D.C.Cir.), cert. denied, 449 U.S. 1042 , 101 S.Ct. 621 , 66 L.Ed.2d 503 (1980). Similarly, the failure of the EPA in this case to respond to the four references to the garages, three of which were written by officials of the two petitioners in this case, and which are scattered in a record which fills four 10 inch by 12 inch by 15 inch boxes does not require us to overturn the final rulemaking.

B. Acceptance of Nevada’s modeling analysis demonstrating attainment.

Both California and Nevada prepared modeling analyses projecting future carbon monoxide pollution in the Nevada side of the Tahoe Basin. Each state used different assumptions in its analysis regarding atmospheric condition, peak traffic, traffic speed, cold start factor and temperature. See 49 . Fed.Reg. 6898-99 (Feb. 24, 1984). California’s assumptions were more conservative than Nevada’s and therefore resulted in predictions of higher levels of carbon monoxide. Under California’s modeling analysis, Nevada would not attain the required standard, while Nevada’s analysis predicted that the standard would be met. California asserts that in determining that Nevada had demonstrated attainment, EPA failed to explain adequately the inconsistency between the two analyses and to articulate the reasons why it accepted Nevada’s.

The record shows, however, that the EPA did explain the difference. It pointed out that the discrepancy was due to the use by California of more conservative estimates, and further explained in its technical support document for the Nevada SIP why it considered Nevada’s assumptions adequate. In addition, the record also reflects that Nevada’s model had been tested and modified by monitoring during a period when conditions were likely to be at their worst. Therefore, EPA’s approval of Nevada’s modeling analysis cannot be said to have been arbitrary or capricious.

C. Maintenance of the carbon monoxide standard in Nevada’s implementation plan.

California argues that the EPA gave inadequate consideration to the requirement that a Plan provide for maintenance of air quality standards once they are attained. See 42 U.S.C. § 7410 (a)(1), (2)(B); 40 C.F.R. § 52.22 (1984). The EPA’s final rulemaking notice squarely responded to a comment by the League to Save Lake Tahoe about maintenance:

49 Fed.Reg. 6899 (col. 1) (Feb. 24, 1984).

The underlying record shows that the modeling done by Nevada included a five *1442 percent compounded growth rate to account for new sources of pollution, and demonstrated that air quality standards would be maintained through 1988 with such growth. The statute requires the Administrator to approve an SIP if it includes emissions limitations, schedules and timetables for compliance with such limitations and other measures as may be necessary to ensure attainment and maintenance of the standards. Nevada’s analysis, based on a projected five percent compound growth rate, supports the agency’s determination that the statutory requirements have been met.

California points to the EPA’s reference to additional control measures and argues that such measures should have been clearly identified. Since the EPA determined that the Plan was sufficient even without such measures, California’s argument lacks merit.

California also argues that the EPA should have taken into account the effect of the parking garages on air quality in assessing maintenance. For the reasons discussed above in Part A, the argument fails here as well.

II. Nevada’s Challenge to EPA Approval of California’s Implementation Plan.

Nevada challenges the EPA’s approval of California’s SIP on four grounds: (1) California’s modeling analysis is inadequate; (2) California’s SIP does not achieve expeditious attainment of national ambient air quality standards in violation of section 7502(a)(2); (3) it will not achieve national ambient air quality standards by December 31, 1987 because the modeling shows that one site will not be within the limits for carbon monoxide attainment; and (4) California’s SIP has an unenforceable vehicle emission control inspection and maintenance program (I/M).

Nevada first challenges California’s modeling analysis by asserting that there was no factual support for the inputs used. Nevada does not specify which inputs are inaccurate or point to anything in the record indicating their inaccuracy. Since Nevada maintains that its own analysis is adequate, and California’s analysis is based upon more conservative inputs than is Nevada’s, Nevada cannot plausibly maintain that California’s SIP is inadequate on this ground. We agree with the Sixth Circuit that the EPA may approve a Plan based on an analysis more conservative than the Act requires. See Cleveland Electric Illuminating Co. v. EPA, 572 F.2d 1150 , 1163-64 (6th Cir.), cert. denied, 439 U.S. 910 , 99 S.Ct. 278 , 58 L.Ed.2d 256 (1978).

Nevada next contends that California’s SIP does not provide for attainment as expeditiously as possible because it does not include some of the measures adopted by the California Air Resources Board and which were included in an earlier implementation plan. Nevada does not explain how adoption of the other strategies would result in earlier attainment. We see no basis for holding, as a matter of law, that a state which has once adopted certain measures thereby commits itself to their inclusion in future plans. Such a result would discourage experimentation with pollution control methods. The statute requires attainment “as expeditiously as practicable,” a standard which permits some flexibility. See, e.g., Train v. Natural Resources Defenses Council, Inc., 421 U.S. 60, 86-87 , 95 S.Ct. 1470, 1485 , 43 L.Ed.2d 731 (1975); National Steel Corp., Great Lakes Steel Division v. Gorsuch, 700 F.2d 314, 322 (6th Cir.1983); Florida Power and Light Co. v. Costle, 650 F.2d 579, 587 (5th Cir.1981); Northern Ohio Lung Association v. EPA, 572 F.2d 1143, 1149 (6th Cir.1973).

Nevada’s next argument is that the California Plan cannot achieve attainment by 1987 because its modeling analysis shows that one site, called the Travel Lodge site, will not be in compliance by that time. The modeling shows that compliance would be lacking by only a very small margin. 2

*1443 Given the conservative nature of California’s modeling inputs, we cannot hold that its prediction of minimal noneompliance at one site in the state brings down the entire Plan. See Alabama Power Co. v. Costle, 636 F.2d 323, 360 (D.C.Cir.1979) (holding that a violation of attainment standards may be considered “de minimis”). In this case it is not even clear that any violation will occur.

Nevada’s final challenges to the California Plan relate to vehicle emission controls. The Clean Air Act requires that an implementation program “establish a specific schedule for implementation of a vehicle emission control ... program,” 42 U.S.C. § 7502 (b)(ll)(B). It also provides that Plans, like California’s, which are required to be submitted before July 1,1982, contain “enforceable measures to assure attainment of the applicable standard not later than December 31, 1987.” 42 U.S.C. § 7502 (c).

The State of Nevada argues that the California Plan has no schedule for implementation of a vehicle emission control program and contains no enforceable measures to assure attainment. This is incorrect. The State of California has committed itself to implementation of an I/M Program both in its revised Plan and as supplemented in a letter by the Tahoe Regional Planning Agency, the organization charged with implementing such a Plan.

Nevada next argues that, even though California has made an enforceable commitment to implement the program, the statute requires that the program be in place by July 1, 1982, the deadline for submission of the Plan. The statute, however, requires that by July 1,1982 the Plan must be submitted, and must contain enforceable measures. It does not require that the measures be in place at that time. See 42 U.S.C. § 7502 (c). The deadline for achieving attainment is December 31,1987. States are thus given five years to put their full program into place.

The decision of the Fifth Circuit in City of Seabrook v. EPA, 659 F.2d 1349, 1364 (5th Cir.1981), cert. denied, 459 U.S. 822 , 103 S.Ct. 51 , 74 L.Ed.2d 57 (1982), the only authority upon which Nevada relies, is not to the contrary. The court there held that the “schedule” requirement in section 7502(b)(11)(B) did not require submission of an enforceable program when original SIP’s were to be submitted in 1979. The court held that the terms of an enforceable program were to be submitted in 1982 pursuant to 42 U.S.C. § 7502 (c). The court thus held that submission of a Plan containing enforceable measures was not required until 1982. Neither the statute, its history, nor the Fifth Circuit’s opinion in City of Seabrook supports the proposition that an entire maintenance program had to be put in place between 1979, when the statute required a schedule, and 1982, when the statute required enforceable measures to be spelled out.

The petitions for review are denied.

. Petitions to review interim approval by the EPA have also been consolidated in this case, but are rendered moot by our decision on the final approval. Those petitions, Nos. 82-7410, 82-7427, 82-7473, and 82-7474, are therefore dismissed.

. In 1981 the carbon monoxide concentration at the site was 28.4 parts per million. Full compli-anee requires this to be reduced to 9.0. The *1443 California modeling analysis shows that the figure would be reduced to 11.1 by 1987.