Petitioner Indianapolis Power & Light Company (IPL) challenges regulations promulgated by the Environmental Protection Agency (EPA) pursuant to the acid rain program begun under 42 U.S.C. Secs. 7651-7651o (Clean Air Act). IPL argues that the regulations relating to the calculation and allocation of "extension allowances" should allow for the adjustment of a utility unit's 1988-1989 emissions data if the unit was out of operation for an extended period during those two years. Because the Clean Air Act is silent on the issue of adjusting a utility unit's 1988-1989 emissions data and the EPA's interpretation is permissible, we deny IPL's petition for review.

Title IV of the Clean Air Act establishes an acid rain program to reduce emissions of sulfur dioxide and nitrogen oxides, the primary precursors of acid rain. See 42 U.S.C. Secs. 7651-7651o. Title IV imposes a national cap of 8.90 million tons of sulfur dioxide emissions per year on electric utilities. 42 U.S.C. Secs. 7651b(a)(1), 7651d(a)(3). The emissions reductions are to occur in two phases. In Phase I, at issue here, 110 utilities with the largest coal-fired utility electric generating units must reduce their emissions to 2.50 pounds of sulfur dioxide per million British thermal units (Btus) of fuel consumed by each power-generating unit in the "baseline" years of 1985, 1986 and 1987. 42 U.S.C. Sec. 7651c.1 Each utility unit is allocated a number of fully marketable pollution allowances so that it may emit sulfur dioxide at this level. 42 U.S.C. Sec. 7651c. Each allowance authorizes the emission of one ton of sulfur dioxide during one calendar year and may be bought, sold, traded or banked for future use or resale. 42 U.S.C. Secs. 7651a(3), 7651b(b).

Beginning in 1995, the emissions from each Phase I unit may not exceed the number of allowances that unit holds. 42 U.S.C. Sec. 7651c(a). Emissions from a Phase I unit that exceed the number of allowances allocated to it are unlawful and are subject to various fines and penalties. 42 U.S.C. Secs. 7651b(g), 7651j. To comply with Title IV requirements and ensure that its emissions do not exceed its allowances, a utility has three options. A utility can switch from coal with a high sulfur content to low-sulfur coal, purchase allowances from other utilities or install costly sulfur dioxide control technology known as "scrubbers." Congress encouraged the installation of scrubbers by establishing an extension allowance program. 42 U.S.C. Sec. 7651c(d).

Under the program, utilities that install and operate scrubbers on their Phase I units to reduce sulfur dioxide emissions qualify for "extension allowances" in addition to the allowances initially allocated. 42 U.S.C. Sec. 7651c. The EPA is authorized to "approve an extension proposal in whole or in part, and with such modifications or conditions as may be necessary, consistent with the orderly functioning of the allowance system...." 42 U.S.C. Sec. 7651c(d)(3). Utilities whose extension proposals are approved are not exempt from the Title IV prohibition on emissions in excess of allowances. Instead, the EPA raises their emissions limitation and grants a corresponding number of extension allowances for a two-year period according to the following formula:

[T]he difference between the lesser of the average annual emissions in calendar years 1988 and 1989 or the projected emissions tonnage for calendar year 1995 [and 1996] of each eligible phase I extension unit ..., and the product of the unit's baseline multiplied by an emission rate of 2.50 lbs/mmBtu, divided by 2,000.

42 U.S.C. Sec. 7651c(d)(4)(A); 40 C.F.R. Sec. 72.42(c). That is, the EPA bases the number of extension allowances on the difference between the maximum emissions level set by Congress (i.e., 2.50 pounds of sulfur dioxide per million Btus of fuel consumed by the unit in the baseline years of 1985-1987) and the lesser of the unit's expected emissions in 1995 and 1996 or its actual emissions in 1988 and 1989.

Seventeen utilities, including IPL, decided to install scrubbers and applied for extension allowances. According to the formula, the utilities were entitled to 4.0-4.5 million extension allowances. The Clean Air Act, however, capped the number of available extension allowances at 3.5 million. 42 U.S.C. Sec. 7651c(a)(2). The EPA therefore held a lottery in March 1993 to determine the selection priority for granting extension allowances to the applicants. The seventeen utilities had prepared for this eventuality, however, by entering an extension allowance pooling agreement in March 1992 (Agreement). Under the Agreement, the utilities agreed to pool the extension allowances and to re-allocate them, based in part on the formula used by the EPA, so that each applicant, regardless whether it was selected in the lottery, received a percentage of the number of allowances it would have received in the lottery had it been selected. Although IPL was not selected in the lottery it nonetheless received extension allowances under the Agreement.

IPL's Petersburg, Indiana Unit # 2, a Phase I unit, was out of operation from October 1, 1988, through April 6, 1989, for unexpected major repairs. As a result, the unit's sulfur dioxide emissions during the two-year period were lower than usual.2 According to the EPA's calculations using actual emissions data for 1988 and 1989, Petersburg Unit # 2 is entitled to approximately 14,000 extension allowances. According to IPL's calculations, however, if the emissions data were normalized or annualized to reflect what emissions would have been in 1988 and 1989 but for the outage, Petersburg Unit # 2 would be entitled to at least 20,000 additional extension allowances. Consequently, the number of extension allowances to which IPL is entitled under the EPA formula, and thus under the Agreement, varies dramatically according to whether actual or normalized emissions data for 1988 and 1989 are used.

The EPA considered the issue of adjustments to emissions data in its proposed rulemaking. 56 Fed.Reg. 63,002 (1991) (Joint Appendix (JA) 2). In December 1991, in requesting comment "on the appropriateness of allowing an adjustment in the limited situations where a forced outage occurred," the EPA explained that allowing adjustments was not warranted for several reasons:

First, the Phase I Extension is voluntary. If a utility finds that it is not favorable to apply for a Phase I Extension, it can chose [sic] another compliance option. Second, the information is historical, actual data reported by the utility itself and the statute does not provide for adjustment.... Third, the Agency is attempting to respond to industry's need for certainty at an early date regarding allocation of allowances from the Phase I Extension reserve. Allowing an opportunity for adjustments to be made to essential historical data would delay certainty.

56 Fed.Reg. 63,017 (JA 9). The EPA did not allow for adjustments in the final rules issued on January 11, 1993. 58 Fed.Reg. 3,590, 3,669 (1993) (JA 57, 63); 40 C.F.R. Sec. 72.42(c)(3) (JA 63).3 In the final acid rain permit issued to IPL, the EPA affirmed that it intended to allocate allowances without adjusting the emissions data to account for the fact that Petersburg Unit # 2 was out of operation for six months in 1988 and 1989. IPL petitions the court to vacate the regulations which do not provide for adjustments and to remand the matter to the EPA with instructions to adjust the emissions data for Petersburg Unit # 2.

IPL makes three principal arguments. First, it claims that the Clean Air Act obliges the EPA to adjust a utility unit's 1988-1989 emissions data to account for an unexpected prolonged outage. Second, IPL argues that even if the statute is silent on the issue, the EPA mistakenly interpreted the statute as prohibiting it from adjusting the 1988-1989 emissions data. Finally, IPL maintains that the EPA interpretation is unreasonable. We reject each argument. We may set aside the regulations promulgated pursuant to the acid rain program only if they are:

(A) arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law;

(B) contrary to constitutional right, power, privilege, or immunity;

(C) in excess of statutory jurisdiction, authority, or limitations, or short of statutory right; or

(D) without observance of procedure required by law....

42 U.S.C. Secs. 7607(d)(9), 7607(d)(1)(F). The EPA's interpretation of the Clean Air Act must be reviewed under Chevron, U.S.A., Inc. v. Natural Resources Defense Council, 467 U.S. 837, 104 S. Ct. 2778, 81 L. Ed. 2d 694 (1984) (Chevron ). Applying these standards, we uphold the EPA regulations.

The Clean Air Act provides that extension allowances are to be allocated in part on the "average annual emissions in calendar years 1988 and 1989." 42 U.S.C. Sec. 7651c(d)(4)(A). It does not define "average annual emissions" nor does it explicitly authorize or prohibit adjustment of a utility unit's 1988-1989 emissions data for any reason, including an unexpected prolonged outage. Accordingly, the language cannot be read, as IPL argues, to compel EPA to adjust emissions data.

We conclude that the statutory language reasonably supports only two readings. First, the language can be construed to require that the EPA use actual emissions data. Under this reading, the EPA may not exercise any discretion but must mechanically calculate a utility unit's average annual emissions by dividing the total emissions during calendar years 1988 and 1989 by two.4 But the EPA did not read the language this way.5 Rather, the EPA concluded that the statute was silent regarding whether emissions data could be adjusted, thus allowing it to exercise its discretion, apply its expertise and adopt and implement a reasonable interpretation. See Ohio v. EPA, 997 F.2d 1520, 1527 (D.C.Cir.1993) (where agency regulation "represents a reasonable and permissible construction of the statute," court will defer under Chevron ). We agree that the statute is silent on the issue and uphold as a permissible construction of the statute the EPA's decision not to adjust emissions data.

The EPA articulated a reasoned basis for not adjusting a utility unit's emissions data. It sought to explain to the affected utilities in unequivocal terms how it proposed to allocate extension allowances. By relying solely on the historical emissions data previously reported to the Department of Energy,6 the EPA made it possible for each affected utility to calculate how many extension allowances could be claimed by other utilities and how many extension allowances it could claim and would likely obtain for itself. Each affected utility could then compute the costs and the benefits of installing a scrubber as opposed to other means of complying with the Title IV emissions reduction requirements. By contrast, if the EPA were to permit adjustments to the reported emissions data in special circumstances, utilities, lacking information on how the EPA would define the special circumstances and whether other utilities' units would meet the specified criteria for adjustment, would be less able to determine whether participation in the extension allowance program would be beneficial. It was reasonable, then, for the EPA to conclude that allowing adjustments to emissions data could significantly reduce the incentives to participate in the voluntary extension allowance program.

We reject IPL's argument that section 404(d)(3) of the Clean Air Act indicates that Congress intended the EPA to base its extension allowance calculations on emissions data adjusted to reflect "normal" operating circumstances. Section 404(d)(3) does authorize the EPA to "approve an extension proposal in whole or in part, and with such modifications or conditions as may be necessary, consistent with the orderly functioning of the allowance system, and to ensure the emissions reductions contemplated by the subchapter." 42 U.S.C. Sec. 7651c(d)(3). The provision, however, notably does not refer to outages, emissions or adjustments because it relates only to whether a utility qualifies for the extension allowance program, not to the allocation of extension allowances to utilities that qualify for and participate in the program. It authorizes the EPA to review and amend the terms of a utility's proposal to ensure that the utility in fact installs scrubbers and makes the structural changes which qualify it for the program and which extension allowances are designed to reward.

Congress did not need to specify "average annual actual emissions" for the court to uphold EPA's decision not to adjust emissions data. If Congress had intended the EPA to use only emissions data that took into account the time a unit was not in operation in 1988 or 1989, it could have expressly mandated that extension allowances be calculated according to adjusted data as it did in determining fuel consumption during the baseline years under 42 U.S.C. Sec. 7651c. But Congress remained silent on the issue. We therefore uphold under Chevron the EPA's reasoned decision not to adjust emissions data. Accordingly, the petition for review is

Denied.

I write separately not because I disagree with the majority's conclusion that the EPA's interpretation of 42 U.S.C. Sec. 7651c(d)(4)(A) is a reasonable one, but rather because I agree with the majority's conclusion more firmly than the majority itself does. That is, I would hold that the EPA's interpretation is not merely a reasonable construction, but the only reasonable construction. I do not agree that Congress was silent on the question of whether the EPA should "mechanically calculate a utility unit's average annual emission by dividing the total emission during calendar years 1988 and 1989 by two." Maj. op. at 86. Rather, I do not see how Congress could have said any more explicitly than it did that this is precisely what it was mandating. We have observed in the past that "some will find ambiguity even in a 'No Smoking' sign...." International Union, United Auto. Aerospace & Agric. Implement Workers of America v. General Dynamics Land Sys. Div., 815 F.2d 1570, 1575 (D.C.Cir.1987). I see little more ambiguity in the present statute on the question before the EPA than in such a sign. I would therefore reach the same conclusion as the majority, but I would reach it under the first step of the Chevron analysis without proceeding to the second.