Wisconsin Electric Power Company, Petitioner, v. Douglas M. Costle, Administrator, and the United Statesenvironmental Protection Agency, Respondents.wisconsin Electric Power Company, Petitioner, v. Anne M. Gorsuch, Administrator, and the United Statesenvironmental Protection Agency, Respondents, 715 F.2d 323 (7th Cir. 1983)

We have consolidated these appeals to review two decisions of the Environmental Protection Agency ("EPA"). In No. 80-2734 we review the EPA's rule designating portions of the city of Milwaukee, Wisconsin as a nonattainment area for the pollutant sulfur dioxide ("SO sub2 "). In No. 82-1724 we review the EPA's decision to deny Wisconsin Electric Power Company's ("WEPCO") request for a rulemaking to redesignate the Milwaukee area as attainment. Finding the EPA's actions to be consistent with the Clean Air Act and the Administrative Procedure Act ("APA"), we decline to disturb the decisions under review.

Pursuant to the provisions of the Clean Air Act, the Administrator of EPA establishes national ambient air quality standards for pollutants such as SO sub2 . See 42 U.S.C. § 7409. Each state is initially responsible for classifying its geographical areas as satisfying the national ambient air quality standards ("attainment") or not satisfying the standards ("nonattainment"). See id. § 7407(d) (1). The EPA's Administrator may accept or modify the classifications submitted by a state. See id. § 7407(d) (2). If the need arises, the EPA or a state may institute proceedings to revise a classification list. See id. §§ 7407(d) (2), 7407(d) (5).

In 1978, the EPA approved the Wisconsin Department of Natural Resources' ("Wisconsin DNR") classification of Milwaukee County as attainment for SO sub2 . On May 3, 1979, however, the Wisconsin DNR recommended redesignating portions of Milwaukee as nonattainment. This recommendation was accompanied by a technical document describing monitored violations of the primary standard for SO sub2 that occurred in 1978. Based on the Wisconsin DNR's submission, the EPA began a rule-making procedure designed to change Milwaukee's classification. WEPCO objected to the change but on October 5, 1980, the EPA's Administrator signed the final rule designating portions of Milwaukee as nonattainment areas for SO sub2 . See 45 Fed.Reg. 67348 (1980). WEPCO subsequently filed in this Court a timely petition seeking judicial review of the EPA's rule.

While the petition for review was pending in this Court, WEPCO asked the EPA to reconsider its rule and redesignate Milwaukee as an attainment area. WEPCO's request to the EPA was purportedly made pursuant to 42 U.S.C. § 7607(d) (7) (B), which defines the procedural rights of a party seeking reconsideration of certain EPA rules. That section, however, does not apply with respect to rules designating an area as attainment or nonattainment. The EPA therefore properly construed WEPCO's request as a petition made pursuant to § 4(d) of the APA which states that each "agency shall give an interested person the right to petition for the issuance, amendment, or repeal of a rule." 5 U.S.C. § 553(e).

Our review of the EPA's rule classifying Milwaukee as a nonattainment area was postponed, by agreement of the parties, until the EPA acted on WEPCO's petition to repeal the rule and to promulgate a new rule designating the area as attainment. On April 5, 1982, the EPA's Administrator signed the order denying WEPCO's petition to redesignate the Milwaukee area. The EPA's decision not to institute a rulemaking designed to redesignate Milwaukee as an attainment area was based on data submitted by the Wisconsin DNR. The Wisconsin DNR had performed a modeling analysis of the air quality in the Milwaukee area. This analysis incorporated power plant characteristics and emissions data provided by WEPCO, utilized an EPA-approved model, and complied with applicable EPA guidelines. Because the modeling analysis predicted numerous violations of the primary ambient air quality standard for SO sub2, the EPA decided to abide by its 1980 rulemaking and retain Milwaukee's classification as a nonattainment area. See 47 Fed.Reg. 15813 (1982). WEPCO petitioned for judicial review of this decision and, for the purpose of rendering decisions, the petition was consolidated with the petition seeking review of the 1980 rule designating Milwaukee a nonattainment area for SO sub2 .

This procedural history reveals that although these review petitions are connected by a single problem--SO sub2 pollution in Milwaukee--there are two distinct administrative decisions under review. Our review of the decisions is governed by the judicial review section of the APA, see 5 U.S.C. § 706. In each appeal, therefore, we can disturb the EPA's decision only if the agency acted beyond the scope of its statutory authority, or in an arbitrary or capricious manner, or in violation of WEPCO's procedural rights. See id.; Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 415-17, 91 S. Ct. 814, 823-824, 28 L. Ed. 2d 136 (1971). Moreover, we are guided by our decision in U.S. Steel Corp. v. EPA, 605 F.2d 283 (7th Cir. 1979), cert. denied, 444 U.S. 1035, 100 S. Ct. 710, 62 L. Ed. 2d 672 (1980), that EPA rulemakings will be disturbed on procedural grounds only in rare circumstances. See id. 605 F.2d at 290-91.

(No. 80-2734)

Because WEPCO tends to merge the two EPA decisions under review, it is difficult to identify WEPCO's specific objections to the 1980 rulemaking which classified portions of Milwaukee as nonattainment. We have, however, discerned two objections to the rule. First, WEPCO contends that the EPA illegally relied on monitored SO sub2 -concentration data reported by "running averages" as opposed to "block averages." WEPCO's second argument, which takes various forms, is that monitored air quality data made available to the EPA since the rule was promulgated undermine the validity of the nonattainment designation. We hold that both objections are without merit.

The ambient air quality standard applicable in this case is written in terms of maximum average SO sub2 concentrations for 24-hour periods.¹  Monitors that measure SO sub2 concentrations operate continuously. Therefore, average concentrations may be reported for a 24-hour period starting at midnight--"block average"--or any 24-hour period commencing at any time during the day--"running average." The running-averages method is, of course, more likely to detect violations of the air quality standard.

Relying on the decision in PPG Industries, Inc. v. Costle, 659 F.2d 1239 (D.C. Cir. 1981), WEPCO contends that the EPA was required to use block averages, not running averages, in deciding whether there had been monitored violations of the SO sub2 ambient air quality standard. The EPA, however, found that "under either averaging procedure ... the annual 1978 SO sub2 monitoring data indicate that there are 2 or more exceedances of the 24 hour SO sub2 " air quality standard. See 45 Fed.Reg. 67348 (1980). WEPCO does not now challenge this finding; therefore, even if the EPA were precluded from using running averages, the rule under review in this case is supported by data reported using concededly-legitimate block averages.

In any event, WEPCO attaches too much significance to the decision in PPG Industries. The court in that case did not hold that the EPA is precluded from using running averages. On the contrary, the court indicated that the EPA could, consistent with the Clean Air Act, rely on running averages. See 659 F.2d at 1248-49. Because of inadequate notice, however, the court in PPG Industries remanded the case to the EPA to allow for comments on a proposed rule requiring monitored data to be reported in running averages. See id. at 1250-51. There is no procedural infirmity in this case requiring remand; WEPCO was aware of the use of running averages, commented on this reporting method, and the EPA expressed its rational view that running averages should be used "because people breath in SO sub2 from the ambient air on a continuous basis rather than in midnight 'blocks'." 45 Fed.Reg. 67348 (1980). The EPA's reliance on data reported in running averages, therefore, provides no basis for disturbing the rule designating portions of Milwaukee as a nonattainment area.

WEPCO's other attack on the nonattainment designation is that data made available to the EPA since the promulgation of the rule render the rulemaking itself arbitrary or capricious. As noted above, Milwaukee's nonattainment designation was founded on monitored violations of the SO sub2 air quality standard. The violations were measured by the Wisconsin DNR's monitors in 1978 but the nonattainment designation was not finalized by the EPA's Administrator until October 5, 1980. In response to WEPCO's comments, the EPA explained that the period of no monitored violation did not preclude a nonattainment designation because the EPA generally requires eight quarters without a monitored violation before designating an area as attainment. It is WEPCO's position before this Court that eight quarters have now passed without a monitored violation and, therefore, the rule classifying Milwaukee as a nonattainment area must be set aside.

We disagree. WEPCO's briefs in this Court evince a familiarity with the administrative law principle that an agency may not offer to a reviewing court post hoc bases for a prior decision; the decision must stand or fall based on the record before the agency when the decision was announced, see Columbus and Ohio Southern Electric Co. v. Costle, 638 F.2d 910, 912 (6th Cir. 1980). What WEPCO fails to recognize, however, is that the rules of administrative law apply across the board, to agencies and interested parties alike.

We agree with the Ninth Circuit that it is not appropriate "for either party to use post-decision information as a new rationalization either for sustaining or attacking the Agency's decision."²  Association of Pacific Fisheries v. EPA, 615 F.2d 794, 811-12 (9th Cir. 1980) (emphasis added). Three reasons support this conclusion. First, post-decision information "reaches a reviewing court untested by any procedures ... designed to assure its accuracy and completeness." National Association of Demolition Contractors, Inc. v. Costle, 565 F.2d 748, 752 (D.C. Cir. 1977) (quoting Amoco Oil Co. v. EPA, 501 F.2d 722, 729 n. 10 (D.C. Cir. 1974)). Although the monitored data reported after the nonattainment decision was issued may be accurate in some sense, it also may be incomplete (e.g., the monitors may not present a total picture of Milwaukee's air quality). Second, because the world's phenomena and our understanding of those phenomena are always changing, an administrative rule could always be subject to post hoc attacks. Such a besieged rule would never be "final" and the benefits of a stable, consistent administrative policy would be dissipated. Third, when we review a rulemaking our task is to ensure that the agency "articulated a rational connection between the facts found and the choice made." Baltimore Gas and Electric Co. v. Natural Resources Defense Council, Inc., 103 U.S. ----, ----, 103 S. Ct. 2246, 2257, 76 L. Ed. 2d 437 (1983). We do not decide whether in light of the problem addressed, the rule is wise and continues to be wise. In sum, the monitored data made available to the EPA after the nonattainment designation was issued do not undermine the rule, which was rational in light of the record before the agency during the rulemaking. The proper use of the post-decision data was to petition the EPA, as WEPCO did, for a change in the rule designating Milwaukee a nonattainment area. See K. Davis, Administrative Law Treatise 111 (Supp.1982).

It may be WEPCO's position that if a complete and fresh record had been before the EPA at the time of the final rulemaking (Oct. 5, 1980), the EPA would not have designated Milwaukee a nonattainment area. This is so because for the period ending October 1, 1980, there had been no monitored violation of the SO sub2 air quality standard for eight quarters and, with eight "clean" quarters, the EPA generally classifies an area as attainment. The failure to have a complete record at the time of the designation, WEPCO concludes, requires setting the rule aside.

We cannot agree that the EPA's rule was based on such a stale record as to render the rule arbitrary or capricious. As the Supreme Court stated in ICC v. Jersey City, 322 U.S. 503, 64 S. Ct. 1129, 88 L. Ed. 1420 (1944), " [a]dministrative consideration of evidence ... always creates a gap between the time the record is closed and the time the administrative decision is promulgated. This is especially true if the issues are difficult, the evidence intricate, and the consideration of the case deliberate and careful." Id. at 514, 64 S. Ct. at 1134; see Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, 435 U.S. 519, 554-55, 98 S. Ct. 1197, 1217, 55 L. Ed. 2d 460 (1978). Even if monitored air quality data could be collected, compiled, validated, and transmitted instantly to the EPA in a rulemaking proceeding, the record must be closed at some point to allow for public comment and careful deliberation. The time period between closing the record and issuing a decision is lengthened, of course, because the processes of validation and the like cannot be completed in an instant. We will not set aside the EPA's nonattainment designation, therefore, because the EPA did not possess on October 5, 1980, measurements made by Wisconsin DNR's monitors five days previously.

III. THE PETITION TO CHANGE THE NONATTAINMENT DESIGNATION

(No. 82-1724)

After the EPA announced the final rule designating parts of Milwaukee a nonattainment area, WEPCO petitioned the agency to change the classification. WEPCO's petition, alternatively labeled one for reconsideration or redesignation, actually was a petition under 5 U.S.C. § 553(e) for the EPA to institute a rulemaking, governed by the APA's notice and comment provisions,³  designed to repeal the nonattainment designation and redesignate Milwaukee as attainment.

Before the EPA acted on WEPCO's petition, the Wisconsin DNR performed a "RAM-urban modeling analysis" of the air quality in the Milwaukee region. RAM is a dispersion model that predicts the concentration of SO sub2 in the ambient air by taking into account such factors as power plant characteristics, surrounding terrain, and meteorlogical conditions. The RAM study performed in this case predicted numerous violations of the primary air quality standard for SO sub2 . Based on this modeling data, therefore, the EPA denied WEPCO's petition to begin a rulemaking to reclassify the Milwaukee area.

It has not always been clear that an agency decision denying a petition to institute a rulemaking is reviewable under the APA. Nevertheless, we are persuaded, based on the scholarly opinion in WWHT, Inc. v. FCC, 656 F.2d 807 (D.C. Cir. 1981), that the EPA's decision denying WEPCO's petition is reviewable under the standards set forth in the APA section on judicial review, 5 U.S.C. § 706. Our review is guided, however, by the proposition that "only in the rarest and most compelling circumstances" will this Court act "to overturn an agency judgment not to institute rulemaking." WWHT, Inc. v. FCC, 656 F.2d 807, 818 (D.C. Cir. 1981).

The APA does not detail procedures for petitions made pursuant to § 553(e). Therefore the EPA does not violate the APA by not having detailed procedures governing petitions to begin rulemakings. See Laminators Safety Glass Ass'n v. Consumer Product Safety Commission, 578 F.2d 406, 411 n. 15 (D.C. Cir. 1978). Moreover, given the Supreme Court's admonition that reviewing courts must refrain from creating administrative procedures not required by positive law, Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, 435 U.S. 519, 548, 98 S. Ct. 1197, 1214, 55 L. Ed. 2d 460 (1978), WEPCO faces a formidable task in challenging the EPA's decision on procedural grounds.

Because we must decide whether the EPA acted arbitrarily or capriciously in denying WEPCO's petition for a rulemaking, perhaps the agency was required to "respond to the petition" and, in denying the request, "set forth its reasons." Oljato Chapter of the Navajo Tribe v. Train, 515 F.2d 654, 666 (D.C. Cir. 1975). This minimal procedural requirement may be implicit in the APA's structure if our review is to be meaningful. Id. In this case, however, it is undisputed that the EPA expressed its clear rationale of decision (i.e., the RAM analysis predicted violations).

WEPCO would have us create new procedures applicable to an agency's consideration of a § 553(e) petition to begin a rulemaking. WEPCO maintains that before the EPA denied the petition to perform a rulemaking reclassifying Milwaukee, the EPA was required: (1) to give notice that it intended to rely on the Wisconsin DNR modeling analysis, and (2) to make the analysis available for public comment. WEPCO has unwittingly described the notice and comment procedures governing informal rulemakings under the APA, 5 U.S.C. § 553, see Portland Cement Ass'n v. Ruckelshaus, 486 F.2d 375, 393 n. 67 (D.C. Cir. 1973), cert. denied, 417 U.S. 921, 94 S. Ct. 2628, 41 L. Ed. 2d 226 (1974). In effect, therefore, WEPCO is arguing that all petitions for a rulemaking must be granted, at least provisionally, and a rulemaking must be conducted even if ultimately the agency decides not to adopt a new rule.

Apart from violating the command of Vermont Yankee, WEPCO's proposal is fraught with flaws. An agency could be constantly engaged in considering endless § 553(e) petitions, thus depleting resources that could be used to fulfill the agency's mandate. We will not, therefore, create new procedural requirements but will rather espouse the view that "agencies 'should be free to fashion their own rules of procedure and to pursue methods of inquiry capable of permitting them to discharge their multitudinous duties.' " FCC v. Schreiber, 381 U.S. 279, 290, 85 S. Ct. 1459, 1467, 14 L. Ed. 2d 383 (1965) (quoting FCC v. Pottsville Broadcasting Co., 309 U.S. 134, 143, 60 S. Ct. 437, 441, 84 L. Ed. 656 (1940)).

In any event, there was substantial, if not total, compliance with WEPCO's view of the proper procedure. We find incredible WEPCO's assertion that it was unaware that the EPA might rely on Wisconsin DNR's modeling analysis in deciding to retain Milwaukee's nonattainment designation. On March 11, 1981, the Wisconsin DNR sent the EPA a letter (also mailed to WEPCO) stating Wisconsin DNR's objection to WEPCO's request for redesignation. This letter noted that no monitored violation of the SO sub2 air quality standard had been detected in the last two years but stated that the accuracy and completeness of the monitored data "cannot be fully evaluated without a rigorous air quality modeling analysis." On July 27, 1981, the Wisconsin DNR sent the EPA another letter stating that in connection with WEPCO's request for redesignation, Wisconsin DNR was performing a RAM modeling analysis of SO sub2 concentrations. The EPA replied by a letter, also transmitted to WEPCO, which stated that "action on [WEPCO's] redesignation request will be deferred until we have received your SO sub2 attainment analysis." The EPA response was consistent with and even mandated by 42 U.S.C. § 7407(d) (2), which provides that the states must be given an opportunity to demonstrate why a redesignation is inappropriate. The record is replete with other documents demonstrating that WEPCO was aware that the EPA was waiting on the RAM modeling analysis before acting on WEPCO's petition. Indeed, during this "waiting period," the Wisconsin DNR sent WEPCO letters detailing the RAM model, the data used in the analysis, and the preliminary results of the modeling.

WEPCO was thus able to comment on why the EPA should not have relied on the modeling analysis when responding to WEPCO's petition. From the briefs in this Court, WEPCO appears to object to the use of the RAM model in general, not to the model's results in particular. WEPCO was free to cite its objections to the EPA and to argue that the monitored data, not the modeled data, should be used in classifying the Milwaukee area. To the extent that WEPCO takes issue with particular aspects of the modeling analysis that were not known (or knowable) to WEPCO before the EPA's decision, we perceive no reason why WEPCO cannot file another § 553(e) petition for a rulemaking to reclassify the Milwaukee area.

The Clean Air Act explicitly authorizes the EPA to classify an area as nonattainment based on "monitored data" or "air quality modeling." 42 U.S.C. § 7501(2). The EPA's choice to rely on an air quality model is a policy judgment deserving great deference. See Batavia v. FERC, 672 F.2d 64, 82 (D.C. Cir. 1982); see also Baltimore Gas and Electric Co. v. Natural Resources Defense Council, Inc., --- U.S. ----, ----, 103 S. Ct. 2246, 2255-2256, 76 L. Ed. 2d 437 (1983) (reviewing court must be at its most deferential when reviewing predictions made by an expert agency). Moreover, the model used in this case, a multi-source model for SO sub2 is an EPA-approved model that has been upheld by the Sixth Circuit as a legitimate tool for making attainment/nonattainment determinations. See Cleveland Electric Illuminating Co. v. EPA, 572 F.2d 1150, 1164 (6th Cir. 1978), cert. denied, 439 U.S. 910, 99 S. Ct. 278, 58 L. Ed. 2d 255 (1978). In the face of these circumstances WEPCO does not contend that reliance on a RAM model is itself arbitrary or capricious. Rather it is WEPCO's view that in the context of this case, the EPA was required to eschew modeled results and redesignate Milwaukee as an attainment area for SO sub2 .

WEPCO directs our attention to a statement made in 1978 by an EPA official that " [i]f there is a conflict between adequate monitoring data and modeling results, monitored values should be used." WEPCO also points to a 1979 memorandum written by the EPA's Director of Control Programs Development that states: "Generally, eight quarters of ambient air quality data are required showing no violation before an attainment designation can be approved." Based on these documents and the fact that for two years there had been no monitored violation of the SO sub2 standard, WEPCO contends that the EPA acted arbitrarily and capriciously in not redesignating Milwaukee an attainment area.

A review of the EPA's documents, however, reveals that WEPCO has overstated its case and mischaracterized EPA policy. Under EPA guidelines there is no requirement that an area be designated attainment after eight quarters of no monitored violations. The memorandum on which WEPCO relies in arguing that there is such a requirement states that the "eight quarters" policy "is merely a simple extension of current Agency policy concerning the selection of control strategy design values where recent air quality shows improvement over previous concentrations and real, not paper, emission reductions have occurred in the nonattainment area." (Emphasis added). By "real" reductions the memorandum is referring to emission reductions that are the "result of legally enforceable actions." Because there have been no enforceable emission limits applicable to SO sub2 sources in the Milwaukee area, the EPA, consistent with its guidelines, chose not to redesignate the area based on monitored data.

Moreover, we agree with the Sixth Circuit that the EPA "is not required ... by its own policies to prefer monitoring to modeling in making attainment status designations." PPG Industries, Inc. v. Costle, 630 F.2d 462, 467 (6th Cir. 1980) (footnote omitted). One EPA document provides that states should use "either modeling or monitoring" data in considering whether to redesignate an area. See Procedures for Handling Future Section 107 Redesignations (Oct. 10, 1978). Another EPA document states that modeling "will be extremely supportive" of a nonattainment designation for SO sub2 . See State Guidance Geographic Designation of Attainment Status (1977). And, in designating a portion of Ohio nonattainment, the EPA noted that the "Agency has determined that where air quality modeling results are available, such results will be used to determine the designation, taking precedence over air quality monitoring data which is usually not sufficiently comprehensive to cover any given area." 43 Fed.Reg. 45998 (1978).

Finally, the EPA's decision in this case is consistent with the statement made by the EPA official in 1978: " [i]f there is a conflict between adequate monitoring data and modeling results, monitored values should be used." According to this statement, monitored data are to be preferred only if two conditions are met--the monitored data are adequate and the data conflict with modeled results. The EPA's decision reveals that neither condition was satisfied.

In denying WEPCO's petition, the EPA noted that there was reason to believe that the monitored data was inadequate because the Wisconsin DNR's model demonstrated that the monitors did not accurately reflect air quality near WEPCO's Valley power plant. 47 Fed.Reg. 15813 (1982). More important, the modeled and monitored data do not conflict. The EPA has interpreted the Clean Air Act as authorizing a nonattainment designation based on predicted (future) violations of air quality standards. This interpretation, which WEPCO does not challenge here, has been upheld as reasonable and valid. See Columbus and Southern Ohio Electric Co. v. Costle, 638 F.2d 910, 912 (6th Cir. 1980); PPG Industries, Inc. v. Costle, 630 F.2d 462, 464-65 (6th Cir. 1980); see generally 42 U.S.C. § 7407(d) (1) (B). A modeling analysis, such as the one in this case, is the method that the EPA employs to predict violations; monitors, on the other hand, merely record historical concentrations at discrete locations. Furthermore, the RAM model is based on hypothetical facts concerning power plant operations and weather conditions that are legitimate, see Republic Steel Corp. v. Costle, 621 F.2d 797, 805 (6th Cir. 1980), yet perhaps different from the historical facts, Cleveland Electric Illuminating Co. v. EPA, 572 F.2d 1150, 1164 (6th Cir. 1978), cert. denied, 439 U.S. 910, 99 S. Ct. 278, 58 L. Ed. 2d 255 (1978). Because models and monitors describe the world at different times, and perhaps under different circumstances, the modeled and monitored data in this case do not conflict.⁴ 

For the reasons expressed in this opinion, the petitions for review are denied.