United States Court of Appeals
                                                                      Fifth Circuit
                                                                     F I L E D
                                                                     October 28, 2003
                     UNITED STATES COURT OF APPEALS
                          FOR THE FIFTH CIRCUIT
                                                                Charles R. Fulbruge III
                     _______________________________                    Clerk

                               No. 02-60017
                     _______________________________


         BCCA APPEAL GROUP; NATIONAL RESOURCES DEFENSE COUNCIL;
         ENVIRONMENTAL DEFENSE; DAVID MARRACK; JANE W. ELIOSEFF;
       GALVESTON-HOUSTON ASSOCIATION FOR SMOG PREVENTION (GHASP);
     GALVESTON BAY CONSERVATION AND PRESERVATION ASSOCIATION (GBCPA);
                SIERRA CLUB; AND BRAZORIA COUNTY, TEXAS,

                                                                      Petitioners,

                                    v.

             UNITED STATES ENVIRONMENTAL PROTECTION AGENCY;
                  CHRISTINE T. WHITMAN, ADMINISTRATOR; AND
GREGG A. COOKE, ADMINISTRATOR UNITED STATE ENVIRONMENTAL PROTECTION AGENCY
                                  REGION 6,

                                                                     Respondents,

                                    v.

                  STATE OF TEXAS; HARRIS COUNTY, TEXAS;
               BCCA APPEAL GROUP; ENVIRONMENTAL DEFENSE;
               DAVID MARRACK; JANE W. ELIOSEFF; GALVESTON-
                HOUSTON ASSOCIATION FOR SMOG PREVENTION;
              GALVESTON BAY CONSERVATION AND PRESERVATION
              ASSOCIATION; HOUSTON-GALVESTON AREA COUNCIL;
                    CITY OF HOUSTON; AND SIERRA CLUB,

                                                                      Intervenors.

          _____________________________________________________

                      On Petitions for Review of Final Rule
              of the United States Environmental Protection Agency
         ______________________________________________________


                                    1
Before DAVIS and BENAVIDES, Circuit Judges, and RESTANI,* Judge

DAVIS, Circuit Judge and RESTANI, Judge:

       Before the court are consolidated petitions for direct review of a decision of the United

States Environmental Protection Agency (“EPA”) approving (1) the State of Texas’s one-hour

ozone attainment demonstration state implementation plan (“SIP”) for the Houston-Galveston

severe ozone nonattainment area, and (2) the state’s control strategy (collectively referred to as

the “Houston SIP”). See Approval and Promulgation of Implementation Plans, 66 Fed. Reg.

57,160 (EPA Nov. 14, 2001) (final rule). Petitioners consist primarily of industries subject to the

emissions controls in the Houston SIP, a county government affected by some of the SIP’s

provisions, environmental groups, and individuals who live and work in the Houston-Galveston

area. Finding that the EPA’s final rule approving the Houston SIP is not arbitrary, capricious, or

otherwise not in accordance with law, we deny the petitions for review.

                                        I. BACKGROUND

A.     Statutory Background

       The Clean Air Act (“CAA” or “the Act”), 42 U.S.C. §§ 7401–7671q (2000), establishes a

comprehensive program for controlling and improving the nation’s air quality through state and

federal regulation. The EPA is responsible for, among other things, identifying air pollutants that

endanger the public health and welfare and formulating National Ambient Air Quality Standards

(“NAAQS”) that specify the maximum permissible concentrations of those pollutants in the




       *
         The Honorable Jane A. Restani, U.S. Court of International Trade, sitting
by designation.
                                                 2
ambient air. 

The EPA administrator has promulgated NAAQS for various

pollutants, including ozone. See 40 C.F.R. § 50.9(a) (promulgating the one-hour ozone standard

relevant to the present dispute). The one-hour NAAQS for ozone is .12 parts per million

(although EPA often refers to it as 124.5 parts per billion (“ppb”)), and an area attains the

standard when maximum measured hourly average ozone concentrations exceed the NAAQS no

more than one day per calendar year. See 

       While the EPA determines the standards of air quality, the CAA imposes upon the states

the primary responsibility for ensuring that the ambient air meets the NAAQS for the identified

pollutants. 42 U.S.C. § 7407(a). The Act requires each state to prepare a one-hour ozone

attainment demonstration SIP providing for the implementation, maintenance, and enforcement of

the NAAQS for each air quality control region within the state. 

The CAA

requires each SIP to, among other things: (1) “include enforceable emission limitations and other

control measures, means, or techniques . . . as may be necessary or appropriate” to meet the

statutory attainment deadline; (2) establish and operate “appropriate devices, methods, systems,

and procedures” to “monitor, compile, and analyze data on ambient air quality;” (3) include an

enforcement program; (4) contain “adequate provisions” to prohibit emissions activity that will

significantly interfere with attainment or deteriorate air quality; (5) provide “necessary

assurances” that the state has adequate resources “and authority under State (and, as appropriate,

local) law to carry out [its] implementation plan;” (6) require owners and operators of stationary

sources to monitor emissions and provide reports to the state; (7) provide for revision of the SIP

as may be necessary to comply with revised NAAQS or to implement improved methods of

attaining such standard; and (8) provide for the performance of air quality modeling as the EPA


                                                  3
directs in order to predict an air pollutant’s effect on ambient air quality. 42 U.S.C. § 7410(a)(2);

see 

(requiring attainment of the NAAQS by the statutory deadline in designated

nonattainment areas like Houston-Galveston).

       Thus, the CAA supplies the goals and basic requirements of state implementation plans,

but the states have broad authority to determine the methods and particular control strategies they

will use to achieve the statutory requirements. See Union Elec. Co. v. EPA, 

, 266

(1976) (“So long as the national standards are met, the state may select whatever mix of control

devices it desires”). After reasonable notice and public hearings, a state must adopt the SIP and

submit it to the EPA for review and approval. 42 U.S.C. § 7410(a)(1). The EPA in turn must

approve the SIP if it meets all of the CAA’s requirements. 

B.     Facts and Procedural History

       The Houston-Galveston area, a large geographic area consisting of eight counties, is one

air quality control region in Texas. Houston-Galveston has one of the most serious ozone

problems in the country. In order to comply with the CAA’s requirement for attainment of the

one-hour ozone standard by 2007, Texas adopted the attainment demonstration SIP at issue in

this case, which the EPA approved in its final rule.

       An “attainment demonstration SIP” has two components: (1) the attainment

demonstration, which is based on computer modeling that predicts whether the area will meet the

ozone standard by the statutory deadline of 2007; and (2) the state’s control strategy, which is its

plan for achieving the actual emissions reductions needed for attainment. Modeling efforts for the

Houston-Galveston SIP were complicated by a number of unique environmental factors and a




                                                  4
shortage of readily-available control options sufficient to provide the needed reductions.1

Nevertheless, Texas developed an attainment demonstration and control strategy that its analyses

confirmed would reach attainment by 2007.2

       The EPA evaluated the State’s modeling and associated analyses and determined that they

were consistent with the CAA and EPA’s implementing regulations. EPA also conducted an

exhaustive review of the State’s control strategy and found that it was as stringent or more

stringent than any other SIP in the country. See 66 Fed. Reg. at 57,178. Because the EPA

determined that the control strategy would reach attainment by 2007 and that the SIP met other

applicable requirements of the Act, EPA fully approved the Houston SIP as required by federal

law. 

see 42 U.S.C. § 7410(k)(3).




       1
        The Houston-Galveston area’s unique “land-sea breeze” meteorological
condition affects ozone formation and movement around the region, adding a “level
of complexity . . . not seen anywhere else in the country.” 66 Fed. Reg. at 57,164.
This condition causes emissions in the Houston-Galveston to begin to form ozone in
the local atmosphere, “later emissions and ozone formed are transported out over
the warm air over the Gulf of Mexico where the warmer temperatures further
activate the chemistry to form more ozone which is then transported back inland
over the area.” 

to overcoming this meteorological complexity, the
Houston SIP had to reduce nitrogen oxide (“NOx”) emissions by 71 percent to
reach attainment by 2007, despite a shortage of control options. 

       2
        The State’s sophisticated computer modeling simulated ozone formation in
the Houston-Galveston area despite the land-sea breeze phenomenon, and additional
analyses confirmed that the control strategy would achieve attainment by the
statutory deadline.
                                                  5
       Petitioners, most of whom participated in the underlying administrative proceedings, now

challenge EPA’s approval of the Houston SIP and Texas’s control strategy.3 Two petitioners,

BCCA Appeal Group (“BCCA”)4 and Brazoria County,5 essentially oppose the Houston SIP

because they believe some of its control measures are too stringent and will nevertheless fail to

attain the NAAQS for ozone. BCCA raises numerous legal challenges to the EPA’s approval of

Texas’s modeling and the State’s use of other analytical methods used to demonstrate attainment,

claims that the control strategy will not achieve the statutory standard, and asserts that the EPA

erred in approving the SIP while withholding final action on certain state-adopted rule provisions

integral to the SIP’s control strategy. Brazoria County maintains that certain components of

Texas’s emissions control strategy violate state law and that EPA’s approval of those measures

was arbitrary and capricious.

       Conversely, the environmental petitioners believe the Houston SIP does not go far enough

in adopting sufficient control measures to achieve attainment by the statutory deadline.

Petitioners Environmental Defense, David Marrack, Jane Elioseff, Galveston-Houston

Association for Smog Prevention, and the Galveston Bay Conservation and Preservation

Association (collectively “Environmental Defense” or “ED”) challenge Texas’s attainment


       3
       Petitioners apparently divided up the issues for briefing, and the
environmental petitioners generally adopt each other’s arguments.
       4
        BCCA Appeal Group is comprised of owners and operators of stationary
sources of air pollution that are subject to Texas’s control measures.
       5
        Brazoria County is one of eight counties within the Houston-Galveston area.
Fort Bend County initially joined Brazoria County’s petition for review, but later
voluntarily withdrew from the action. Fort Bend County’s petition was terminated
on November 26, 2002.
                                                 6
demonstration, the SIP’s inclusion of an enforceable commitment to adopt additional control

measures in order to achieve attainment, and the motor vehicle emissions budgets (“MVEB”)

included in the SIP. Petitioner Natural Resources Defense Counsel (“NRDC”) raises various

challenges to EPA’s approval of the SIP’s Voluntary Mobile Emissions Reductions Program

(“VMEP”) and the Texas Emissions Reduction Program (“TERP”). Petitioner Sierra Club’s

claims focus on the Houston SIP’s treatment of reasonably available control measures (“RACM”).

                          II. JURISDICTION AND STANDARD OF REVIEW

        The court has jurisdiction to review the EPA’s approval of the Houston SIP pursuant to

section 307(b)(1) of the CAA, 42 U.S.C. § 7607(b)(1). The court must uphold the EPA’s

findings, conclusions, and ultimate action approving the Houston SIP unless “arbitrary,

capricious, an abuse of discretion, or otherwise not in accordance with law.” Administrative

Procedure Act, 5 U.S.C. § 706(2)(A) (2000); see Sierra Club v. United States EPA, 

, 739 (5th Cir. 2002).

                                          III. DISCUSSION

        The APA’s standard of review is narrow. A rule is “arbitrary and capricious” only where

the agency has considered impermissible factors, failed to consider important aspects of the

problem, offered an explanation for its decision that is contrary to the record evidence, or is so

irrational that it could not be attributed to a difference in opinion or the result of agency expertise.

Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 

, 43 (1983). Thus,

agency decisions will be upheld so long as the agency “examine[s] the relevant data and

articulate[s] a satisfactory explanation for its action including a ‘rational connection between the

facts found and the choice made.’” Burlington Truck Lines v. United States, 

, 168


                                                   7
(1962) (citation omitted). A reviewing court must be “most deferential” to the agency where, as

here, its decision is based upon its evaluation of complex scientific data within its technical

expertise. Baltimore Gas & Elec. Co. v. NRDC, 

, 103 (1983).

       The court applies the two-step Chevron analysis to questions involving the EPA’s

interpretations of the statutes it administers. See Chevron U.S.A., Inc. v. NRDC, 

,

842–43 (1984); Texas Office of Pub. Util. Counsel v. FCC, 

, 320 (5th Cir. 2001). If

Congress “has directly spoken to the precise question at issue,” the agency and the court “must

give effect to the unambiguously expressed intent of Congress.” 

–43.

If the statute, however, is “silent or ambiguous with respect to the specific issue,” 

the

court must first assess the administrative decision-making process itself to determine whether the

agency’s action is entitled to Chevron deference. United States v. Mead Corp., 

,

226–31 (2001). Under Mead, Congress must have “delegated authority to the agency generally

to make rules carrying the force of law,” and the agency interpretation claiming deference must

have been promulgated in the exercise of that authority. 

If the agency’s decision is

a result of a sufficiently formal and deliberative process to warrant deference, see 

the

second step of Chevron requires the court to assess whether the agency’s interpretation is “based

on a permissible construction of the statute.” 

. If the agency’s

interpretation is reasonable, it will be upheld. See Smiley v. Citibank, N.A., 

,

744–45 (1996) (explaining that a “permissible” interpretation is a “reasonable” one); Texas Office

of Pub. Util. 

(“The question is not whether we might have preferred

another way to interpret the statute, but whether the agency’s decision was a reasonable one.”).




                                                  8
Federal courts accord “great deference” to the EPA’s construction of the Clean Air Act. Union

Elec. Co. v. EPA, 

, 256 (1976) (citing Train v. NRDC, 

, 75 (1975)).

        In the present case, Congress’s enactment of the CAA delegated authority to the EPA to

review SIPs for their compliance with the statute and EPA’s implementing regulations. 42 U.S.C.

§ 7410(k). The EPA exercised its statutory authority in approving the Houston SIP through

notice-and-comment rulemaking proceedings. Because notice-and-comment rulemaking is a

formal process, EPA’s final rules approving the Houston SIP, to the extent they involve the

reasonable resolution of ambiguities in the CAA, will be afforded Chevron deference. See

Whitman v. Am. Trucking Ass’ns, 

, 484 (2001) (citations omitted).

A.      Whether BCCA Has Standing to Challenge the SIP Approval

        As an initial matter, EPA argues that BCCA lacks standing to challenge EPA’s approval

of the attainment demonstration. BCCA bears the burden of establishing Article III standing. See

Lujan v. Defenders of Wildlife, 

, 561 (1992); Am. Forest & Paper Ass’n v. United

States EPA, 

, 296 (5th Cir. 1998). BCCA must demonstrate that (1) it has suffered

an “injury in fact” that is (2) fairly traceable to the defendant’s conduct and not “some third party

not before the court,” and (3) it is “likely . . . that the injury will be redressed by a favorable

decision.” 

–61 (citations omitted).

        According to the EPA, BCCA is injured, if at all, as a result of the specific control

measures Texas adopted in the Houston SIP. Accordingly, EPA argues that BCCA’s injury is not

fairly traceable to EPA and would not be redressed by a favorable decision. EPA also claims that

BCCA has not suffered an injury in fact with respect to certain point source NOx control

provisions and, therefore, lacks standing to challenge those provisions. BCCA accuses the EPA


                                                    9
of mis-characterizing the legal nature of its claims by confusing its motive with its method. While

BCCA’s motive may be to alter the mix of various air quality control measures in the Houston

SIP, its legal challenge is based on the EPA’s statutory authority to approve a SIP when it is

allegedly based on flawed modeling, fails to demonstrate attainment of the ozone standard, and

materially differs from the proposed SIP. Thus, BCCA maintains that its injury, being subject to a

SIP that fails to meet minimum statutory standards, was caused by the EPA and would be

redressed by a decision in its favor. We agree.

       This court has held that companies forced to comply with EPA regulatory requirements

meet each prong of the standing test. American Forest & Paper Association involved a group of

industries that were subject to a state’s pollutant discharge permit 

.

Although permitting authority was statutorily vested in the EPA, the agency could delegate its

authority to a state if the state’s program demonstrated compliance with statutorily enumerated

requirements. The court noted that “EPA does not enjoy wide latitude in deciding whether to

approve or reject a state’s proposed permit program. ‘Unless the Administrator . . . determines

that the proposed state program does not meet [the specified] requirements, he must approve the

proposal.’” 

(quoting Save the Bay, Inc. v. EPA, 

, 1285 (5th Cir. 1977)).

The court held that petitioners subject to the EPA-approved permit program had standing to

challenge the EPA’s approval of the program. 

       The present case is analogous to that in American Forest & Paper. Like the permitting

program at issue there, the EPA’s role in approving air pollution control plans is limited. The

EPA must approve a plan if it meets minimum statutory requirements, and states are free to

impose stricter measures. See 42 U.S.C. § 7410(k)(3); Union Elec. 

(“the


                                                  10
Administrator ‘shall approve’ the proposed state plan” if it satisfies the statutory criteria, and the

Administrator “is not to be concerned with factors other than those specified”). If, however, the

EPA approves a plan that does not meet the statutory standards, the regulated industry has

standing to challenge that action.

       Contrary to EPA’s characterization of BCCA’s claims, the legal basis for BCCA’s

challenge is the SIP’s compliance with the minimum requirements of the CAA, not the SIP’s

stringency.6 BCCA has suffered an injury in fact in the form of the costs of compliance with the

Houston SIP’s control measures, along with the added risk that its members will be responsible

for any penalties assessed for the Houston SIP’s failure to attain the ozone standard by the

statutory deadline. BCCA’s injury is fairly traceable to EPA because, if BCCA’s claims are true,

EPA unlawfully approved a SIP that failed to meet minimum statutory requirements. Finally,

BCCA’s injury is redressable by a decision in its favor because EPA would be forced to

disapprove the portions of the Houston SIP the court finds incompatible with the CAA. See 42

U.S.C. § 7410(k)(3). In addition, the approved SIP is federally enforceable against BCCA’s

members, so that a favorable decision would redress that threat of injury as well. See 

        Thus, this case is distinguishable from cases where petitioners attempted to
challenge SIPs that not only met the CAA’s minimum requirements, but also
exceeded them. Because the states can adopt more stringent air pollution control
measures than federal law requires, the EPA is empowered to disapprove state plans
only when they fall below the level of stringency required by federal law. See, e.g.,
Duquesne Light Co. v. United States EPA, 

, 611, 613 (3d Cir. 1999)
(holding that petitioner lacked standing to sue the EPA when its injury was being
subject to state control measures that were more stringent than the CAA requires).
                                                  11
7413(b)–(c) (subjecting violators to civil and criminal penalties). Accordingly, the court finds that

BCCA has established its right to sue the EPA under Article III.

B.      Whether BCCA and Brazoria County Waived Arguments Not Properly Raised
        During the Administrative Process

        EPA argues that even if BCCA has standing to pursue its challenge to EPA’s action,

BCCA failed to raise, or properly present, its challenges to the Houston SIP during the comment

period. EPA also asserts that, because none of the arguments advanced by petitioner Brazoria

County were raised in the administrative proceedings before the EPA, the County is prohibited

from seeking judicial review of these matters.

        1.      BCCA

        Although the EPA has an obligation to give full consideration to significant comments

raised during the administrative proceedings, “it is still incumbent upon intervenors who wish to

participate [in the proceeding] to structure their participation so that it is meaningful, so that it

alerts the agency to the intervenors’ position and contentions.” Vermont Yankee Nuclear Power

Corp. v. NRDC, 

, 533 (1978). A commenter in administrative proceedings before

the EPA has the burden of clarifying its position by specifying “why and how” the submitted

documentation is relevant to its position. Northside Sanitary Landfill, Inc. v. Thomas, 

, 1520 (D.C. Cir. 1988). In Northside, the D.C. Circuit held that a commenter who

submitted 420 pages of technical documentation without explaining the relevancy of the

documents, or the “specific relationship between any of the documents and Northside’s

objections,” had not properly raised the objections during the comment 

1518–20. “[T]he mere submission of voluminous documentation to the EPA” is not enough. 

As the Supreme Court has explained,

       administrative proceedings should not be a game or forum to engage in unjustified
       obstructionism by making cryptic and obscure references to matters that “ought to be”
       considered and then, after failing to do more to bring the matter to the agency’s
       attention, seeking to have that agency determination vacated on the ground that the
       agency failed to consider matters “forcefully presented.”

Vermont 

–54.

       EPA concedes that the documentation supporting BCCA’s challenges to its approval of

the Houston SIP—deposition transcripts, hearing transcripts, and exhibits compiled in connection

with a state court action—was submitted to the EPA during the comment period. EPA argues,

however, that BCCA, like the commenter in Northside, failed to alert the EPA to the relevant

portions of the voluminous material it submitted to the agency and to explain how that material

relates to the specific objections in its comments. EPA maintains that BCCA’s comments, much

like those in Northside, provided only cursory descriptions of the documents themselves. EPA

provides only one concrete example of this phenomena in its brief, claiming that the only direction

in BCCA’s comments with regard to more than 1,000 pages of hearing transcripts was: “In

particular, the Group points to the testimony, and related exhibits, of Dr. Harvey Jeffries and Dr.

Tom Tesche.” BCCA Comments, J.A. Tab 1a, at 3. EPA claims that this vague reference to two

experts’ testimony fails to explain how the referenced testimony is relevant to the issues BCCA

raised in its comments, where the testimony could be located in the voluminous transcripts, or

which “related exhibits” were relevant and where they could be found.

       The court has reviewed BCCA’s comments and finds that petitioner sufficiently clarified

its position for the EPA. BCCA expressed the Group’s “very serious concerns . . . about the


                                                 13
feasibility and economic consequences of the . . . NOx point source controls,” stated that the

photochemical grid modeling that formed the basis for the State’s control strategy was

“fundamentally flawed,” and that the proposed control strategy “will not attain the standard.”

BCCA Comments, J.A. Tab 1a, at 2. Contrary to EPA’s characterization of the comments

regarding the expert testimony, the court finds that BCCA sufficiently explained the issue to

which the testimony was relevant. See 

It is apparent when reading the entire paragraph

from which the EPA cherry picks its argument that the testimony relates to BCCA’s objections to

the SIP’s NOx controls, which were based on the photochemical modeling results. BCCA raised

objections to the modeling as well, claiming that its failure to account for ozone “spikes” rendered

it fundamentally flawed, and BCCA commented on the statistical and graphical tests used to

validate the model. See infra Part III.C. The comments also criticized the EPA’s approval of the

supplemental analyses Texas employed in formulating its control strategy, i.e., the quadratic

equation and its inputs, as well as the EPA’s weight-of-evidence analysis. Accordingly, the court

concludes that BCCA was a meaningful participant in the administrative proceedings, and EPA’s

arguments to the contrary fail.

        2.      Brazoria County

        Petitioner Brazoria County challenges for the first time speed limit reductions,7 vehicle

inspection/maintenance provisions, and lawn equipment operating restrictions8 in the Houston



        7
            Under the SIP, the speed limit would be lowered to 55 mph effective May 1,
2002.
        8
         Use of lawn equipment would be limited to certain afternoon hours between
April 1 and October 31 of each year.
                                                 14
SIP. The County’s central argument is that these provisions were adopted by Texas in violation

of state law and, therefore, the EPA’s approval of the SIP is arbitrary and capricious. EPA

responds that, because the County, or any other petitioner, failed to raise these issues during the

administrative proceedings, the arguments are waived.9 The court agrees.

       Generally, in considering a petition for review from a final agency order, this court will not

consider questions of law which were neither presented to nor passed on by the agency. E.g.,

Bass v. United States Dep’t of Agriculture, 

, 964 (5th Cir. 2000); Southwestern Bell

Tel. Co. v. Public Util. Comm’n, 

(5th Cir. 2000); Myron v. Martin, 

, 51

(5th Cir. 1982); Inst. for Tech. Dev. v. Brown, 

, 449 n.3 (5th Cir. 1995). In 1998,

the court specifically held that challenges to EPA action are waived by the “failure to raise the

objections during the notice and comment period.” Texas Oil & Gas Ass’n v. United States EPA,

, 933 n.7 (5th Cir. 1998) (citing United States v. L.A. Tucker Truck Lines, Inc., 

, 35–37 (1952)).10 For the federal courts to review a petitioner’s claims in the first

       9
         The City of Houston and the State of Texas, intervenors in this action,
concur with EPA’s waiver argument. These intervenors also argue that this court
lacks jurisdiction to consider the alleged errors of state law raised by the County’s
petition for review. Because Brazoria County’s claims can be disposed of upon the
court’s finding of waiver, see discussion infra, the difficult question of whether the
federal courts should properly adjudicate state law issues relating to state
implemention plan development need not be addressed.
       10
         Circuit precedent does conflict in this area, see City of Seabrook v. United
States EPA, 

(5th Cir. 1981) (concluding that a party’s failure to raise
challenges to a SIP during the administrative process did not waive those arguments
on appeal) and Am. Forest & Paper Ass’n v. EPA, 

, 295 (5th Cir.
1998) (following City of Seabrook), but the court chooses to follow Texas Oil &
Gas as the most closely analogous case and the better rule of law. In Seabrook, the
                                                                         (continued...)
                                                 15
instance would “usurp[] the agency’s function” and “deprive the [EPA] of an opportunity to

consider the matter, make its ruling, and state the reasons for its action.” Unemployment Comp.

Comm’n v. Aragan, 

, 155 (1946). This is a basic tenent of administrative law

generally. E.g., L.A. Tucker Truck 

(“Simple fairness . . . requires as a

general rule that courts should not topple over administrative decisions unless the administrative


       10
         (...continued)
court’s primary basis for finding no waiver was that “[t]he EPA has cited no
authority for the proposition that an argument not raised during the comment period
may not be raised on 

n.17. This is not the case here. In
addition, the question presented in Seabrook was a purely legal one, requiring the
appellate court’s evaluation of whether EPA complied with the statute. The court
reasoned, “Since the EPA is required by statute to make these determinations in
approving or disapproving a state plan, we do not think that petitioners needed to
notify the agency that it had acted arbitrarily.” 

By contrast, in the
present case Brazoria County argues that EPA was arbitrary and capricious in
failing to independently verify that the Houston SIP complied with state law. As
this is not an obligation expressly imposed by the CAA, the County needed to raise
its objections during the administrative proceeding and provide EPA an opportunity
to consider the issue before asserting, after the fact, that EPA was arbitrary for
failing to do so. Therefore, Seabrook is distinguishable from this case on its facts
and the law. Furthermore, in recent years the court has stepped back from
Seabrook’s holding on waiver, and has even applied the waiver doctrine to bar its
consideration of claims that were not raised before the EPA in similar situations as
the present dispute. See Texas Oil & Gas 

n.7 (recognizing,
without mentioning Seabrook, that petitions for review of EPA decisions may not
raise issues for the first time on appeal based on Supreme Court precedent); but see
Am. Forest & Paper 

(following Seabrook earlier that year).
Other federal circuits have specifically questioned Seabrook’s reasoning in light of
the general consensus in other circuits that the waiver doctrine applies to EPA
decisions. See 1000 Friends of Maryland v. Browner, 

, 228 n.7 (4th
Cir. 2001). Because the present case is distinguishable from Seabrook on the law
and the facts, the court need not resolve the conflict in the circuit at this time.
Rather, the court finds Texas Oil & Gas controlling here.
                                                16
body not only has erred but has erred against objection made at the time appropriate under its

practice.”); 

(“Practical notions of judicial efficiency, administrative

autonomy and encouraging effective agency procedures provide the basis for the general rule”)

(citing McKart v. United States, 

, 195 (1969)).

       Therefore, only in exceptional circumstances should a court review for the first time on

appeal a particular challenge to the EPA’s approval of a state implementation plan that was not

raised during the agency proceedings. See, e.g., 1000 

–28; Michigan

Dep’t of Envtl. Quality v. Browner, 

, 183 n.1 (6th Cir. 2000); Military Toxics

Project v. EPA, 

, 956–57 (D.C. Cir. 1998); NRDC v. United States EPA, 

, 1073–74 (D.C. Cir. 1994).

       Brazoria County does not dispute that it failed to raise its arguments before the EPA. The

County provides no justification for its failure to properly raise these claims during the

administrative proceedings. In addition, rather than challenging EPA’s failure to perform a

statutory duty, the County asks us to declare that the EPA acted arbitrarily in failing to perform a

function not mandated by statute. 

n.11. We decline the invitation. The County has

waived its state law-based arguments by failing to present them to the EPA.11


       11
         As mentioned above, the County claims that certain provisions in the
Houston SIP were invalid under Texas state law and, therefore, EPA’s approval of
those provisions was arbitrary and capricious. The court notes that the CAA only
requires that the states provide “necessary assurances that the State . . . will have
adequate . . . authority under State (and, as appropriate, local) law to carry out such
implementation plan (and is not prohibited by any provision of . . . State law from
carrying out such implementation plan or portion thereof).” 42 U.S.C. §
7410(a)(2)(E)(i). There is no statutory requirement that the EPA review SIP
                                                                           (continued...)
                                                 17
C.     Whether EPA’s Approval of Texas’s Attainment Demonstration Is Supported by the
       Record and Consistent with the CAA

       Section 182(c)(2)(A) of the CAA requires Texas to demonstrate that the Houston SIP will

achieve attainment of the ozone NAAQS by the statutory deadline. 42 U.S.C. § 7511a(c)(2)(A).

“This attainment demonstration must be based on photochemical grid modeling12 or any other

analytical method determined . . . to be at least as effective.” 

EPA

approved Texas’s attainment demonstration for the Houston-Galveston area because the agency

concluded, based on all the evidence, that the area would reach attainment of the NAAQS for



       11
          (...continued)
submissions to ensure compliance with state law, contrary to the County’s position
in this litigation. Such a requirement would be extremely burdensome and negate
the rationale for having the state provide assurances in the first instance. The EPA
is entitled to rely on a state’s certification unless it is clear that the SIP violates state
law, and proof thereof, such as a state court decision, is presented to the EPA during
the SIP approval process. See Ohio Envtl. Council v. United States EPA, 

, 27–29 (6th Cir. 1979) (holding that EPA was not arbitrary and capricious in
relying upon a state’s certification that its action was lawful and enforceable under
state law).
       12
          Photochemical grid modeling is computerized air quality modeling “that
evaluates how emissions from various sources combine in the atmosphere and
predicts the concentration of pollutants that likely will result.” 1000 

n.4. This technique employs complex computer models that can predict
ozone levels as of the statutory attainment date based on monitoring data,
meteorology, the area’s projected growth, planned emission reductions, and other
factors. 

        The photochemical grid model overlays a grid on the geographical area being
studied and simulates emissions and ozone concentrations in each “cell” of the grid.
66 Fed. Reg. at 57,165. Each cell is a three-dimensional block of space, about 4
kilometers by 4 kilometers in area, with a depth of 50 meters. 

grid
cell, the model simulates the average air quality conditions present in that space. 

by 2007 and that no additional measures would advance the attainment date. See 66 Fed.

Reg. at 57,160.

       Texas’s attainment demonstration includes both photochemical grid modeling and

supplemental analyses that EPA considered in its “weight-of-evidence” analysis. As an initial

matter, BCCA contends that Texas’s photochemical grid modeling was flawed because it failed to

simulate rapidly-forming ozone peaks, known as “spikes,”13 and overestimated ozone formation in

other parts of the Houston-Galveston area.14 Thus, BCCA argues that EPA’s reliance on the

faulty modeling renders its approval of the Houston SIP arbitrary and capricious. In addition,

BCCA and Environmental Defense both question the legality of EPA’s weight-of-evidence

approach that considers other analytical methods in addition to the attainment demonstration

itself. Part 1 of this section discusses petitioners’ challenges to Texas’s photochemical grid

modeling and EPA’s approval thereof. Part 2 addresses the legality of EPA’s decision to approve




       13
         Ozone “spikes” are dramatic increases in measured ozone over short
periods of time, ranging from several minutes to an hour. BCCA points out in its
brief that from 1990 to 1998, there were 106 instances of ozone spikes measured by
monitors in the Houston-Galveston area. That amounts to approximately one spike
per month in an area covering 220,000 square kilometers. BCCA argues that the
model’s failure to predict the exact location of spikes renders it irreparably flawed.
BCCA similarly argues that the EPA’s approval of the Houston SIP was contrary to
law because the control strategy does not control for spikes and, therefore, cannot
bring the area into attainment with the ozone standard. 

These arguments
are rejected for the reasons explained in Part III.C.1, infra.
       14
        The agency’s final rule considered, but rejected, identical criticisms of
Texas’s modeling raised during the comment period. See 66 Fed. Reg. at
57,164–65.
                                                 19
the Houston SIP after determining that, based on all the record evidence, the SIP would achieve

attainment of the ozone NAAQS by 2007.

        1.      EPA’s Approval of Texas’s Attainment Demonstration

        Texas used an EPA-approved photochemical grid model, the Comprehensive Air Quality

Model with Extensions, in its attainment demonstration for the Houston-Galveston area. Texas

applied the model to a large geographic region, covering over 220,000 square kilometers, to

ensure that all the major emission sources were included in the model’s results. Texas adjusted

the model to account for the unique land-sea breeze phenomenon conducive to ozone formation

in the region. 

n.1 and accompanying text. Before using the model to predict future

ozone concentrations, however, Texas validated the model by performing a test run (“the base

case”) that compared the model’s predictions with actual air quality data for a chosen time period.

The base time period covered four days, September 8–11, 1993, that featured both high ozone

concentrations and the land-sea breeze weather patterns characteristic of the Houston-Galveston

area. Following EPA-accepted protocols, Texas then entered emissions data into the model for

the base time period, ran the model, and compared the predicted results with actual ozone

concentrations measured at 34 air quality monitors in Houston-Galveston during the base time

period. Texas then applied a battery of tests and analyses set forth in EPA guidance, including

diagnostic and sensitivity analyses, graphical displays, and statistical tests, which collectively

demonstrated that the model’s base case performance was acceptable. See 66 Fed. Reg. at

57,164.

        Once it validated the model, Texas used it to predict ozone concentrations for the 2007

attainment date based on anticipated changes in the number and type of emissions sources. The


                                                  20
state then introduced its proposed control strategy and ran the model’s “attainment test,” which

compared the predicted ozone levels to the NAAQS for ozone for all the grid cells in the selected

episode. The model showed that Texas’s proposed strategy would significantly reduce ozone

concentrations in the Houston-Galveston area, but not enough to meet the applicable ozone

standard.

         In response, Texas, following an approach set forth in EPA guidance, estimated the

additional emissions reductions necessary to bring the area into attainment, and adopted a revised

control strategy that would provide for attainment by 2007. 66 Fed. Reg. at 57,160. The revised

control strategy includes additional control measures and an enforceable commitment15 to adopt

even more control measures after a mid-course review in 2004 and submission of a SIP revision

to the EPA. 

the attainment demonstration, including the photochemical grid

modeling, as part of the final rule under review in these petitions. See 

grid models are imperfect tools for predicting future air quality, a

modeled attainment demonstration “provide[s] a reasonable expectation that the measures and

procedures outlined will result in attainment of the NAAQS by [the statutory deadline].” 1996

Modeling Guidance, J.A. Tab 43, at 3. “[A] reviewing court must remember that the [agency] is

making predictions, within its area of special expertise, at the frontiers of science. When

examining this kind of scientific determination, as opposed to simple findings of fact, a reviewing

court must generally be at its most deferential.” Baltimore Gas & Elec. Co. v. NRDC, 

, 103 (1983). The court’s role is to evaluate whether the EPA’s projections represent arbitrary


         15
              The legality of the state’s enforceable commitment is discussed infra Part
III.E.
                                                 21
or capricious exercises of its authority, not whether they are accurate. Texas v. EPA, 

, 301 (5th Cir. 1974). Finally, there is a presumption of regularity to the EPA’s choice of

analytical methodology, so challenging parties must overcome a “considerable burden.” Am.

Petroleum Inst. v. EPA, 

, 983 (5th Cir. 1986).

       The court finds that EPA’s reliance on the model’s results was not arbitrary, capricious, or

contrary to law. In its final rule, the EPA addressed BCCA’s concern that the model fails to

account for ozone spikes. See 66 Fed. Reg. at 57,165. EPA explained that while air quality

monitors measure ozone concentration at one fixed point in space, ozone concentrations can vary

significantly over a grid cell. 

grid model, by definition, averages natural

conditions over the volume of each grid cell. 

EPA does not expect comparisons

between model predictions and monitor observations to exactly match. 

found that

while the Texas model has difficulty replicating rapid increases in ozone, “[t]his is to be expected

and does not necessarily call into question the model’s utility as a tool to predict the level of

emission reductions needed to reach attainment.” 

that the Texas model

“provides reasonable predictions of ozone levels as confirmed by comparisons with monitoring

data and therefore can provide an acceptable estimate of the amount of emissions needed for

attainment.” 

is reasonable and is supported by the record, and, therefore,

EPA’s determination is entitled to deference.

       The EPA’s final rule also addressed BCCA’s concern that the photochemical grid model

both over- and under-predicted ozone in some areas. See 

EPA’s final rule explains

that the model was validated by a battery of diagnostic and sensitivity analyses and graphical and




                                                  22
statistical performance measures. 

“Unpaired Peak Accuracy,”16 “Normalized

Bias,”17 and “Gross Error”18 statistical test results were all within the suggested limits in the EPA



       16
           The “Unpaired Peak Accuracy” test measures the difference between the
highest observed ozone concentration at any time at any monitoring station and the
highest model-predicted ozone concentration at any time in any grid cell. 1991
Modeling Guidance, J.A. Tab 42, at 54 & 81. BCCA asserts that Texas’s passing
statistic for this test was achieved by manipulating the model between 1997 and
2000. EPA responds that this point was not raised, and that BCCA relies upon
matierials and testimony not properly identified, during the comment period.
Regardless, the record reflects only that Texas improved its modeling in 1997,
which resulted in an overprediction of ozone concentrations in one grid cell. As
discussed infra, EPA found that the overprediction resulted from the model’s
difficulty in simulating wind speed and direction, and that opinion is entitled to
deference.
        Furthermore, BCCA’s claim that the model should not have passed the
Unpaired Peak Accuracy test because it predicted an ozone peak approximately 45
miles from the actual peak location is misplaced. This test measures the difference
between the highest observed ozone value and the highest model-predicted value
over all hours and monitoring stations. Geographic location is simply not a factor in
that test. Rather, the Normalized Bias and Gross Error tests take geographic
proximity into account, and the model passed those tests for all days in the base
time period.
       17
          The “Normalized Bias” test measures the model’s ability to replicate
observed patterns during times when available monitored data and modeling data
are most likely to represent similar spatial scales. 1991 Modeling Guidance, J.A.
Tab 42, at 54. EPA concluded that Texas’s model performed acceptably on the
Normalized Bias test because it produced results within EPA’s suggested ranges for
all days in the base time period. BCCA argues that these passing statistics mask
“compensating errors” in the model, i.e., an ozone over-prediction in one area
compensates for an under-prediction in another area. EPA’s Modeling Guidance
recognizes that this test can be subject to compensating errors. 1991 Modeling
Guidance, J.A. Tab 42, at 82. This is why EPA recommends that states apply a
variety of statistical and other tests to validate models. See 

53–57.
                                                                          (continued...)
                                                 23
Guideline for each day of the base case period, which demonstrated that the model was predicting

ozone concentrations in the Houston-Galveston area with a reasonable degree of accuracy. 66

Fed. Reg. at 57,164. EPA considered all model performance measures and concluded that the

model performed well. 

EPA recognized that the graphical model performance for

one day of the base period indicated that the model underestimated ozone at some locations and

overestimated ozone at others, the EPA attributed the error to the model’s difficulty replicating

wind speed and direction due to the area’s unique land-sea breeze phenomenon.19 Id.; 

n.1. Nevertheless, because the diagnostic and sensitivity tests revealed no flaws in model

formulations, and statistical measures confirmed that the model generally predicted the right

magnitude of ozone peaks, the EPA determined that the model provided an acceptable tool for


       17
        (...continued)
Regardless, the notion that the Texas model masked compensating errors in the
Normalized Bias test is directly refuted by the fact that the model passed the Gross
Error statistical test. See infra n.18.
       18
         The “Gross Error” test provides an overall assessment of the model’s
precision. 1991 Modeling Guidance, J.A. Tab 42, at 54. In contrast to the
Unpaired Peak Accuracy test, the Gross Error test compares predicted and
monitored ozone values in the same geographic area. Under-predictions and over-
predictions do not offset each other in the Gross Error test because the test assigns
positive values to both. 

The Texas model passed the Gross Error test for
each of the four days in the modeled episode, which showed that the model was
estimating ozone concentrations with an acceptable degree of geographic precision.

       19
         Although it apparently did not raise the issue during the comment period,
BCCA now asserts that Texas should have considered two additional graphical
tests. The court declines to address the issue as EPA did not have the opportunity
to do so in the first instance. 

Part III.B.2 (applying the waiver doctrine to
bar our consideration of issues not raised during the administrative proceedings).
                                                24
estimating the amount of emissions reductions needed for attainment. 66 Fed. Reg. at 57,165–66.



       Because EPA considered BCCA’s arguments during the administrative process and

offered a rational explanation for its reliance on the model despite the model’s inability to exactly

replicate Houston-Galveston’s unique meteorological conditions, the court upholds EPA’s

approval of Texas’s photochemical grid model. See Burlington Truck 

;

Motor Vehicle Mfrs. 

. EPA recognized the model’s shortcomings in the

final rule and provided plausible explanations that were supported by the record. In light of the

reasonable explanation for the model’s discrepancies, and given the fact that the model

nevertheless performed well on the full battery of validation tests, EPA’s acceptance of Texas’s

modeling was neither arbitrary nor capricious. Furthermore, because EPA’s decision was based

upon its evaluation of complex scientific data within its technical expertise, the court is mindful of

its obligation to be “most deferential” to the EPA’s approval of Texas’s photochemical grid

modeling. See Baltimore Gas & Elec. 

. Therefore, the court concludes that

EPA’s acceptance of Texas’s photochemical grid model is supported by substantial record

evidence and is otherwise in accordance with law.

       2.      EPA’s Weight-of-Evidence Analysis

       The statute requires that an attainment demonstration be “based on photochemical grid

modeling or any other analytical method determined . . . to be at least as effective.” 42 U.S.C. §

7511a(c)(2)(A). Due to the inherent uncertainties in air quality modeling, EPA has interpreted the

statute to allow states to supplement their photochemical modeling results with additional

evidence to demonstrate attainment. 66 Fed. Reg. at 57,170; see 40 C.F.R. 51.112(a)(1)


                                                 25
(referencing Appendix W as listing the applicable requirements for gauging the adequacy of a

control strategy); 40 C.F.R. Pt. 51 App. W § 10.0 (recognizing limitations of modeling and

allowing consideration of evidence in addition to modeled attainment demonstrations). EPA’s

modeling guidance was updated in 1996, and again in 1999, to permit states to use the weight-of-

evidence approach to assess additional emissions reductions that are part of its SIP but were not

modeled. See 1996 Modeling Guidance, J.A. Tab 43, at 1–2; 1999 Modeling Guidance, J.A. Tab

44, at 1–2. According to EPA, this approach is consistent with the statute “because the modeling

results constitute the principal component of EPA’s analysis, with supplemental information

designed to account for uncertainties in the model.” 66 Fed. Reg. at 57,170. The question is

whether EPA’s interpretation of the statute it administers is entitled to deference under Chevron

and its progeny.

       As previously discussed, Texas’s modeled control strategy alone did not demonstrate

attainment of the NAAQS for ozone by the 2007 statutory deadline. In accordance with EPA

guidance, Texas supplemented its modeling with other evidence to show that the area would

reach attainment by 2007. Texas first used a quadratic equation to calculate the “gap” in NOx

emissions reductions between those achieved by the modeled control strategy and the levels

required to achieve the NAAQS.20 See 66 Fed. Reg. at 57,172. Texas then revised its final



       20
         A commenter before the EPA, and petitioners here, criticize EPA’s
technique for estimating the ambient impact of additional emissions reductions not
modeled on grounds that EPA employed a rollback modeling technique that is
precluded under EPA regulations. See 66 Fed. Reg. at 57,172. The “proportional
rollback” approach is derived from a purely empirical/mathematical relationship that
assumes that if emissions from each source are decreased by the same percentage,
                                                                       (continued...)
                                                26
control strategy to eliminate that gap and provide for attainment by 2007. EPA weighed all of the

evidence and concluded that the control measures Texas adopted would more likely than not lead

to attainment. 

rule indicated that a majority of the control measures were

modeled, and EPA evaluated the impact of additional emissions reductions that were not modeled

by reviewing the model’s response to changes in emissions and observed air quality changes. 

its decision as “strengthened by the state’s commitment to check progress towards

attainment in 2004 and to adopt additional measures, if the anticipated progress is not being

made.” 

and Environmental Defense challenge the EPA’s approval of the

Houston SIP based on its weight-of-evidence determination. The court finds, however, that the




       20
         (...continued)
ambient air quality concentrations decrease proportionately. 

20 percent
improvement in ozone is needed for attainment, it is assumed that a 20 percent
reduction in emissions would be required. 

not rely on this approach in its evaluation of the Houston-Galveston
attainment demonstration. 

a locally-derived relationship between
emissions reductions and improvement in ozone levels as determined by the model
and/or observed changes in air quality. 

if modeled or monitored
results indicated that ozone was reduced by 25 ppb during a particular period, and
that VOC emissions fell by 20 tons per day and NOx emissions fell by 10 tons per
day during that period, EPA would develop a relationship for ozone improvement
related to those reductions in emissions. 

formula assumes a quadratic
relationship between the emissions and ozone for a relatively small amount of ozone
improvement, this is not a “proportional rollback” technique. 

makes limited use of these locally-derived adjustment factors to estimate the extent
in which additional emissions reductions, not Texas’s core control strategies, would
reduce ozone levels. 

commitment to perform a mid-course review
further convinced the EPA that Texas’s overall plan would result in attainment by
2007. 

                                                27
EPA properly interprets the CAA to allow for a weight-of-evidence analysis, and that the EPA’s

conclusions in its determination here were reasonable and entitled to deference.

        The statute requires that an attainment demonstration be “based on photochemical grid

modeling;” the statute does not require that an attainment demonstration be based solely or

directly on photochemical grid modeling. See 42 U.S.C. § 7511a(c)(2)(A). As such, the statute

is ambiguous as to how the photochemical grid modeling may be used. As the statute does not

specifically govern the precise question at issue, the court must determine whether EPA’s

interpretation of it is entitled to deference.

        EPA’s weight-of-evidence approach to approving the Houston SIP was set forth in notice-

and-comment rulemaking, and as such it is entitled to deference if it is reasonable. See 

–31. EPA has interpreted the “based on” language to allow for the assessment of

additional emissions controls, not modeled, as part of the weight-of-evidence analysis.21 See 66

Fed. Reg. at 57,171. The model’s results are the “principal component” of EPA’s weight-of-

evidence determination, but the weight-of-evidence approach allows Texas to supplement the

modeled results with additional control measures. 

1996 Modeling Guidance, J.A.

Tab 43, at S-1 (“In a weight of evidence determination, model results are weighed heavily.”).

Furthermore, because the statute also grants EPA the broad authority to approve equally effective

alternatives to photochemical grid modeling, Congress could not have intended to bar EPA from



        21
         Contrary to Environmental Defense’s suggestion to the contrary, nothing in
the CAA or in EPA’s regulations or guidance requires a state to model its entire
control strategy. See 1000 

(concluding that the CAA does
not require a state to model its entire control strategy so long as EPA reasonably
concludes that attainment is likely based on the weight of the evidence).
                                                 28
considering data in addition to modeled results. See 42 U.S.C. § 7511a(c)(2)(A) (giving EPA

Administrator broad discretion to approve “other analytical method[s]”).22 As such, EPA’s

conclusion that the weight-of-evidence approach to approving attainment demonstrations is

consistent with the CAA is reasonable and is entitled to deference. See 1000 

(upholding EPA’s weight-of-evidence analysis and explicitly rejecting the argument that the

CAA requires new modeling upon the state’s revision of a SIP).

       Texas modeled 88 percent of the emissions reductions necessary for attainment, and the

results from Texas’s modeling formed the starting point for EPA’s weight-of-evidence analysis.

Petitioners contest EPA’s approval of Texas’s use of the quadratic equation to calculate the

additional reductions needed for attainment. Texas, however, followed the general approach

described in EPA’s 1999 Modeling Guidance to identify the amount of additional emissions

reductions needed for attainment. Texas evaluated the model’s reaction to three hypothetical

control strategies imposed on the modeled episode day with the highest ozone levels. Dec. 2000

SIP, J.A. Tab 12a, at 3-47. Based upon the results, Texas and EPA developed a quadratic

equation that represented the relationship between NOx emissions and ozone levels. 66 Fed. Reg.

at 57,174. Thus, the quadratic equation used in Texas’s supplemental analysis was “based on,”

and was an extension of, actual photochemical grid modeling results from Houston-Galveston,



       22
          Environmental Defense asserts that the weight-of-evidence determination is
an improper “other analytical method” under 42 U.S.C. § 7511a(c)(2)(A) because
the EPA Administrator did not make a specific determination that this approach is
“at least as effective” as photochemical grid modeling. Because weight-of-evidence
analysis is based on photochemical grid modeling, EPA did not use an “other
analytical method” that would have required the Administrator’s effectiveness
determination.
                                                29
and it was used only to assess the 12 percent portion of the state’s control strategy that was not

included in the modeling. See 

Accordingly, EPA found that the quadratic equation

was a sound scientific method for estimating the additional NOx reductions, beyond the modeled

controls, that would be necessary for the Houston-Galveston area to reach attainment. 

       As discussed, the quadratic equation was used to estimate what the model would predict

to be the peak ozone concentration for the future episode if Texas had modeled its complete

control strategy. Because the EPA’s approval of the model itself was reasonable, 

Part

III.C.1, EPA reasonably concluded that additional analyses based on and derived from the model

also predicted peak ozone levels with reasonable accuracy. Petitioners have not met their

“considerable burden” of overcoming the presumption of regularity that attaches to EPA’s choice

of analytical methodology. See Am. Petroleum 

. As Petitioners have failed

to establish that the weight-of-evidence approach to approving the Houston SIP’s attainment

demonstration is arbitrary, capricious, or otherwise not in accordance with law, we reject their

general and as applied challenges to the EPA’s weight-of-evidence determination.

D.     Whether EPA Erred in Withholding Action on Certain Provisions of the Houston
       SIP

       On November 14, 2001, the EPA approved the Texas Mass Emissions Cap and Trade

program, one element of Texas’s control strategy, as a revision to the Houston SIP. Approval

and Promulgation of Air Quality State Implementation Plans (SIP), 66 Fed. Reg. 57,252 (EPA

Nov. 14, 2001) (final rule). The program is mandatory for stationary facilities that emit NOx in

the Houston-Galveston area, and facilities are required to meet annual maximum NOx allowances.



                                                 30


bank, sell, or purchase their allowances. 

, the program has a

provision that allows facilities to use emission reduction credits, discrete emission reduction

credits (“DERCs”), and mobile discrete emission reduction credits (“MDERCs”) “in lieu of

allowances if they are generated in the [Houston-Galveston] area.” 

deferred

action on DERCs and MDERCs so that neither credit can be used until EPA approves of those

measures. See 

argues that EPA acted contrary to law and failed to observe proper procedure in

approving the Houston SIP while withholding final action on the DERC and MDERC provisions,

which would allow compliance with the mandated NOx reductions through the use of emissions

reductions credits.23 EPA responds that its decision to withhold action on the DERC and



       23
         BCCA’s opening brief also challenges EPA’s failure to act on an alternate
80 percent reduction in emissions from industrial point sources (“ESADs”), as
opposed to the EPA-approved 90 percent reduction. The use of the lower
reductions in the Texas SIP was expressly conditioned upon further study and
submission of a SIP revision to EPA. EPA’s final rule approving the SIP therefore
found that the issue was not ripe for its review because the record did not provide
the needed scientific data or modeling to show that the alternate ESAD would
achieve the NAAQS for ozone. 66 Fed. Reg. at 57,232. This is a reasonable
explanation and EPA did not err in withholding action on the alternate ESAD
reduction. Regardless, as a result of a consent order entered into between TNRCC
and BCCA in a related state court proceeding, TNRCC recently substituted the 90
percent ESAD with the 80 percent rate. TNRCC submitted the new regulations to
the EPA as a proposed revision to the Houston SIP on January 28, 2003. Letter
from Savage to Fulbruge of 5/7/2003, at 1–2. While EPA contests BCCA’s
submission of these materials and argues that they do not qualify as appropriate
supplemental authorities under Fed. R. App. P. 28(j), see Letter from Amaditz to
Fulbruge of 5/20/2003, as a result of these recent developments, the court need only
address the question of whether EPA erred in deferring action on the DERC and
                                                                         (continued...)
                                                 31
MDERC rules was within the agency’s authority under the CAA. In responding to comments

advanced during the administrative process, EPA explained that the DERC and MDERC rules are

“separate and independent” from the cap-and-trade program because they were not submitted by

the State of Texas for emission credit in the attainment demonstration. 66 Fed. Reg. at 57,255.

Further, EPA found that the use of DERCs and MDERCs is not necessary for the cap-and-trade

program to achieve emissions reductions necessary to reach attainment. 

explained that the purpose of the DERC and MDERC provisions, as defined in the Texas

Administrative Code, is to provide emission point sources with a “voluntary compliance option.”

will be evaluated for compliance with the CAA and may be approved as a SIP

revision after notice and comment. 

finds BCCA’s reasoning unpersuasive and its reliance on case law from other

circuits misplaced. Riverside Cement Co. v. Thomas, 

, 1247 (9th Cir. 1988),

concerned a state rule regulating the permissible NOx emissions from cement kilns that also

provided that, prior to the rule’s effective date, a feasibility hearing would be held to evaluate the

efficacy of the proposed emissions limits. EPA, however, approved the rule without waiting for

the feasibility hearing, thereby imposing an absolute emissions limit without regard to the

contingency the state built into the rule. 

Circuit held that this action was an

arbitrary and capricious amendment of the state’s proposal. 

Similarly, in Indiana &

Michigan Elec. v. EPA, 

, 491 (7th Cir. 1984), EPA approved a SIP provision while



       23
    (...continued)
MDERC provisions.


                                                  32
deferring action on an exception to its application, which effectively revised the state’s plan in a

manner that made it more stringent than the state intended. See also Bethlehem Steel Corp. v.

Gorsuch, 

, 1036 (7th Cir. 1984) (finding impermissible the EPA’s action on an

emissions limitation that eliminated a “blow-off allowance,” thereby increasing the stringency of

that control strategy).

        These cases are inapposite to support BCCA’s position. Unlike the situation in Riverside

Cement and Indiana & Michigan Elec., EPA did not, in effect, amend Texas’s proposal in a way

that eliminated a prerequisite to, or an exception to, the approved rule’s application. In fact,

Indiana & Michigan Elec. actually supports EPA’s action because the court stated, “if some

provisions in a plan are independent of others, there is no reason why the agency must consider all

of the provisions at the same 

. The EPA’s final rule characterized the

DERC and MDERC provisions as “separate and independent” from the cap-and-trade program it

approved, 66 Fed. Reg. at 57,255, and BCCA did not contest this finding in either its comments

before the agency or its brief to this court. Finally, the postponement of action on DERCs and

MDERCs did not make the SIP more stringent than the state intended, as was the case in

Bethlehem Steel.24 Therefore, BCCA failed to demonstrate that EPA’s action was arbitrary,

capricious, or otherwise not in accordance with law. Thus, EPA’s deferral of action on these

provisions is upheld.



        24
          While EPA’s action may take away one option for compliance, it does not
alter the obligation to reduce emissions to specified levels. This view is supported
Texas’s express agreement to the timetable EPA proposed for considering the
DERC and MDERC provisions. See Approval and Promulgation of Implementation
Plans, 66 Fed. Reg. 38,231, 38,232 (EPA July 23, 2001).
                                                  33
E.     Whether EPA’s Approval of the SIP’s “Enforceable Commitment” to Adopt
       Additional Controls on a Fixed Schedule is Consistent with the CAA

       The EPA-approved control measures in the Houston SIP achieve 94 percent of the NOx

reductions needed for attainment. 66 Fed. Reg. at 57,178 (explaining that adopted measures

include, among other things, industrial point source NOx controls and the TERP and VMEP

programs, see infra Part III.G–H). The final element of Texas’s control strategy is an enforceable

commitment25 to adopt and implement additional NOx controls on a fixed schedule to achieve an

additional 56 tons/day of NOx reductions.26 

57,178. Thus, the commitment


       25
         These commitments are enforced by the EPA and citizens under the CAA.
See 42 U.S.C. §§ 7413 (federal enforcement) and 7604 (citizen suits). EPA has
approved enforceable commitments in the past and courts have enforced them, as
well as SIP provisions generally, against the states that failed to comply with those
commitments. See, e.g., 1000 

–37 (upholding EPA’s
approval of a SIP provision “when considered with the emissions reduction
programs in place and when considered with Maryland’s enforceable commitment
to implement all other control measures necessary to reach attainment.”) (emphasis
added)); Am. Lung Ass’n v. Kean, 

(D.N.J. 1987), aff’d, 

(3d Cir. 1989) (concluding that the court has jurisdiction to adjudicate
citizen suits against the state in its regulatory capacity and enforcing state’s SIP
commitments); Friends of the Earth v. Carey, 

(2d Cir. 1976)
(enforcing provisions of state implementation plan in citizen suit); NRDC v. New
York State Dept. of Envtl. Conservation, 

(S.D.N.Y. 1987)
(finding state agency liable for failure to implement control strategies by the
timetable set in an enforceable commitment). Finally, if a state fails to meet its
commitments, EPA could find a failure to implement the SIP under section 179(a) of
the CAA, which would trigger an 18-month period for the state to begin
implementation before mandatory sanctions are imposed. 66 Fed. Reg. at 57,177
n.11.
       26
        Under the schedule, Texas was required to adopt additional controls to
provide 25 percent of the reductions covered by the enforceable commitment, or 14
                                                                     (continued...)
                                                34
addresses only 6 percent of the emission reductions necessary to attain the NAAQS for ozone.

Petitioner Environmental Defense asserts that the EPA lacks authority under the

CAA and EPA’s regulations and guidance to approve a SIP containing an enforceable

commitment to adopt unspecified control measures. EPA maintains that the SIP’s limited use of

the enforceable commitment is permissible under the statute.27

       Nothing in the CAA speaks directly to enforceable commitments. The CAA does,

however, provide EPA with great flexibility in approving SIPs. A SIP may contain “enforceable

emission limitations and other control measures, means, or techniques . . . as well as schedules

and timetables for compliance, as may be necessary or appropriate” to meet the CAA’s

requirements. 42 U.S.C. § 7410(a)(2)(A) (emphasis added); see 42 U.S.C. § 7502(c)(6) (using

the same language to describe nonattainment area plan requirements).28 Thus, according to the


       26
        (...continued)
tons/day of NOx, by December 2002. 66 Fed. Reg. at 57,161. Texas must adopt
the remainder of the necessary controls by May 1, 2004. 

City of Houston and BCCA Appeal Group support EPA’s position in
their intervenor briefs.
       28
         Environmental Defense claims that 42 U.S.C. § 7410(k)(4), which was
enacted in 1990 amendments to the CAA and gives EPA discretionary authority to
conditionally approve a SIP “based on a commitment of the State to adopt specific
enforceable measures by a date certain,” was meant to supplant EPA’s practice of
using enforceable commitments under § 7410(a)(2)(A). There is nothing in the
statute or in the legislative history, however, to support this theory. Furthermore,
the cases Environmental Defense relies upon, which address conditional SIP
approvals or construe statutory provisions before the 1990 amendments, are
inapposite.
       The legislative history of the 1990 amendments of § 7410(a)(2)(A)
                                                                           (continued...)
                                                35
plain language of the statute, SIPs may contain “means,” “techniques,” and/or “schedules and

timetables for compliance” that the EPA considers “appropriate” for attainment so long as they

are “enforceable.” See 

“Schedules and timetables” is broadly defined as “a

schedule of required measures including an enforceable sequence of actions or operations leading

to compliance with an emission limitation, prohibition or standard.” 42 U.S.C. § 7602(p). The

remaining terms are not defined by the Act. Because the statute is silent on the issue of whether

enforceable commitments are appropriate means, techniques, or schedules for attainment, EPA’s

interpretation allowing limited use of an enforceable commitment in the Houston SIP must be

upheld if reasonable.

       EPA interprets § 7410(a)(2)(A) to mean that enforceable emission limitations and other

control measures do not necessarily need to generate reductions in the full amount needed to

attain. Rather, EPA interprets the statutory language to allow limited use of other “means” and

“techniques,” such as enforceable commitments, so long as the entire package of measures and




       28
         (...continued)
acknowledged EPA’s practice of fully approving SIPs that contained limited
enforceable commitments, yet Congress made no changes to that section precluding
the practice. See H.R. Rep. No. 101-490, pt. 1 at 218 (1990). In fact, the 1990
amendments expanded EPA’s authority under § 7410(a)(2)(A) by adding the
“means” and “techniques” and “as . . . appropriate” language. This strengthens,
rather than limits, the statutory basis for EPA’s practice of fully approving SIPs that
contain enforceable commitments as part of an overall control strategy. Finally,
because § 7410(a)(2)(A) is silent on the issue of whether an enforceable
commitment is an “appropriate” “means” or “technique” to reach attainment, EPA’s
interpretation must be upheld if the court finds it a permissible construction of the
statute.
                                                36
rules provides for attainment.29 66 Fed. Reg. at 57,177. EPA generally considers three factors in

determining whether to approve a SIP’s enforceable commitment: (1) whether the commitment

addresses a limited portion of the statutorily-required implementation plan; (2) whether the state

is capable of fulfilling its commitment; and (3) whether the commitment is for a reasonable and

appropriate period of time. 

based upon its consideration of these factors,

present circumstances in Houston-Galveston, along with New York City, Philadelphia, and

Baltimore, warrant EPA’s consideration of enforceable commitments).

       In the present case, EPA determined that Texas’s limited use of the enforceable

commitment as part of its overall control strategy was appropriate within the meaning of CAA

sections 110(a)(2)(A) and 172(c)(6), 42 U.S.C. §§ 7410(a)(2)(A) and 7502(c)(6). See 66 Fed.

Reg. at 57,177. In applying the three-factor test to the Houston SIP, the EPA found that the first



       29
         Environmental Defense argues that EPA’s regulations and guidance
preclude the use of enforceable commitments. ED bases this argument in part on
the agency’s definitions in the Code of Federal Regulations of “control strategy”
and “control measures.” The defintion of control stragegy is broad and does not
support ED’s position, see 40 C.F.R. § 51.100(n)(1)–(8), and EPA has never
construed the enforceable commitment to be a “control measure” within the
meaning of the CAA. It is, rather, a “means” or “technique” specifically permitted
under the Act. ED also claims that the enforceable commitment does not constitute
a “measure[], rule[], or regulation[],” and therefore violates EPA’s regulation
requiring each SIP to “demonstrate that the measures, rules, and regulations
contained in it are adequate to provide for the timely attainment and maintenance of
the national standard it implements.” See 40 C.F.R. § 51.112(a). Contrary to ED’s
contention, however, the enforceable commitment is a “rule” as defined by the
Administrative Procedure Act. See 5 U.S.C. § 551(4). It creates specific rights,
imposes specific obligations on, and is enforceable against the State of Texas.
Therefore, we reject ED’s arguments that EPA lacks discretion under EPA
regulations and guidance to approve enforceable commitments.
                                                 37
factor supported the use of an enforceable commitment here because it only addressed a small

portion of the overall plan. In fact, the enforceable commitment addresses only six percent of the

total emission reductions needed to attain the standard. See 

The second factor also

weighed in favor of approving the commitment because Texas “provided EPA with sufficient

information to assure EPA that it will be capable of adopting controls to achieve the necessary

level of emission reductions.” 

EPA with a list of soon-to-be-available,

cutting-edge technologies that would achieve at least 56 tons/day of NOx emission reductions by

the statutory deadline,30 thereby justifying its use of the enforceable commitment as opposed to

adopted control measures. See 

Texas was in the process of exploring,

developing, and assessing the capabilities of those cutting-edge technologies, some of which were

further along in the development process than others, EPA approved the two-tiered timetable for

adopting the additional controls covered by the commitment. Id.; 

n.26. EPA

considered this timetable to be as expeditious as possible given the technological circumstances, in

addition to the time Texas would need to adopt the measures that would achieve the necessary

emission reductions. 66 Fed. Reg. at 57,178. Finally, such a commitment is enforceable and

binding upon the State of Texas. 

n.25. Therefore, EPA determined that the

enforceable commitment was an appropriate mean, technique, or timetable for compliance that

would provide for attainment by 2007. See 66 Fed. Reg. at 57,177.


       30
         The developing technologies submitted to EPA included, for example,
diesel emulsion, fuel cells, diesel NOx reductions systems, energy efficiency
measures, and several innovative ideas, such as marine loading operations and
episodic emission controls, all of which required further scientific study. Approval
and Promulgation of Implementation Plans, 66 Fed. Reg. 36,656, 36,665 (EPA July
12, 2001) (proposed rule).
                                                38
       The court finds that EPA reasonably concluded that an enforceable commitment to adopt

additional control measures on a fixed schedule was an “appropriate” means, technique, or

schedule or timetable for compliance under the statute.31 See 42 U.S.C. §§ 7410(a)(2)(A) and

7502(c)(6). The statute is silent or ambiguous on this issue. Because the EPA’s approval of the

enforceable commitment in the Houston SIP was promulgated under notice-and-comment

rulemaking, EPA’s interpretation is entitled to Chevron deference if it is based on a permissible

construction of the statute. EPA’s three-factor test reasonably evaluates whether an enforceable

commitment would be “appropriate” in a given situation. This test guarantees that a state’s use of

an enforceable commitment as part of its overall control strategy is limited in scope and that the

state is capable of fulfilling the commitment pursuant to an expeditious, yet practicable, schedule.

In the present case, EPA carefully considered each factor and found that Texas’s use of the

commitment was appropriate. Despite the uncertainty as to the exact technologies Texas will

actually employ to achieve attainment by 2007, EPA considered the possibilities Texas submitted

with its SIP and determined that the state is capable of adopting these “cutting-edge” controls to

achieve the standard. See 66 Fed. Reg. at 57,178. Because that determination was a reasonable




       31
         For over 20 years, EPA has consistently maintained its interpretation of the
CAA as allowing states to submit, and EPA to approve in appropriate
circumstances, a SIP that contains an enforceable commitment to adopt additional
controls as part of an overall control strategy. See 66 Fed. Reg. at 57,177 (citing
the use of such commitments in Arizona, California, New Jersey, and New York).
Therefore, Environmental Defense’s assertion in its brief that EPA “conjured up
‘enforceable commitments’ as a new theory” to allow its approval of the Houston
SIP is completely false.
                                                 39
one given the facts and circumstances of the severity of the ozone problem in Houston-Galveston,

it is entitled to deference.

F.      Whether EPA Properly Approved the Motor Vehicle Emissions Budget in the
        Houston SIP

        A Motor Vehicle Emissions Budget or MVEB establishes the maximum level of on-road

emissions that, when considered with emissions from all other sources, still provides for

attainment of the ozone NAAQS by the statutory deadline.32 40 C.F.R. § 93.101; 66 Fed. Reg. at

36,666. The MVEB is a necessary component of a SIP that demonstrates attainment because it

identifies those vehicle emissions that can be produced without jeopardizing an area’s attainment

status. The Act’s conformity provisions integrate the MVEB with the transportation planning

process. 42 U.S.C. § 7506(c)(1). Transportation activities may only receive federal approval and

funding upon a demonstration of compliance with the SIP. 

conforms with the SIP

only if it is consistent with eliminating or reducing the severity and number of air quality violations

and achieves expeditious attainment. 

Once the EPA approves the

MVEBs or issues an adequacy determination,33 state agencies may rely on that budget when



        32
         The EPA’s regulations specifically define the motor vehicle emission budget
as “that portion of the total allowable emissions defined in the submitted or
approved control strategy implementation plan revision or maintenance plan for a
certain date for the purpose of meeting reasonable further progress milestones or
demonstrating attainment or maintenance of the NAAQS, for any criteria pollutant
or its precursors, allocated to highway and transit vehicle use and emissions.” 40
C.F.R. § 93.101.
        33
        An adequacy determination allows transportation planners to use the
MVEBs in a submitted, but not yet approved, SIP for conformity purposes. E.g.,
1000 

.
                                                  40
determining if a proposed transportation project conforms to the relevant SIP. 1000 

. The Houston SIP contains two MVEBs, one for NOx and one for VOCs.34 66 Fed.

Reg. at 36,666. The EPA determined that the budgets, when considered with all other measures,

provided for attainment by the statutory deadline.35 

       Plaintiff Environmental Defense disagrees with the EPA’s approval of the MVEB for the

Houston SIP, claiming that (1) the EPA could not approve the budgets because the Houston SIP

does not provide for attainment, and (2) the EPA violated the transportation conformity

provisions of the CAA and its regulations. For the reasons that follow, these arguments fail.

       First, because we have already affirmed the EPA’s approval of the Houston SIP’s

attainment demonstration, which includes an enforceable commitment to adopt additional control

measures on a fixed schedule, Environmental Defense’s first challenge to the EPA’s approval of

the MVEBs must similarly be rejected. See 1000 

–37 (upholding EPA’s

approval of the MVEB “when considered with the emissions reduction programs in place and

when considered with Maryland’s enforceable commitment to implement all other control

measures necessary to reach attainment . . . .”) (emphasis added)).


       34
         The budgets for the 2007 attainment demonstration SIP are 79.5 tons/day
for VOCs and 156.7 tons/day for NOx. 66 Fed. Reg. at 57,179. Approval of the
attainment budgets was based on the current control measures in the SIP and the
enforceable commitments made for additional controls. 

EPA
recognized that the budgets would have to be revised upon Texas’s adoption of
additional transportation measures as part of the enforceable commitment. 

       35
         Intervenors Houston-Galveston Area Council, a transportation and air
quality planning agency of the State of Texas, and Harris County support EPA’s
approval of the MVEB in the Houston SIP.
                                                41
       Environmental Defense’s arguments based on the transportation conformity provisions of

the CAA and its implementing regulations are equally without merit. Environmental Defense

incorrectly argues that a budget that does not provide for attainment, does not provide the

“necessary emissions reductions” required by 7506(c)(2)(A). The statute cited requires a

transportation plan or transportation improvement program to be consistent with an approved

MVEB; it has nothing to do with EPA’s preceding action of approving the MVEB itself. See 42

U.S.C. § 7506(c)(2)(A). Moreover, 40 C.F.R. § 93.118(e)(4)(iv) provides that the EPA will not

find a MVEB adequate for transportation conformity purposes unless “the motor vehicle

emissions budget(s), when considered together with all other emissions sources, is consistent with

applicable requirements for . . . attainment.” 

finding36 here was consistent

with the regulation because EPA reasonably found that the Houston SIP’s MVEB would achieve

attainment when considered with all other sources. 66 Fed. Reg. at 57,178. Therefore,

Environmental Defense’s arguments based on this statute and its corresponding regulations are

meritless. Because EPA acted in conformity with the CAA and its own regulations, EPA’s action

approving the MVEB was not arbitrary or capricious and must be upheld.

G.     Whether EPA’s Approval of the Texas Emissions Reduction Plan is Supported by
       the Record

       The Texas Emissions Reduction Plan or TERP is a discretionary economic incentive

program to reduce emissions. At issue is TERP’s diesel emission reduction program that provides


       36
        EPA approved the budgets after conducting a detailed examination of the
SIP’s control measures and technical analyses. See 66 Fed. Reg. at 57,180.
Because the process for approving a MVEB in a SIP is more stringent than an
adequacy determination, by approving the budgets, EPA necessarily found them to
be adequate.
                                                42
financial incentives to help private and public entities purchase or lease cleaner diesel technology

for mobile sources. Tex. Health & Safety Code Ann., Title 5, Subtitle C, Ch. 386, SubChap. C.

Texas’s program is modeled after the Carl Moyer Program in California, which was very

successful. Economic incentives, like TERP, are explicitly allowed under the Act as one tool to

achieve attainment. See 42 U.S.C. § 7410(a)(2)(A). Congress directed EPA to promulgate

regulations for economic incentive programs, 

and EPA did so by enacting

regulations for statutory incentive programs and by issuing guidelines for discretionary incentive

programs. 40 C.F.R. Pt. 51, Subpt. U.

       NRDC alleges that EPA’s approval of the TERP diesel program is arbitrary and capricious

because the state did not provide adequate information about resources, implementation, and legal

authority. We agree with the EPA that the state did provide all of the information required by the

Act and that its approval of the program was reasonable on the record.

       In considering TERP, the EPA reviewed the state’s estimated costs, funding mechanisms,

funding allocations, and estimated emission reductions from this program. 66 Fed. Reg. at

57,175. Texas’s TERP program was designed to generate 18.9 tons/day of NOx reductions at an

estimated average cost of $5000 per ton. (The California program projected emissions reductions

at an average cost of $3000 per ton based on data collected through 1999). Id.; 66 Fed. Reg. at

36,665. Funding was to be provided by various fees and surcharges on vehicles. It was

ultimately determined that the Houston area would be allocated $25 million for its diesel emission

reduction program for fiscal year 2002. 66 Fed. Reg. at 57,193. The EPA determined that Texas

would achieve at least 25 tons/day of NOx reductions in Houston from TERP, an amount the

EPA determined to be sufficient to offset the emission reduction shortfall created by the Texas


                                                 43
legislature in enacting TERP. 

The EPA’s past experience with this type of diesel

emission program in California supported its conclusion that substantial NOx reductions could be

achieved with the allocated funds.

       NRDC argues that the state did not comply with certain provisions of the CAA. Section

110(a)(2)(E) of the CAA requires a SIP to “provide . . . necessary assurances that the State . . .

will have adequate personnel, funding, and authority . . . to carry out such implementation plan.”

42 U.S.C. § 7410(a)(2)(E)(I). The NRDC submits that this section requires a state to do so as to

each specific control measure, mean, or technique in a state’s SIP. This would mean that Texas

would have to make the “necessary assurances” for its TERP program. We disagree. The plain

language of the statute only requires a state to give assurances that it has the funding, personnel,

and authority to implement the plan as a whole (“such implementation plan”). It does not require

such assurances for each specific control measure.

       While the Act does not define “necessary assurances,” other circuits have held that

“Congress has left to the Administrator’s sound discretion determination of what assurances are

‘necessary.’” NRDC v. EPA, 

, 884 (1st Cir. 1973); see Friends of the Earth v.

United States EPA, 

, 1126 (2d Cir. 1974) (following First Circuit); NRDC v.

United States EPA, 

, 527 (2d Cir. 1974) (same). Texas provided a general

assurance that its fiscal and manpower resources were adequate to implement the SIP as a whole.

Dec. 2000 Plan at 1-6. In adopting the SIP as a state rule, Texas also certified that “[l]egal

counsel has reviewed and certified that this rule, as adopted, is a valid exercise of the [TNRCC’s]

legal authority.” TNRCC Order No. 2001-0410-RUL (Dec. 20, 2000); TNRCC Order No. 2001-

0410-RUL (Oct. 2, 2001). The EPA was entitled to rely on this certification. See Ohio Envtl.


                                                 44
Council v. EPA, 

, 28 (6th Cir. 1979). The Houston SIP also provided a detailed

discussion about the legal authority of state and local agencies to implement, maintain, and

enforce the plan as a whole, including citations to applicable law. Dec. 2000 Plan at iii. Based on

its past experience with Texas’s air quality program and its relationship with the state, the EPA

determined that these assurances regarding funding, resources, and legal authority met the

minimum requirements of § 7410(a)(2)(E). Furthermore, as discussed above, the EPA also

independently examined TERP’s funding and resources and determined they were adequate.37 As



       37
          The EPA approved the TERP based on, among other things, its evaluation
of Texas’s revised funding projections allocating $25 million to the diesel emission
reduction program in Houston. 66 Fed. Reg. at 57,175. NRDC’s claim that EPA
should not have approved TERP because a state court invalidated one of its funding
mechanisms is without merit. At the time the EPA approved the plan, a judge had
entered a stipulated order temporarily enjoining the disputed funding provision
pending trial. The court made no findings whatsoever on the merits of plaintiffs’
claims. See H.M. Dodd Motor Co. v. Texas Dep’t of Public Safety, No. GN
102585 (200th Judicial District, Travis Co., Oct. 4, 2001). EPA was not aware of
the order when it approved TERP as part of the SIP on October 15, 2001. EPA’s
final rule acknowledged the pending state court litigation, but specifically stated that
EPA cannot “anticipate a court’s findings. If a court finds the funding mechanism
illegal, Texas will have to revise the SIP at that time to address the loss in emission
reductions or find alternative funding sources.” 66 Fed. Reg. at 57,175. Thus, EPA
did respond to NRDC’s comments questioning TERP’s funding because of state
litigation. Therefore, based on the record before the EPA at the time it approved the
SIP, EPA reasonably approved the state’s funding estimates.
        In 2002, plaintiffs in the TERP funding challenge prevailed on the merits.
The state court found that a $225 inspection fee on out-of-state vehicles violated the
Commerce Clause of the U.S. Constitution. H.M. Dodd Motor Co. v. Texas Dep’t
of Public Safety, No. GN 102585 (200th Judicial District, Travis Co., Feb. 21,
2002). EPA then issued a proposed rulemaking finding that the state had failed to
implement the SIP for the Houston Area. Finding of Failure to Implement a State
                                                                           (continued...)
                                                45
such, the EPA’s approval of TERP was in compliance with the CAA and not arbitrary and

capricious.

       NRDC also argues that the CAA requires a state or local agency to provide a step-by-step

description of implementation and an implementation schedule. The statute includes no such

requirements. The Act simply requires state and local agencies to make an early allocation of

responsibility for SIP planning and implementation and that a properly authorized agency prepares

and submits the SIP. See 42 U.S.C. § 7504(a). There is no requirement that states must provide

a step-by-step description of how specific SIP provisions will be implemented or a specific

schedule for each measure’s implementation. See 

that 42 U.S.C. §

7410(a)(2)(E) requires an implementation commitment on the part of state must also fail. This

section provides that where a local or regional entity is responsible for implementation of any

portion of a SIP, the state must ultimately ensure “adequate implementation” of that provision.

As the State of Texas is responsible for TERP, not a local or regional entity, this statute is

inapplicable.

H.     Whether EPA Reasonably Interpreted the CAA as Allowing Emission Reductions
       from the Voluntary Mobile Emission Reduction Program to be Considered in
       Reaching Attainment, and Whether the Record Supports EPA’s Approval of the
       Program in the Houston SIP




       37
         (...continued)
Implementation Plan, 67 Fed. Reg. 49,895 (EPA Aug. 1, 2002) (proposed rule). If
this rule becomes final as proposed, the state will have 18 months to provide
adequate funds for TERP or revise its SIP before sanctions are imposed. 

rule does not, however, provide grounds for invalidating its prior approval
of the SIP before the funding mechanism was invalidated.
                                                 46
        The EPA encourages states to develop voluntary measures to reduce air pollution caused

by vehicle emissions, such as trip reduction programs or growth management strategies, by

granting limited SIP credit in appropriate circumstances for Voluntary Mobile Emission

Reduction Programs or VMEP. See Guidance on Incorporating Voluntary Mobile Source

Emission Reduction Programs in State Implementation Plans at 1 (Oct. 24, 1997) (“VMEP

Guidance”). The EPA believes that such voluntary measures, which rely on the discretionary

actions of public or private parties, can provide emissions reductions that would not be available

through traditional state regulatory programs. 

5. The EPA’s VMEP Guidance provides

a framework for states to obtain credit for such emission reductions. States must identify and

describe the voluntary measures in its VMEP and include supportable projections of emissions

reductions associated with the measures. 

6–7. The state must also make an enforceable

commitment to monitor, assess, and report on the implementation and emissions effects of the

VMEPs, as well as to timely remedy any shortfall in emissions reductions that do not meet the

projected levels. Id.; see 66 Fed. Reg. at 57,190–91. The EPA requires that VMEP emissions

reductions be quantifiable, surplus (i.e., they are not credited twice), enforceable, permanent, and

adequately supported by the states. VMEP Guidance at 6–7. If these requirements are met, the

emission reductions that can reasonably be attributed to VMEPs are included in a state’s control

strategy.38




        38
         EPA’s guidance limits the SIP credit to three percent of the total future year
emissions necessary to reach attainment until EPA gains additional experience in
calculating credits for such voluntary measures. VMEP Guidance at 5.
                                                 47
       The EPA’s authority for the VMEP program is § 7410(a)(2)(A), which, as previously

discussed, allows “enforceable . . . control measures, means or techniques . . . as may be necessary

or appropriate to meet the applicable requirements of this chapter.” Due to increasing levels of

mobile emissions despite improved technology, the EPA determined that limited voluntary

measures that reduce emissions are “appropriate” measures under the CAA if enforceable against

the states as described above. VMEP Guidance at 3–4. Additional authority for EPA’s VMEP

policy can be found in 42 U.S.C. § 7408(f)(1)(B), which allows EPA to publish “information on

additional methods or strategies that will contribute to the reduction of mobile source related

pollutants during periods in which any primary ambient air quality standard will be exceeded.”

The EPA is allowed substantial discretion in its assessment of what constitutes an approvable SIP

and whether the control measures, means, or techniques contained in the SIP are necessary and

appropriate. Based on our review of the record, the EPA’s VMEP Guidance is a reasonable

exercise of that discretion consistent with the CAA.

       The Texas VMEP consists of 14 voluntary measures designed to achieve 23 tons/day of

NOx reductions. The state’s decision to use voluntary measures was necessary due to the large

magnitude of reductions needed for attainment.39 Many of the measures in the VMEP were

existing voluntary programs that the state wanted to take SIP credit for towards its attainment

demonstration. The VMEP added a few new programs that were already in development by

Houston-Galveston Area Council, the regional transportation agency.




       39
      Additionally, vehicle emissions are expected to be the largest category of
NOx emissions in 2007. Dec. 2000 Plan at 6-1.
                                                48
       NRDC argues that the EPA’s approval of Houston’s VMEP is arbitrary and capricious

because the state did not provide adequate information about resources, implementation and legal

authority for the program. Again, we agree with the EPA that the state did provide all of the

information required by the Act and that its approval of the program was consistent with the CAA

and reasonable on the record.

       The state provided supporting documentation for each voluntary measure that included a

description of the measure, the identified or predicted participants, the basis for the quantified

emission reductions, and commitments to monitor, assess, and report emission reductions for the

voluntary measures. While the state did not make a separate specific commitment to remedy any

shortfall in the VMEPs, it believed that the existing enforceable commitment regarding adoption

of additional NOx controls also applied to any shortfall in reductions from VMEPs.40 Based on

its review of each VMEP measure and the SIP’s recognition of the state’s commitment to

“monitor, evaluate, and report” on the VMEP, as well as “to remedy in a timely manner any SIP

credit shortfall if the VMEP program does not achieve projected emission reductions,” the EPA

believed that Texas had fulfilled its requirements for VMEP submissions. The EPA’s final rule



       40
          To reinforce its commitment to VMEPs, the state later submitted a SIP
revision subsequent to EPA’s approval of the SIP clarifying that it has a specific
enforceable commitment to remedy, by the attainment date, any shortfall in
reductions credited to VMEPs. EPA proposed approval of this revision on
September 18, 2002. See Proposed Approval and Promulgation of Implementation
Plans, 67 Fed. Reg. 60,633 (EPA Sept. 26, 2002) (proposed rule). After the EPA
received public comments on its proposal, it responded to comments and approved
the state’s clarification of its enforceable commitment for VMEP. See Approval
and Promulgation of Implementation Plans, 67 Fed. Reg. 68,941 (EPA Nov. 14,
2002) (final rule).
                                                  49
included a detailed discussion of the state’s commitments. See 66 Fed. Reg. at 57,190–92. The

EPA determined that Texas had made the required commitments to monitor, assess, report, and

remedy any credit shortfall from the VMEP measures in accordance with EPA guidance and that

these commitments satisfied the enforceability requirements of the CAA. 

       In summary, the EPA’s VMEP policy is a reasonable interpretation of the statute and the

EPA’s approval of the Texas VMEP is upheld as reasonable on the record. To the extent this

challenge to VMEP is based on a contention that the CAA requires a state to provide necessary

assurances as to each specific measure, mean, or technique or on a contention that the CAA

requires a state or local agency to provide a step-by-step description of implementation and an

implementation schedule, we reject those positions as discussed above.

I.     Whether EPA’s Findings on “Reasonably Available Control Measures” in the
       Houston SIP Were in Accordance with the CAA and Supported by the Record

       All nonattainment area SIPs must provide for implementation of “all reasonably available

control measures as expeditiously as practicable.” 42 U.S.C. § 7502(c)(1). The EPA interprets

this section as imposing a duty to implement only those control measures that contribute to

attainment as expeditiously as practicable. See Sierra Club v. United States EPA, 

,

743 (5th Cir. 2003). In the present case, EPA concluded that Houston’s ozone attainment

demonstration contained all reasonably available control measures (“RACM”) required by the

CAA and, therefore, EPA approved the demonstration.41 66 Fed. Reg. at 57,182.




       41
          BCCA Appeal Group, City of Houston, and Harris County, intervenors on
this issue, support EPA’s interpretation of RACM and approval of Texas’s RACM
demonstration.
                                                50
       Sierra Club challenges EPA’s approval of the RACM analysis on essentially two grounds.

First, Sierra Club argues that EPA’s interpretation of RACM is contrary to the CAA. This court,

however, recently held that EPA’s interpretation of RACM is reasonable. Sierra 

. The court found that EPA “need not require an analysis of all transport control measures”

(“TCMs”) specifically listed in 42 U.S.C. § 7408(f), and that EPA could properly conclude that

the CAA only requires transport control measures that would contribute to expeditious

attainment. 

also affirmed EPA’s interpretation that “potential measures requiring

intensive and costly implementation were not RACMs because they could not be readily

implemented due to excessive administrative burden or local conditions such as high costs.” 

Club’s challenge to EPA’s interpretation of “reasonably available control

measures” must be rejected.

       Second, Sierra Club argues that EPA’s approval of a RACM-deficient plan was arbitrary

and capricious. EPA reviewed what the state submitted and agreed that all RACMs were

included in the Houston SIP. 66 Fed. Reg. at 57,183. EPA even performed its own analysis of

transportation control measures to confirm Texas’s conclusions as to mobile measures. 66 Fed.

Reg. at 57,183. The EPA’s RACM analysis included tables listing potential transportation control

measures, associated potential emissions reductions, and evidence showing that these measures

either would not advance attainment in Houston or were too costly to be considered “reasonably

available.” See 66 Fed. Reg. at 57,184. Based on Texas’s and EPA’s analyses, EPA concluded

that the state’s submission contained sufficient control measures to meet the statutory RACM

requirement. 

                                                51
       EPA’s consideration of whether the potential RACM were economically feasible and

would advance attainment is permitted under the statute. Sierra 

. EPA

provided a detailed evaluation of the RACM issue in its approval of the SIP and appropriately

responded to numerous comments challenging Texas’s RACM in the SIP. See 66 Fed. Reg. at

57,181–90. As such, it can hardly be said that EPA’s approval of the Texas SIP’s RACM, after

its own independent analysis of the TCMs, is arbitrary and capricious.

                                        IV. CONCLUSION

       For the foregoing reasons, the court denies the petitions for review and upholds EPA’s

action approving the Houston SIP.

DENIED.




                                                52