Citizens Coal v. EPA ( 2006 )


Menu:
  •                         RECOMMENDED FOR FULL-TEXT PUBLICATION
    Pursuant to Sixth Circuit Rule 206
    File Name: 06a0162p.06
    UNITED STATES COURT OF APPEALS
    FOR THE SIXTH CIRCUIT
    _________________
    X
    -
    CITIZENS COAL COUNCIL and KENTUCKY
    Petitioners, -
    RESOURCES COUNCIL, INC.,
    -
    -
    No. 02-3628
    ,
    v.                                       >
    -
    -
    -
    UNITED STATES ENVIRONMENTAL PROTECTION
    Respondent. -
    AGENCY,
    -
    N
    On Appeal from the United States
    Environmental Protection Agency.
    No. 40 CFR Parts 9 & 434.
    Argued: June 1, 2005
    Decided and Filed: May 15, 2006
    Before: BOGGS, Chief Judge; MARTIN, SUHRHEINRICH, BATCHELDER, DAUGHTREY,
    MOORE, COLE, CLAY, GILMAN, GIBBONS, ROGERS, SUTTON, and COOK, Circuit
    Judges.
    _________________
    COUNSEL
    ARGUED: Thomas J. FitzGerald, KENTUCKY RESOURCES COUNCIL, INC., Frankfort,
    Kentucky, for Petitioners. David A. Carson, UNITED STATES DEPARTMENT OF JUSTICE,
    Denver, Colorado, for Respondent. ON BRIEF: Thomas J. FitzGerald, KENTUCKY
    RESOURCES COUNCIL, INC., Frankfort, Kentucky, for Petitioners. David A. Carson, UNITED
    STATES DEPARTMENT OF JUSTICE, Denver, Colorado, for Respondent. Dennis A. Whitaker,
    ASSISTANT COUNSEL, COMMONWEALTH OF PENNSYLVANIA, DEPARTMENT OF
    ENVIRONMENTAL PROTECTION, Harrisburg, Pennsylvania, for Amicus Curiae.
    SUHRHEINRICH, J., delivered the opinion of the court, in which BOGGS, C. J.,
    BATCHELDER, GILMAN, GIBBONS, ROGERS, SUTTON, and COOK, JJ., joined. MARTIN,
    J. (pp. 25-39), delivered a separate dissenting opinion in which DAUGHTREY, MOORE, COLE,
    and CLAY, JJ., joined.
    1
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                    Page 2
    _________________
    OPINION
    _________________
    SUHRHEINRICH, Circuit Judge. Petitioners Kentucky Resources Council, Inc. (“KRC”)
    and Citizens Coal Council (“CCC”) (collectively “Petitioners”) challenge a final rule of the
    Administrator of the United States Environmental Protection Agency (“Administrator” or “EPA”)
    promulgated under the Federal Water Pollution Control Act, commonly known as the Clean Water
    Act (“CWA” or the “Act”), 33 U.S.C. §§ 1251- 1387. The rule, found at 67 Fed. Reg. 3370 (Jan.
    23, 2002), amends the existing effluent limitations guidelines for the Coal Mining Point Source
    Category at 40 C.F.R. Part 434 by adding two new subcategories: the Coal Remining Subcategory
    and the Western Alkaline Coal Mine Subcategory (collectively the “Final Rule”). Petitioners argue
    that the Coal Remining regulations conflict with the specific language adopted by Congress in the
    Rahall Amendment, 33 U.S.C. § 1311(p), governing pollution abatement at mining sites abandoned
    before 1977 that companies want to reopen for mining. Petitioners also argue that the creation of
    the Western Alkaline Mining Subcategory violates the CWA by eliminating numeric pollution limits
    and that the EPA Administrator acted arbitrarily and capriciously in preferring best management
    practices to numeric effluent limits for sediment reduction.
    A panel of this Court invalidated the EPA’s final rule establishing effluent limitations under
    the CWA for the two subcategories. The panel unanimously rejected Petitioners’ arguments that the
    Rahall Amendment deprived the EPA of the authority to promulgate the Coal Remining regulations.
    A majority nonetheless held that the Coal Remining regulations were invalid on grounds not raised
    by Petitioners or addressed by the EPA; namely, that the EPA failed to follow procedures the
    majority deemed required by law under 33 U.S.C. § 1314. The Court granted en banc review.
    For the reasons that follow, we find that the EPA did not act contrary to law or arbitrarily
    or capriciously in promulgating regulations for the Coal Remining Subcategory.
    I. Background
    A. Regulatory Landscape
    The United States is divided into three major coal producing regions, termed the
    Appalachian, Interior, and Western regions. Historically, the Appalachian region accounted for
    approximately three-fourths of the total annual coal production. The Western Coal Region contains
    extensive deposits of sub-bituminous, low sulfur-content coal, which occurs in thick coal seams and
    shallow overburden conditions that allow for extraction at relatively low cost. Surface mine coal
    production has increased by more than ninety percent since 1970, and there have been dramatic
    changes in the domestic production of coal due to environmental concerns and market demands.
    The lack of environmental controls, until recently, has produced hundreds of thousands of acres of
    abandoned mine land. 65 Fed. Reg. 19,440, 19,444 (Apr. 11, 2000). The EPA estimates that there
    are currently over 1.1 million acres of abandoned coal mine lands in the United States, which have
    produced over 9,709 miles of streams polluted by acid mine drainage. 
    Id. Roughly ninety
    percent
    of this acid mine drainage comes from coal mines abandoned prior to the passage of federal laws
    regulating pollution and reclamation of mined lands. 
    Id. There are
    two principal pieces of federal legislation governing water pollution caused by coal
    mines. The Surface Mining Control and Reclamation Act (“SMCRA”), 30 U.S.C. §§ 1201-1328,
    was enacted on August 3, 1977, to regulate the environmental impacts of coal mining. The SMCRA
    is administered by the Office of Surface Mining Reclamation and Enforcement (“OSMRE”) within
    the Department of the Interior. Title V of the statute gives the OSMRE broad authority to regulate
    specific management practices before, during, and after mining operations. The OSMRE has
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                                  Page 3
    promulgated comprehensive regulations to control both surface coal mining and the surface effects
    of underground coal mining. See 30 C.F.R. §§ 700-955 (2005).
    The CWA was enacted in 1972 “to restore and maintain the chemical, physical, and
    biological integrity of the Nation’s waters.” 33 U.S.C. § 1251(a); see also BP Exploration & Oil,
    Inc. v. EPA, 
    66 F.3d 784
    , 789 (6th Cir. 1995). The CWA prohibits the “discharge of any pollutant”
    by “any person” into navigable waters except as authorized by the Act. 33 U.S.C. § 1311(a). The
    Act is enforced through effluent limitations guidelines and National Pollutant Discharge Elimination
    System (“NPDES”) permits that set technology-based discharge 1limits for categories and
    subcategories of water pollution point sources. See 33 U.S.C. § 1342. The EPA is charged with
    formulating and enforcing effluent limitations guidelines for classes and categories of pollutants.
    See 33 U.S.C. § 1311(b). An “effluent limitation” is “any restriction established by a State or the
    Administrator on quantities, rates, and concentrations of chemical, physical, biological, and other
    constituents which are discharged from point sources into navigable waters, the waters 2of the
    contiguous zone, or the ocean, including schedules of compliance.” 33 U.S.C. § 1362(11).
    The Act sets progressively more stringent technological standards that the EPA must use in
    setting those discharge limits. 33 U.S.C. § 1311(b)(1); 3 BP 
    Exploration, 66 F.3d at 789
    (“CWA
    directs EPA to institute progressively more stringent effluent discharge guidelines in stages.”). The
    EPA is directed to determine the degree of effluent reduction attainable for three levels of
    technology–best practicable control technology (“BPT”), best available technology economically
    feasible (“BAT”), and best conventional pollutant control technology (“BCT”). 33 U.S.C.
    § 1314(b)(1)(A), (b)(2)(A), (b)(4)(A). The EPA must also identify the factors it will consider when
    it makes these determinations. 33 U.S.C. § 1314(b)(1)(B), (b)(2)(B), (b)(4)(B).
    On April 26, 1977, the EPA issued final BPT effluent limitations guidelines for the Coal
    Mining Point Source Category. See 40 C.F.R. § 434. The regulations established limitations for the
    discharge of iron, manganese, suspended solids, settleable solids, and the pH of wastewaters for
    existing sources. 42 Fed. Reg. 21,380 (Apr. 26, 1977). In 1985, the EPA amended the effluent
    limitations guidelines and created four subcategories for the Coal Mining Point Source Category.
    They were:
    1) Coal Preparation Plants and Coal Preparation Plant Associated Areas
    2) Acid or Ferruginous Mine Drainage
    1
    The CWA also imposes water-quality based requirements. See generally Arkansas v. Oklahoma, 
    503 U.S. 91
    , 101 (1992). Water-quality based requirements are authorized under section 303 of the Act to protect the quality of
    all state waters. Water quality standards are based on the desired uses and conditions of the particular waterway
    involved, and are supplemental to technology-based requirements. See EPA v. California ex rel. State Water Res.
    Control Bd., 
    426 U.S. 200
    , 205 n.12 (1976).
    2
    A “point source” is defined as “any discernible, confined, and discrete conveyance . . . from which pollutants
    are or may be discharged.” 33 U.S.C. § 1362(14). The CWA requires the EPA to identify and categorize all point
    sources warranting effluent guidelines. See 33 U.S.C. §§ 1314(m), 1316(b)(1)(A).
    3
    Section 301 directs that point sources discharging toxic and nonconventional pollutants apply best available
    technology economically achievable (“BAT”) to meet BAT effluent limitations. 33 U.S.C. § 1311(b)(2)(A). Similarly,
    point sources discharging conventional pollutants must apply best conventional pollutant control technology (“BCT”)
    to meet BCT effluent limitations. 33 U.S.C. § 1311(b)(2)(E). New sources of pollutants must meet best available
    demonstrated control technology (“BADT”). 33 U.S.C. § 1316(a)(1); (b)(1)(B).
    The five conventional pollutants are: biochemical oxygen demand (“BOD”), total suspended solids (“TSS”),
    fecal coliform, pH, and “oil and grease.” See 33 U.S.C. § 1314(a)(4); 40 C.F.R. § 401.16. Toxic pollutants are identified
    by Congress and the EPA. 33 U.S.C. § 1311(b)(2)(C) & (D). The EPA has published a list of all such pollutants
    pursuant to section 307(a)(1) of the Act, 33 U.S.C. § 1317(a)(1). It is found at 40 C.F.R. § 401.15.
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                                 Page 4
    3) Alkaline Mine Drainage
    4) Post Mining Areas.
    50 Fed. Reg. 41,296 (Oct. 9, 1985).
    The 1985 regulations did not include a subcategory for coal mining operations that sought
    to remine previously mined, but later abandoned, lands. Thus, the regulations for mining virgin
    lands applied to previously mined, abandoned lands. The applicable effluent limitation guideline
    for remining operations in the Appalachian region was 40 C.F.R. Pt. 434, Subpt. C, Acid or
    Ferruginous Mine Drainage, and the applicable guideline for reclamation activities in the West was
    40 C.F.R. Pt. 434, Subpt. E, Post-Mining Areas. Further, for western coal mining operations,
    numeric limitations were based on the use of sedimentation pond technology that applied to all
    reclamation areas throughout4 the United States, without consideration of differences in climate,
    topography, or other factors.
    Application of these guidelines created a disincentive to potential reminers of abandoned
    mines because of the prohibitive cost of bringing pre-mined lands to the same standard as virgin
    lands, often resulting in untreated pollution for “pre-existing discharges.”
    B. The Rahall Amendment
    To counter this disincentive, Congress added as part of the 1987 amendments to the CWA
    section 301(p), known as the Rahall Amendment after its sponsor, West Virginia Rep. Nick Rahall.
    This amendment sought to provide incentives for remining and reclaiming abandoned mine lands
    that pre-dated the SMCRA in 1977 by exempting certain remining operations from effluent
    limitations, thereby making remining economically feasible. Section 301(p) provides in pertinent
    part:
    Subject to paragraphs (2) through (4) of this subsection, the Administrator, or the
    State in any case which the State has an approved permit program under section
    1342(b) of this title, may issue a permit under section 1342 of this title which
    modifies the requirements of subsection (b)(2)(A) of this section with respect to the
    pH level of any pre-existing discharge, and with respect to pre-existing discharges
    of iron and manganese from the remined area of any coal remining operation or with
    respect to the pH level or level of iron or manganese in any pre-existing discharge
    affected by the remining operation. Such modified requirements shall apply the best
    available technology economically achievable on a case-by-case basis, using best
    professional judgment, to set specific numerical effluent limitations in each permit.
    33 U.S.C. § 1311(p)(1). In other words, the Rahall Amendment allows the Administrator or a State
    with an approved NPDES permitting program to issue a discharge permit that modifies the
    categorical coal mining effluent limitations for pH, iron, and manganese with respect to any “pre-
    existing discharges” from the remined area of any remining operation or affected by the remining
    operation. Where such a modified requirement is allowed, BAT is to be determined on a case-by-
    case basis, using the permit writer’s “best professional judgment” (“BPJ”), “to set the specific
    numerical effluent limitations in each permit.” 
    Id. At the
    same time, under the Amendment, the Administrator may not issue a permit unless
    “the applicant demonstrates . . . that the coal remining operation will result in the potential for
    improved water quality from the remining operation.” 
    Id. § 1311(p)(2).
    Further, the Administrator
    4
    Sedimentation ponds in reclamation areas are designed to capture and store water from a precipitation event
    and then slowly release that water in a continuous, low-velocity discharge. 65 Fed. Reg. 19,440, 19,453 (Apr. 11, 2000).
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                              Page 5
    is prohibited from issuing a permit that allows discharges that exceed pre-existing discharge levels.
    Id.5 The Rahall Amendment also provides that no discharge “shall exceed State water quality
    standards established under section 1313 of this title.” 
    Id. The Amendment
    defines “coal remining
    operation” as one which begins after February6 4, 1987, at a site on which coal mining was conducted
    before August 3, 1977,” 
    id. § 1311(p)(3)(A)
    , and a “remined area” as “only that area of any coal
    remining operation on which coal mining was conducted before August 3, 1977,” 
    id. § 1311(p)(3)(B).
    Finally, the Amendment states that nothing in the coal remining provision was
    intended to “affect the application of the Surface Mining Control and Reclamation Act of 1977 to
    any coal remining operation, including the application of such Act to suspended solids.” 33 U.S.C.
    § 1311(p)(4).
    Congressman Rahall described the Amendment’s purpose as follows:
    Throughout the Appalachian region abandoned coal mine lands exist which, due to
    erosion and acidic discharges, pose a serious threat to water quality.
    ....
    However, in many instances, coal remining is not economically and
    technically feasible because industry becomes liable for treating the preexisting
    water discharges under stringent national effluent guidelines. This coal remining
    provision will enable industry to enter abandoned coal mine sites and engage in
    mining under modified water quality standards established on a case-by-case basis.
    The end result of this effort will be the reclamation of the site and as such, as [sic]
    improvement in water quality over that which existed at the site prior to remining.
    133 Cong. Rec. H 168 (daily ed. Jan. 8, 1987) (statement of Rep. Rahall).
    C. The Final Rule
    Despite the Rahall Amendment, coal mining companies and most states remained reluctant
    to pursue remining without formal EPA approval and guidelines. Thus, after notice and comment,
    the EPA promulgated the Final Rule on January 23, 2002.7 The Final Rule created two new
    subcategories under the Coal Mining Point Source Category and promulgated regulations for both.
    5
    “Pre-existing discharge” “means any discharge at the time of permit application under this subsection.” 33
    U.S.C. § 1311(p)(3)(C).
    6
    August 3, 1977 is the effective date of the SMCRA.
    7
    On April 11, 2000, 65 Fed. Reg. 19,440, the EPA published proposed amendments to effluent limitations
    guidelines and new source performance standards for the coal mining point source category. At that time the EPA
    proposed two new subparts to the existing regulations at 40 C.F.R. Pt. 434 applicable to Coal Remining (Subpt. G) and
    Western Alkaline Coal Mining (Subpt. H). 65 Fed. Reg. 19,440 (Apr. 11, 2000).
    On July 30, 2001, 66 Fed. Reg. 39,300, the EPA published a Notice of Data Availability to discuss options
    relating to two issues raised by commenters on the Coal Remining Subcategory that had not been presented in the
    proposal. The EPA presented these comments, data collected since the proposal, and options being considered for the
    final rulemaking in the notice. The EPA also solicited comment on (1) the expansion of applicability of the Coal
    Remining Subcategory to sites abandoned after 1977, and (2) alternative effluent limits for solids in pre-existing
    discharges. 66 Fed. Reg. 39,300 (July 30, 2001).
    The final rule was then issued on January 23, 2002, 67 Fed. Reg. 3370. 67 Fed. Reg. 3370 (January 23, 2002).
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                     Page 6
    1. Coal Remining Subcategory
    The first subcategory, known as the Coal Remining Subcategory, applies to “pre-existing
    discharges that are located within or are hydrologically connected to pollution abatement areas of
    a coal remining operation.” 40 C.F.R. § 434.71(a) (2005). “Pre-existing discharge” is defined as
    “any discharge resulting from mining activities that have been abandoned prior to the time of a
    remining permit application.” 
    Id. § 434.70(c).
    A coal remining operation is “a coal mining operation
    at a site on which coal mining was previously conducted and where the site has been abandoned or
    the performance bond has been forfeited.” 
    Id. § 434.70(a).
            The Coal Remining Subcategory requires every remining operator to create a site-specific
    Pollution Abatement Plan and submit it to the EPA or the authorized state agency. 40 C.F.R.
    § 434.72(a). The Pollution Abatement Plan must “identify characteristics of the pollution abatement
    area and the pre-existing discharges,” “be designed to reduce the pollution load from pre-existing
    discharges,” and “identify the selected best management practices (BMPs) to be used.” 
    Id. The regulation
    further requires the plan to “describe the design specifications, construction
    specifications, maintenance schedules, criteria for monitoring and inspection, and expected
    performance of the BMPs.” 
    Id. The Coal
    Remining Subcategory sets the effluent limitations for four pollutants: total iron,
    total manganese, net acidity, and TSS. 40 C.F.R. § 434.72(b)(1). The effluent limitation for each
    pollutant “[m]ay not exceed baseline loadings,” or conditions that exist when remining commences.
    
    Id. However, if
    the EPA or state permitting authority finds that measuring baseline pollutant
    loadings is infeasible, and that remining “will result in significant improvement that would not
    otherwise occur,” then no measurable effluent limitations apply. 
    Id. § 434.72(b)(2).
          Under the Final Rule, these Pollution Abatement Plans qualify as BPT, 40 C.F.R. § 434.72,
    and BAT, 40 C.F.R. § 434.73, and BCT, 40 C.F.R. § 434.74, for remining operators.
    Thus, the Coal Remining Subcategory differs from the Rahall Amendment in four ways.
    First, the definition of “coal remining operation” is broader than the definition found in the Rahall
    Amendment, which limits sites eligible for the modified effluent limits to those sites on which the
    coal remining operation began after February 4, 1987, on a site where coal mining had been
    conducted before August 3, 1977, the effective date of the SMCRA. The Final Rule allows any coal
    mining operation at a site on which coal mining was previously conducted and where the site has
    been abandoned or the performance bond has been forfeited, to be eligible regardless of whether
    prior mining occurred before August 3, 1977 or thereafter. Second, although the Rahall Amendment
    lists only three pollutants for which modified requirements are available (pH, iron, and manganese),
    the Final Rule added suspended solids to the list of requirements that could be modified. Third, the
    Rahall Amendment bounded the modified requirements by requiring that where the categorical
    effluent limitations were not required, specific numerical effluent limitations representing a case-by-
    case determination of the BAT would be required in each permit. On the other hand, the Final Rule
    allows modified permits without setting numeric limitations representing case-by-case BAT.
    Finally, the Final Rule differs from the Rahall Amendment by applying the more stringent general
    standard instead of the modified requirements to pre-existing discharges that reminers commingle
    with discharges from active mining operations.
    2. Western Alkaline Coal Mining Subcategory
    The second new subcategory, the Western Alkaline Coal Mining Subcategory, applies to
    “mine drainage from applicable areas of western coal mining operations.” 40 C.F.R. § 434.82(a).
    A “western coal mining operation” is “a surface or underground coal mining operation located in
    No. 02-3628                Citizens Coal Council, et al. v. EPA                                                    Page 7
    the interior western United States, west of the 100th meridian west longitude, in an arid or semiarid
    environment with an average annual precipitation of 26.0 inches or less.” 40 C.F.R. § 434.80(f).
    The Final Rule requires the operator to submit to the permitting authority “a site-specific
    Sediment Control Plan . . . designed to prevent an increase in the average annual sediment yield
    from pre-mined undisturbed conditions.” 40 C.F.R. § 434.82(a). Similar to the Pollution Abatement
    Plans, each Sediment Control Plan “must identify best management practices (BMPs) and also must
    describe design specifications, construction specifications, maintenance schedules, criteria for
    inspection, as well as expected performance and longevity of the best management practices.” 
    Id. The Western
    Alkaline Coal Mining Subcategory does not set any uniform standards for drainage at
    western coal mining operations; it merely requires a Sediment Control Plan that is incorporated into
    the NPDES permit. However, operators must demonstrate “[u]sing watershed models,” that the
    Sediment Control Plans “will result in average annual sediment yields that will not be greater than
    the sediment yield levels from pre-mined, undisturbed conditions.” 40 C.F.R. § 434.82(b). The
    operator’s only other obligation is to implement and maintain the best management practices
    described in the Sediment Control Plan. 
    Id. § 434.82(c).
    These include BPT, 
    id., § 434.82;
    BAT,
    
    id. § 434.83;
    and BCT, 
    id. § 434.84.
            The Western Alkaline Coal Mining Subcategory differs from the Rahall Amendment in that
    the effluent limitations on the concentration of settleable solids and pH have been eliminated, and
    replaced only with a requirement that the operator submit a site-specific Sediment Control Plan that
    is designed to prevent an increase in the “average annual sediment yield from pre-mined,
    undisturbed conditions” using “best management practices.”
    D. Litigation
    Petitioners brought this appeal under section 509(b) of the CWA, challenging the Final
    Rule.8 Petitioners claimed that the remining guidelines exceeded EPA’s statutory authority under
    the Rahall Amendment by expanding both the number of sites and the number of pollutants eligible
    for modified pollutant limits, and that the new rule impermissibly discourages remining at Rahall-
    eligible mines by subjecting commingled waste streams to the most stringent effluent limitation
    applicable to the constituent waste streams. Regarding the Western Alkaline Coal Mining
    Subcategory, Petitioners first argued that BMPs are not appropriate effluent limitations because they
    are not numeric criteria. Second, they claimed that the Administrator has not shown that continued
    imposition of the currently applicable effluent limitations is infeasible. Third, they argued that the
    entire subcategory of Western Alkaline Coal Mining is irrational. Finally, Petitioners contended that
    the EPA’s goal in creating the subcategory is inconsistent with the CWA and the SMCRA.
    The EPA defended the Final Rule on the grounds that the Rahall Amendment does not limit
    its authority to issue the regulations. The EPA argued that the requirement of “best management
    practices” in lieu of numeric effluent limitations was a reasonable exercise of its authority. As to
    the Western Alkaline Coal Mining Subcategory, the EPA argued that numeric limitations were
    infeasible. The EPA also maintained that the creation of the subcategory was not arbitrary or
    capricious because it had relied on scientific information to support its creation. Finally, the agency
    asserted that setting effluent limitations at premined background levels was consistent with the CWA
    and the SMCRA.
    8
    This Court has jurisdiction over this appeal pursuant to section 509(b) of the CWA, 33 U.S.C. 1369(b). The
    petition for review is proper in this Circuit pursuant to 33 U.S.C. § 1369(b)(1), which directs that review may be had by
    application to the appeals court for the federal judicial district in which a person directly affected by the action resides
    or transacts business. The Western Alkaline Coal Mining Subcategory applies to reclaimed mining sites in several
    western states, none in this Circuit. Petitioners state that some of the members of the CCC reside in this Circuit, and that
    some members are affected by the Western Subcategory rule.
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                                   Page 8
    E. Panel Decision
    A panel of this Court unanimously rejected Petitioners’ arguments that the Rahall
    Amendment deprived the EPA of authority to promulgate the Coal Remining regulations. See
    Citizens Coal Council v. EPA, 
    385 F.3d 969
    (6th Cir. 2004), vacated, reh’g en banc granted,
    February 23, 2005. It reviewed the EPA’s construction of the statute under Chevron, U.S.A. v.
    Natural Res. Def. Council., Inc., 
    467 U.S. 837
    (1984), and concluded that neither the Rahall
    Amendment nor the CWA itself addressed whether the EPA could create additional subcategories.
    The unanimous panel further held that the Amendment was simply an opt-out provision, and that
    it did not address the EPA’s authority to promulgate regulations that are broader than its provisions.
    The panel therefore held that the EPA’s interpretation that 33 U.S.C. § 1314(b) grants it independent
    authority to issue the regulations was reasonable under Chevron.
    The majority nonetheless found the regulations invalid on grounds not raised or presented
    by either party. First, the majority found the regulations invalid because the EPA failed to follow
    procedures the majority deemed required by law. Specifically, the majority found that, as the very
    first step in the development of any effluent limitations guidelines under 33 U.S.C. § 1314, the EPA
    must “identify which control measures and practices are available to the various categories and
    classes of point sources” under 33 U.S.C. § 1314(b)(3). The majority found that the EPA had
    impermissibly skipped this mandatory step for the Coal Remining regulations because the EPA first
    defined the desired level of pollution (baseline conditions) and then worked backwards to find
    technological tools (site-specific plans) that would achieve this goal. The majority further reasoned
    that the EPA’s approach shirked its obligation under 33 U.S.C. § 1314(b)(1)(A), (b)(2)(A) and
    (b)(4)(A) to determine the degree of effluent reduction attainable. The majority faulted the EPA for
    setting the required reduction at zero (the baseline) without determining if a greater reduction was
    possible. The court held that “where the EPA fails to determine how much reduction in pollution
    is possible, and to do so by reference to the amounts of pollutants,” it deviates from the statutory
    directive. The majority also faulted the Coal Remining regulations because the regulations
    impermissibly set effluent limitations through site-specific factors. Finally, the majority invalidated
    the Coal Remining regulations because the record contained no evidence that the EPA considered
    each of the factors enumerated in 33 U.S.C. § 1314(b)(1)(B), (b)(2)(B), and (b)(4)(B).
    The panel majority invalidated the Western Alkaline Coal Mining Subcategory on the same
    bases.
    One member dissented from those portions of the majority opinion that found the EPA’s
    regulations invalid, and would have upheld the regulations in their entirety.
    The EPA sought en banc review, arguing that the court should not have reached the issues
    upon which the panel majority ruled against the EPA, because those issues were neither raised to
    the EPA in the underlying notice-and-comment rulemaking proceeding, nor in Petitioners’ initial
    brief. The EPA also contended that the panel erred by not deferring to its reasonable construction
    of ambiguous provisions of a statute it administers. The EPA therefore asked that this Court vacate
    those portions of the panel majority’s decision in which it found the regulations invalid. 9
    9
    The Interstate Mining Compact Commission (“IMCC”) has filed a brief as amicus curiae in support of the EPA
    in the en banc proceedings. The IMCC is an agency of its party states created to assist those states in the establishment,
    implementation, and administration of regulatory programs governing coal mining operations and related environmental
    issues within their respective borders.
    The IMCC consists of its member states, which as of 2005 accounted for 70 percent of the total national coal
    production.
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                                  Page 9
    II. Standards of Review
    A. Statutory Interpretation
    In Chevron, the Supreme Court articulated a two-step test for judicial review of an agency’s
    interpretation of a statute. First, the reviewing court must ask “whether Congress has directly
    spoken to the precise question at issue.” 
    Id. at 842.
    If so, “that is the end of the matter; for the court,
    as well as the agency, must give effect to the unambiguously expressed intent of Congress.” 
    Id. at 842-43
    (footnote omitted). If Congress has not directly spoken to the issue at hand, however, the
    court must decide if the agency’s action under the statute is based on a permissible construction of
    the statute. 
    Id. at 843.
    In this second step, the court must accord considerable weight to the
    agency’s construction of the statute and it may not substitute its own construction of the statute for
    the agency’s reasonable interpretation. 
    Id. at 843-44.
                                                  B. Agency Discretion
    The EPA promulgated the Final Rule through informal rulemaking. The scope of our review
    over the informal rulemaking process is generally governed by section 10(2)(e) of the
    Administrative Procedure Act (“APA”), 5 U.S.C. § 706(2) (1996). BP 
    Exploration, 66 F.3d at 792
    .
    Under § 706(2), a court must invalidate an informal rule if it is “arbitrary, capricious, an abuse of
    discretion, or otherwise not in accordance with law”; “in excess of statutory jurisdiction, authority,
    or limitations, or short of statutory right”; or “without observance of procedure required by law.”
    5 U.S.C. § 706(2)(A), (C), (D).10
    An agency rule is “arbitrary or capricious”
    if the agency has relied on factors which Congress has not intended it to consider,
    entirely failed to consider an important aspect of the problem, offered an explanation
    for its decision that runs counter to the evidence before the agency, or is so
    implausible that it could not be ascribed to a difference in view or the product of
    agency expertise.
    Motor Vehicle Ass’n v. State Farm Mut. Auto. Ins. Co., 
    463 U.S. 29
    , 43 (1983). The court is
    required to make a “searching and careful review” in its assessment of the agency action, but “the
    ultimate standard of review is a narrow one.” Citizens to Preserve Overton Park, Inc. v. Volpe, 
    401 U.S. 402
    , 416 (1971). Where the rulemaking involves review of the agency’s technical or scientific
    evaluations and determinations, the highest level of deference to the agency is to be applied.
    Baltimore Gas & Elec. v. Natural Res. Def. Council, Inc., 
    462 U.S. 87
    , 103 (1983); BP 
    Exploration, 66 F.3d at 792
    .
    III. Analysis
    A. Coal Remining Subcategory
    1. Chevron Analysis
    In their original merits brief, Petitioners contended that, in creating the Final Rule, the
    Administrator violated step one of Chevron by ignoring the language and intent of section 301(p)
    10
    We recognize that there is support for the proposition that in review of rulemaking the second step of Chevron
    indeed amounts to the same inquiry as arbitrary or capricious review under the APA. See United States v. Mead Corp.,
    
    533 U.S. 218
    , 229 (2001); Nat’l Ass’n of Reg. Util. Comm’rs v. ICC, 
    41 F.3d 721
    , 726-27 (D.C. Cir. 1994); Ronald M.
    Levin, see generally The Anatomy of Chevron: Step Two Reconsidered, 72 Chi.-Kent L. Rev. 1253 (1997).
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                               Page 10
    of the Rahall Amendment. They claim that the Rahall Amendment precisely defined the limited
    circumstances in which a modified permit may be issued in lieu of the otherwise-applicable
    categorical effluent limitations for coal mining operations, and that the Final Rule exceeds these
    boundaries in numerous respects. Specifically, Petitioners contended that Congress bounded the
    EPA’s authority by specifying which pollutants (pH, iron, and manganese) are eligible for relaxed
    standards, which operations (pre-August 3, 1977) are eligible, what limits must apply in place of
    national guideline limits (i.e. specific numerical limits), and when the alternative standards would
    apply.
    Thus, the initial question under step one of Chevron is whether the Rahall Amendment itself
    or the CWA otherwise explicitly addresses the Administrator’s authority to create a subcategory
    under the Coal Mining Point Source Category and to promulgate regulations for that subcategory
    that conflict with the Rahall Amendment. The Rahall Amendment states that the Administrator
    “may issue a permit under section 1342 of this title which modifies the requirements of subsection
    (b)(2)(A) of this section” with respect to certain effluents. 33 U.S.C.A. § 1311(p) (emphasis added).
    Section 301(b)(2)(A) in turn provides for “effluent limitations for categories and classes of point
    sources . . . which will result in reasonable further progress toward the national goal of eliminating
    the discharge of all pollutants, as determined in accordance with regulations issued by the
    Administrator.” 
    Id. § 1311(b)(2)(A).
    The Rahall Amendment, therefore, provides for modifications
    of the effluent limitations guidelines determined by the Administrator for categories and classes of
    point sources pursuant to the requirements of sections 301(b) and 304(b). The Amendment is
    permissive; it does not require remining operations to seek such modified permits, and remining
    operators remain free to seek permits under the generally applicable effluent limitations guidelines
    promulgated by the EPA pursuant to section 304(b). The Rahall Amendment says nothing about
    the EPA’s ability or inability to regulate national effluent limitations. It simply provides express
    authority to modify individual NPDES permits on a case-by-case basis to allow effluent limitations
    for specified pollutants in the pre-existing discharges from remining operations other than those
    required  in the generally applicable effluent limitation regulations found in 40 C.F.R. § 434, Subpt.
    C.11
    Sections 301(b) and 304(b) also do not provide a definitive answer. Section 304(b) states
    that the “Administrator shall . . . publish . . . regulations . . . providing guidelines for effluent
    limitations, and, at least annually thereafter, revise, if appropriate, such regulations” for various
    classes and categories of point sources. 33 U.S.C.A. § 1314(b) (emphasis added). Section 301gives
    these guidelines the force of law, 
    id. at §
    1311(a) (“Except as in compliance with this section . . . the
    discharge of any pollutant by any person shall be unlawful”); § 1311(e) (“Effluent limitations
    established pursuant to this section or section 1312 of this title shall be applied to all point sources
    of discharge of pollutants in accordance with the provisions of this chapter.”), and section 402(a)
    (33 U.S.C. § 1342) directs the Administrator to incorporate these regulations into NPDES permits
    issued to individual polluters. Neither section 301(b) nor 304(b) contains any restriction vis-a-vis
    section 301(p) on the Administrator’s power to create additional subcategories for the Coal Mining
    Point Source Category. In short, these sections do not express a clear congressional intent on the
    issue of whether the EPA has the power to create a subcategory that differs from the Rahall
    Amendment.
    11
    The EPA also points out that section 301(p)(1) uses the term “best professional judgment” when describing
    the modified permit requirements, which is utilized for effluent limitations in permits for which no national standard
    applies. “Best professional judgment” refers to NPDES authority under section 402(a)(1)(B) of the Act, 33 U.S.C.
    § 1342(a)(1)(B), by application of the regulations at 40 C.F.R. § 125.3(c) and (d). This permit authority applies where
    the EPA has not promulgated an applicable guideline. Natural Res. Def. Council v. EPA, 
    863 F.2d 1420
    , 1424 (9th Cir.
    1988).
    No. 02-3628                Citizens Coal Council, et al. v. EPA                                           Page 11
    Petitioners assert that the EPA cannot rely on its general rulemaking authority when specific
    statutory directives define how that standard-setting authority is to be used in a particular instance.
    See, e.g., Am. Petroleum Inst. v. EPA, 
    52 F.3d 1113
    , 1119 (D.C. Cir. 1995) (stating that the “EPA
    cannot rely on its general authority to make rules necessary to carry out its functions when a specific
    statutory directive defines the relevant functions of EPA in a particular area”) (citations omitted).
    That is, they claim that the Rahall Amendment’s directives as to which pollutants, which operations,
    what limits, and when alternative standards apply preclude the EPA from adopting standards that
    are inconsistent with these limitations. However, because the Rahall Amendment does not contain
    a specific statutory directive to limit or prevent the EPA from otherwise exercising its statutory
    authority to establish national standards for the remining subcategory, this argument is unavailing.12
    Because Congress has not directly addressed the precise question at issue, we move to step
    two of the Chevron analysis. 
    Chevron, 467 U.S. at 843
    . The EPA claims that it was authorized to
    promulgate regulations for the coal remining subcategory pursuant to its general rulemaking
    authority under section 304(b), and that nothing in section 301(p) limits that authority, because the
    Rahall Amendment is simply a permissive, opt-out, rather than a generally applicable, scheme.
    As noted, section 304(b) authorizes, indeed requires, the Administrator to create effluent
    limitation guidelines for various classes and categories of point sources. When considered with the
    permissive, opt-out nature of the Rahall Amendment, applicable in individual permits, we cannot
    say that the Administrator’s conclusion that it was authorized to create additional subcategories
    under the Coal Mining Point Source Category, not identical to the Rahall Amendment, was
    unreasonable.
    We also find that the EPA did not act arbitrarily or capriciously in promulgating the Final
    Rule. The record reflects that, despite the Rahall Amendment, remining operations were not
    occurring. The Final Rule was issued fifteen years after the Rahall Amendment. In the first ten
    years after its enactment, only 330 (out of 1072) Rahall-type permits had been issued, with 300 of
    those being issued by Pennsylvania.13 Many states had not been able to establish the guidelines and
    procedures required to issue Rahall permits. At the same time, “IMCC member States [had]
    indicated that they would be able to establish formal remining programs under guidelines set forth
    under an EPA effluent limitations Coal Remining subcategory,” and that with such guidelines, “mine
    operators would be more inclined to enter into remining projects.” 65 Fed. Reg. at 19,445. The
    EPA promulgated the rule in the hopes of providing “better environmental results than the [then-]
    current requirements.” 
    Id. at 19,446.
    Thus, the EPA had a reasonable basis for its action–namely
    evidence that the Rahall Amendment was, as a practical matter, not achieving its goal of
    encouraging remining and, on a larger scale, of restoring the “chemical, physical, and biological
    integrity of the Nation’s waters.” 33 U.S.C. § 1251(a).
    The EPA’s conclusion that remining has many potential benefits is also supported in the
    record. 
    Id. at 3375.14
    Performing statistical analyses to evaluate the effects of remining on water
    12
    The EPA also asserts that it enacted the Final Rule in part to implement the Rahall Amendment. This
    argument must be rejected. The Rahall Amendment authorized the EPA to modify the effluent limitation guidelines in
    individual permits. It does not authorize the EPA to issue generally applicable rules for coal remining.
    13
    Pennsylvania had amended its mining statute to provide for remining incentives three years before Congress
    amended the CWA. See 52 Pa. Stat. Ann. § 1396.4f (West 1998).
    14
    As the Final Rule notes, the regulations are supported by several key documents:
    1. “Coal Remining Best Management Practices Guidance Manual” (EPA 821-B-01-010): This
    document describes abandoned mine land conditions and the performance of Best Management
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                                Page 12
    quality at thirteen remining sites, the EPA analyzed the adverse environmental impacts of current
    practices as a basis for assessing the incremental environmental impacts of the proposed rule. 
    Id. at 3389;
    Kristen L. Strellec, Office of Water, Office of Sci. and Tech., Eng’g and Analysis Div.,
    Envtl. Prot. Agency, EPA 821-B-00-002, Effluent Limitations Guidelines and Standards for the Coal
    Mining Industry: Remining and Western Alkaline Subcategories 8-1(Mar. 2000), available at
    http://www.epa.gov/guide/coal/impact.pdf. The EPA observed that abandoned mine lands are
    associated with a wide range of public health and safety problems and aesthetic degradation,
    including abandoned mine openings, highwalls, unstable spoil piles, and hazardous water bodies.
    In addition, acid mine drainage from abandoned mine lands causes serious water quality problems
    because of its high acidity and other contaminants. 
    Id. Remining improved
    water quality through
    the removal of acid-forming materials and sediment, improved public safety by correcting problems
    associated with abandoned mines, and improved the aesthetic quality of the land and water
    resources. 
    Id. at 8-1
    to 8-3.
    In sum, the decision to establish national standards for the remining subcategory was not
    contrary to the plain language of section 301(p), was within the Administrator’s broad discretion
    under sections 301(b) and 304(b), is supported by the record, and was not unreasonable.
    2. Specific Challenges
    Having rejected the argument that the EPA lacked authority to promulgate the Final Rule,
    and further concluded that the EPA’s interpretation of the Act as allowing it to create subcategories
    under the Coal Mining regulation was reasonable,       we now address Petitioners’ challenges to
    specific aspects of the remining subcategory rule.15
    Practices (BMPs) that have been implemented at remining operations. The BMP Guidance Manual
    is a technical reference document that presents research and data concerning the prediction and
    prevention of acid mine drainage to the waters of the United States. . . .
    2. “Coal Remining Statistical Support Document” (EPA 821-B-01-011): This document describes
    the statistical methodology for establishing and monitoring baseline conditions and setting discharge
    limits at remining sites.
    3. “Development Document for Final Effluent Limitations Guidelines and Standards for the Western
    Alkaline Coal Mining Subcategory” (EPA 821-B-01-012): This document presents EPA’s technical
    conclusions concerning the Western Alkaline Coal Mining Subcategory.
    4. “Economic and Environmental Impact Assessment of Effluent Limitations Guidelines and
    Standards for the Coal Mining Industry: Remining and Western Alkaline Subcategories” (EPA 821-
    B-01-013): This document presents the methodology employed to assess economic and environmental
    impacts of the final rule and the results of the analysis.
    5. “Statistical Analysis of Abandoned Mine Drainage in the Assessment of Pollution Load” (EPA
    821-B-01-014): This document describes pollutant characteristics of pre-existing discharges at
    abandoned mine lands.
    67 Fed. Reg. at 3371.
    15
    It is not clear to what extent Petitioners sought review of these issues under step two of Chevron as to the
    Coal Remining Subcategory. Thus, our conclusion that the Rahall Amendment did not limit the EPA’s authority to issue
    the subcategories should technically end the matter. However, Petitioners’ specific challenges appear to implicate not
    only this type of review, but review under the APA as well.
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                                 Page 13
    a. Definition of “Coal Remining Operation”
    Petitioners complain that the EPA exceeded its authority in defining “coal remining
    operation” to include remining operations on mine lands abandoned after August 3, 1977. For the
    reasons discussed above, the EPA was not constrained to 16use a definition identical to the definition
    of “coal remining operation” in the Rahall Amendment. Furthermore, as explained above, the
    EPA’s reading of the Act as giving it authority under section 304(b), together with section 301(b),
    is not unreasonable.
    We also find that the EPA did not act arbitrarily or capriciously in deciding to extend the
    applicability of the coal mining subcategory effluent limitations guidelines beyond those remining
    operations that had been previously eligible for modified permits under section 301(p). The record
    reflects that the EPA examined whether the broader definition would cause an increase in abandoned
    mines in order to take advantage of the remining effluent limitation guidelines for pre-existing
    discharges at a future remining operation. 67 Fed. Reg. at 3374. However, the EPA concluded that
    there are strong disincentives under the SMCRA against mine operation abandonments. 
    Id. A mine
    operator not only forfeits the reclamation bond, but is also precluded from receiving further mining
    permits until reclamation is completed.17 Thus, it cannot be said that the EPA acted arbitrarily or
    capriciously in adopting a broader definition of “coal remining operation” than that found in the
    Rahall Amendment.
    b. TSS
    Petitioners contend that section 301(p) does not allow modified NPDES permit limits for
    TSS. In support, they cite section 301(p)(4), which states that “[n]othing in this subsection shall
    affect the application of the Surface Mining Control and Reclamation Act of 1977 . . . to any coal
    remining operation, including the application of such Act to suspended solids.” 33 U.S.C.
    § 1311(p)(4).
    Again, nothing in the Rahall Amendment limits or even addresses the EPA’s discretion to
    promulgate national standards. The quoted language refers to the SMCRA only, and Petitioners
    offer no explanation as to how including effluent limitations for suspended solids affects the
    application of SMCRA requirements for suspended solids. Further, the record reflects that the EPA
    considered whether the alternative limits for solids would square with the SMCRA requirements for
    suspended solids, and concluded that “the final regulation is consistent with SMCRA which
    mandates the prevention of additional contribution of suspended solids to streamflow to the extent
    possible using the best technology currently available.” 67 Fed. Reg. at 3388. The EPA explained
    that it
    has adopted what is essentially a compliance schedule so that, during remining and
    reclamation activities, the operator cannot contribute sediment levels beyond the
    baseline discharge loading. After remining and reclamation has been completed, the
    operator must meet the standards for TSS and SS contained in subpart E–Post
    16
    In fact, section 301(p)(3) limits the reach of the definitions to the “purposes of this subsection.” 33 U.S.C.
    § 1311(p)(3).
    17
    The SMRCA regulates the surface mining industry by requiring surface coal mining operators to obtain
    mining permits. 30 U.S.C.A. § 1257 (West 1986 & Supp. 2005). The mining permit must be denied unless the permit
    application contains a reclamation plan, 
    id. at §
    1257(d), which guarantees that reclamation can be achieved, 
    id. at §
    1260. The permitting process also requires the permitting authority to postpone issuance of a mining permit until the
    applicant posts a performance bond. 
    Id. at §
    1259(e). The amount of the performance bond must be “sufficient to assure
    the completion of the reclamation plan if the work is to be performed by the regulatory authority in the event of
    forfeiture.” 
    Id. at §
    1259(a).
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                     Page 14
    Mining areas prior to bond release. EPA concluded that the implementation of
    successful sediment control BMPs should, in most cases, be able to meet the BPT
    standards contained in subpart E–Post Mining areas regardless of whether the area
    has been disturbed due to remining or virgin mining.
    
    Id. Nor is
    including effluent limitations for solids inconsistent with the purpose of section
    301(p). The record reflects that the EPA initially did not propose alternative limits for solids, but
    based on comments and information received on the proposed rule, the EPA decided that alternative
    limits for pH, iron, and manganese, but not for solids, would not provide sufficient incentives for
    remining abandoned mine lands. The EPA noted that the alternative limits for suspended solids were
    consistent with the intent of the Rahall Amendment, “which seeks to encourage remining while
    ensuring that the remining will potentially improve and reclaim [abandoned mine lands].” 
    Id. Thus, the
    EPA’s inclusion of TSS in the coal remining subcategory was not arbitrary or capricious.
    c. BMPs
    Petitioners also argue that the EPA lacks authority to promulgate BMPs in place of numerical
    standards as effluent limitations, because section 301(p) provides that modified permits under the
    Rahall Amendment apply BAT on a case-by-case basis “to set specific numerical effluent limitations
    in each permit.” 33 U.S.C. § 1311(p)(1) (emphasis added). Again, for the reasons articulated above,
    section 301(p) does not limit the EPA’s authority to use BMPs.
    The EPA contends that its interpretation was reasonable because the CWA does not mandate
    the use of numeric limitations only. The EPA therefore claims that it reasonably determined that
    effluent limitations for iron, manganese, acidity, and/or TSS from pre-existing discharges from coal
    remining operations may take the form of BMPs when a baseline pollutant loading cannot be
    calculated.
    The EPA relies on the definition of “effluent limitation,” which is defined under the Act as
    any restriction established by a State or the Administrator on quantities, rates, and
    concentrations of chemical, physical, biological, and other constituents which are
    discharged from point sources into navigable waters, the waters of the contiguous
    zone, or the ocean, including schedules of compliance.
    33 U.S.C. § 1362(11) (emphasis added); see also 33 U.S.C. § 1312(a) (stating that “[w]henever, in
    the judgment of the Administrator . . . discharges of pollutants from a point source or group of point
    sources, with the application of effluent limitations required under section 1311(b)(2) of this title,
    would interfere with the attainment or maintenance of that water quality in a specific portion of
    navigable waters . . . effluent limitations (including alternative effluent control strategies) for such
    point sources or sources shall be established which can reasonably be expected to contribute to the
    attainment or maintenance of such water quality”) (emphasis added). Thus, under the Act, effluent
    limitations are not limited to numeric discharges but encompass “any restriction” on discharges,
    including “schedules of compliance,” or “a schedule of remedial measures including an enforceable
    sequence of actions or operations leading to compliance with an effluent limitation.” 33 U.S.C.
    § 1362(17).
    The EPA’s reading of the Act finds support in the language of the Act and is therefore not
    unreasonable. Other courts have also recognized that the CWA does not require the EPA to set
    numeric limits where such limits are infeasible. See, e.g., Waterkeeper Alliance, Inc. v. EPA, 
    399 F.3d 486
    , 502 (2d Cir. 2005) (stating that site-specific BMPs are effluent limitations under the
    CWA; holding that the EPA provision regulating the emission of water pollutants by concentrated
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                                  Page 15
    animal feeding operations was arbitrary and capricious because it allowed permitting authorities to
    issue permits that did not include terms of nutrient management plans, in violation of CWA
    requirement that effluent limitations be included in NPDES permits); Natural Res. Def. Council, Inc.
    v. EPA, 
    673 F.2d 400
    , 403 (D.C. Cir. 1982) (noting that “section 502(11) defines ‘effluent
    limitation’ as ‘any restriction’ on the amounts of pollutants discharged, not just a numerical
    restriction”; holding that section of CWA authorizing courts of appeals to review promulgation of
    “any effluent limitation or other limitation” did not confine the court’s review to the EPA’s
    establishment of numerical limitations on pollutant discharges, but instead authorized review of
    other limitations under the definition) (emphasis added). In Natural Res. Def. Council, Inc. v.
    Costle, 
    568 F.2d 1369
    (D.C. Cir. 1977), the D.C. Circuit stressed that
    when numerical effluent limitations are infeasible, EPA may issue permits with
    conditions designed to reduce the level of effluent discharges to acceptable levels.
    This may well mean opting for a gross reduction in pollutant discharge rather than
    the fine-tuning suggested by numerical limitations.
    
    Id. at 1380
    (rejecting the EPA’s argument that where it was infeasible to promulgate numeric
    effluent limitations, the EPA was authorized to exclude the relevant point source from the NPDES
    program, because section 402(a) gives the EPA considerable flexibility in framing the permit to
    achieve a desired reduction in pollutant discharges); see also 
    id. at 1380
    n.21 (noting the proposition
    that Congress did not regard numeric effluent limitations as 18  the only permissible limitation was
    supported by section 302(a) of the Act, 33 U.S.C. § 1312(a)).
    Petitioners contend that the suggestion that best management practices are the equivalent of
    effluent limitations “flies in the face of the distinction drawn by Congress in Section 304(e) and
    304(b).” Section 304(e) authorizes BMPs as “supplemental” to effluent limitations. See 33 U.S.C.
    § 1314(e) (stating that the EPA “may publish regulations, supplemental to any effluent limitations
    specified under subsections (b) and (c) of this section for a class or category of point sources, for
    any specific pollutant which the Administrator is charged with a duty to regulate as a toxic or
    hazardous pollutant”). Petitioners therefore claim that BMPs alone may not constitute effluent
    limitations.
    The EPA counters that section 304(e) applies specifically to toxic pollutants, and maintains
    that it derives independent authority under sections 402(a) and 304(b), 33 U.S.C. §§ 1342(a),
    1314(b), and section 502(11), 33 U.S.C. § 1362(11) to incorporate BMPs in the effluent guidelines
    for other pollutants. Thus, the EPA maintains that sections 502(11), 304(b), and 304(e) can be read
    together as allowing BMPs as supplemental to an effluent limitation under § 304(b), or can each
    stand as an effluent limitation by itself. This is certainly not an unreasonable interpretation of the
    Act, and given the ambiguity in the CWA on this point, we will uphold the EPA’s interpretation as
    reasonable.
    In any event, the effluent limitations guidelines are consistent with section 301(p)(1) in that
    they require numerical limitations to be included in the NPDES permits, unless calculation of such
    numeric limits is infeasible. See 40 C.F.R. § 434.72(b)(1) (“Except as provided in . . . paragraph
    (b)(2) of this section, the following effluent limits apply to pre-existing discharges . . . .”). Further,
    18
    The EPA’s NPDES permit regulations reflect the EPA’s longstanding interpretation of the CWA as allowing
    BMPs to take the place of numeric effluent limitations under certain circumstances. 40 C.F.R. § 122.44(k), entitled
    “Establishing limitations, standards, and other permit conditions (applicable to State NPDES programs . . . ),” provides
    that permits may include BMPs to control or abate the discharge of pollutants when: (1) “[a]uthorized under section
    304(e) of the CWA for the control of toxic pollutants and hazardous substances from ancillary industrial activities”;
    (2) “[a]uthorized under section 402(p) of the CWA for the control of storm water discharges”; (3) “[n]umeric effluent
    limitations are infeasible”; or (4) “[t]he practices are reasonably necessary to achieve effluent limitations and standards
    or to carry out the purposes and intent of the CWA.” 40 C.F.R. § 122.44(k).
    No. 02-3628                 Citizens Coal Council, et al. v. EPA                                              Page 16
    the required numerical limitations are consistent with section 301(p)(2), which states that limitations
    for pH, iron, and manganese cannot allow discharge of such pollutants “to exceed the levels being
    discharged from the remined area before the coal remining operation begins.” Under the coal
    remining subcategory, the EPA interpreted the Rahall Amendment’s requirement            not to exceed
    baseline “levels” as a requirement not to exceed a pollutant baseline “loading.”19 67 Fed. Reg. at
    3384. Thus, determining specific effluent limitations requires calculating baseline pollutant loadings
    for each pollutant in the pre-existing discharge, using a standardized procedure for the calculation.
    See 40 C.F.R. § 434, app. B (“Baseline Determination and Compliance Monitoring for Pre-Existing
    Discharges at Remining Operations”). The calculation becomes the numerical limitation for the
    pollutant. Non-numeric limits are allowed only when numeric limits are infeasible. Finally, the
    remining subcategory rule requires non-numeric effluent limitations consisting of a pollution
    abatement plan, which “must be designed to reduce the pollution load from pre-existing discharges
    and must identify the selected best management practices (BMPs) to be used.” 40 C.F.R.
    § 434.72(a). This requirement is consistent with section 301(p), which requires the coal remining
    operator to demonstrate “that the coal remining operation will result in the potential for improved
    water quality from the remining operation.”
    In sum, the EPA’s inclusion of numeric and non-numeric limitations in the guideline for the
    coal remining subcategory was a reasonable exercise of its authority under the CWA.
    d. Intercepted and Commingled Discharges
    Petitioners also complain that the Final Rule ignores the language and intent of the Rahall
    Amendment by requiring a pre-existing discharge that is “intercepted by active mining or that is
    commingled with waste streams from active mining areas for treatment” be subject 20   to the higher
    categorical standards for commingling of waste streams found in 40 C.F.R.§ 434.61. 40 C.F.R.
    § 434.71(b). Thus, the availability of the alternative standards under the Final Rule is limited to
    discharges that are relocated or diverted, and to pre-existing discharges “[a]fter commingling has
    ceased.” 
    Id. Petitioners contend
    that the Final Rule conflicts with the language in the Rahall
    Amendment stating that modifications shall be allowed “from the remined area of any coal remining
    operation . . . or in any pre-existing discharge affected by the remining operation.” 33 U.S.C.
    § 1311(p)(1). In other words, Petitioners contend that this regulation is too strict.
    The EPA responds that this language does not contemplate that active mining discharges will
    necessarily have to be commingled with the pre-existing discharges at remining sites. Thus,
    according to the EPA, Petitioners’ reading would actually broaden the scope of permit modifications
    authorized by section 301(p) by applying the grant of authority to discharges from the entire
    remining operation rather than to pre-existing discharges only. The EPA further contends that the
    permit modification authority applies only to the pre-existing discharges and not to discharges
    19
    A pollutant loading is calculated as a product of a flow measurement and a pollutant’s concentration. 67 Fed.
    Reg. at 3384.
    20
    That provision states that
    [w]here waste streams from any facility covered by this part [Part 434] are combined for treatment or
    discharge with waste streams from another facility covered by this part, the concentration of each
    pollutant in the combined discharge may not exceed the most stringent limitations for that pollutant
    applicable to any component waste stream of the discharge.
    40 C.F.R. § 434.61. This provision was promulgated in its present form when part 434 was substantially amended in
    October 1982. 47 Fed. Reg. 45,382 (Oct. 13, 1982). Thus, the general requirement that a commingled discharge would
    be subject to the more stringent requirements was established at the time section 301(p) was introduced.
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                      Page 17
    resulting from the active mining activities of a remining operation, citing 33 U.S.C. § 1311(p)(1),
    (p)(3)(C).
    As previously discussed, the Rahall Amendment does not prohibit the EPA from applying
    its general commingling rule to commingling of pre-existing discharges. See 40 C.F.R. § 434.61.
    Further, as the EPA notes, the term “pre-existing discharge” is defined in section 301(p)(3)(C) as
    “any discharge at the time of permit application,” or in other words, those discharges in existence
    prior to the start of remining. The phrase “in any pre-existing discharge affected by the remining
    operation” is ambiguous, and we cannot say that the EPA’s interpretation of this provision is
    inconsistent with section 301(p) and therefore unreasonable. We therefore defer to the EPA’s
    construction of its own regulations.
    Lastly, Petitioners assert that section 402(l)(2), 33 U.S.C. § 1342(l)(2), bars applying effluent
    limitations guidelines to pre-existing discharges that are diverted in order to avoid interception by
    or commingling with active mining discharges. This argument makes no sense. Section 402(l)(2)
    provides an exemption from the requirement for a permit for “discharges of stormwater runoff from
    mining operations . . . which are not contaminated by contact with, or do not come into contact with,
    any overburden, raw material, . . . byproduct or waste products located on the site of such
    operations.” 33 U.S.C. § 1342(l)(2). Pre-existing discharges at remining sites are highly
    contaminated as a result of contact with the raw materials, overburden, and waste products of prior
    mining activities. And, section 301(p) inherently recognizes that pre-existing discharges are subject
    to NPDES permit requirements. In short, the EPA’s decision to apply an existing regulation rather
    than the Final Rule to waste streams that have been commingled was within its discretion and not
    unreasonable.
    e. Response to Dissent
    The dissent’s primary argument is that the EPA’s decision to set effluent reduction attainable
    at zero for remining areas was arbitrary and capricious. The dissent asserts that, “[b]y requiring
    remining operators to preserve only the status quo, the EPA undermines the potential for reduction
    that is available and required under the Clean Water Act.” This argument overlooks the whole point
    of not only the Final Rule, but the Rahall Amendment itself. The Rahall Amendment was enacted
    based upon congressional findings that remining and reclamation were not occurring due to
    prohibitive cost of meeting effluent limitations guidelines applicable to virgin lands. The Final Rule
    was promulgated because the Rahall Amendment had not proven effective in encouraging remining.
    In setting the effluent reduction attainable at zero, the Final Rule took its cue from the Rahall
    Amendment itself, which set the baseline at zero. See 33 U.S.C. § 1311(p)(2) (stating that “in no
    event shall such a permit allow the discharges of iron and manganese, to exceed the levels being
    discharged from the remined area before the coal remining operation begins”). Furthermore,
    although the Final Rule sets the effluent limitation for each pollutant as not to exceed “baseline
    loadings,” the Coal Remining Subcategory at the same time requires every remining operator to
    create a site-specific Pollution Abatement Plan that is “designed to reduce the pollution load from
    pre-existing discharges.” 40 C.F.R. § 434.72(a).
    The dissent also contends “there is no evidence that the EPA ever explored the prospect of
    accomplishing more than preserving the status-quo” in attempting to meet the Act’s goal of
    eliminating the discharge of all pollutants. However, as the EPA points out and the record reflects,
    the EPA considered available technology in several ways prior to the enactment of the Coal
    Remining Subcategory. Prior to promulgating the Final Rule, the EPA was aware that the
    technologies underlying the existing effluent limitations for the Coal Mining Category were not
    effective in reducing pollution from abandoned mines, because remining under the regulations and
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                                Page 18
    available technology was not occurring, even after the Rahall Amendment. 65 Fed. Reg. at
    19,448.21
    In addition, the EPA performed a specific analysis of forty remining sites to determine if
    application of the existing effluent limitations would be cost effective. 
    Id. at 19,446.
    The EPA
    performed a cost comparison analysis of the forty remining sites submitted by the State of
    Pennsylvania and determined that in all forty cases, remining was deemed not economically feasible
    if the pre-existing discharges were required to meet the existing effluent limitations. 
    Id. Furthermore, the
    EPA considered the reductions that could be achieved through BMP
    technology. Specifically, it reviewed data submitted from the State of Pennsylvania from 112 closed
    remining sites operating under Rahall permits, representing the most extensive data currently
    available for assessment of the water quality impacts of BMP implementation at remining
    operations. 
    Id. at 19,447.
    This study showed that implementation of appropriate BMPs was
    effective in improving or eliminating acidity loading in 45% of the pre-existing discharges, total iron
    loading in 44% of the discharges, and total manganese in 42% of the discharges. 
    Id. at 19,448.
    The
    annual combined reductions in pollutant loadings from the BMPs equaled 5.8 million pounds of
    acidity, 189,000 pounds of iron, and 11,400 pounds of manganese. 
    Id. The EPA
    also examined
    thirteen active remining operations included in its Coal Remining Database and determined that
    roughly 58% showed a decrease in mean pollutant loadings and approximately 50% showed a
    decrease in pollutant loadings after the implementation of BMPs. 67 Fed. Reg. at 3389.
    Finally, the EPA specifically discussed the available technology for coal remining in its
    consideration of the various levels of control under section 304(b). 67 Fed. Reg. 3370, 3379-80.
    Regarding BAT, the EPA concluded that there was only one available technology that was more
    stringent than the BMPs, namely pre-existing treatment, but that it was not economically feasible
    for remining sites. 
    Id. at 3379.
    This was also true for BPT, where the EPA determined that pollutant
    abatement plans represented the average of the best technology currently available due to
    Pennsylvania’s experience in employing BMPs in Rahall permits. 
    Id. at 3380.
    As for BCT, the EPA
    applied its BCT cost-effectiveness test and determined that there are “no technologies that can
    achieve greater removals of conventional pollutants than established for BPT that are also cost-
    reasonable under the BCT costs test.” 
    Id. Finally, as
    for “elimination” of pollutants, the EPA noted
    that “zero [pollutant] loadings are expected to occur, at least for some remining sites, after regrading
    and contouring when discharge flows may be reduced greatly”; and that “zero flows have been
    observed after remining at some mine sites.” 
    Id. at 3385.
    This, however, had occurred in only a
    small percentage of facilities and was therefore not identified as an attainable effluent limitation for
    BPT, BAT, or BCT. See Office of Water, Office of Sci. and Tech. Eng’g and Analysis Div. Envtl.
    Prot. Agency, Coal Remining–Best Management Practices Guidance Manual, EPA 821-B-01-010,
    at 6-8 (2001), available at http://epa.gov/guide/coal/bmp/section6.pdf (last visited May 4, 2006)
    (chart showing overall percentages of discharges achieving varying levels of water quality
    improvements from the implementation of BMP technology).
    21
    The EPA found:
    The current regulations at 40 CFR part 434 create a disincentive for remining because of their
    high compliance costs. Moreover, the potential of the statutory exemption contained in the Rahall
    Amendment to overcome this disincentive and derive the maximum environmental benefits from
    remining operations has not been fully realized in the absence of implementing regulations. If mining
    companies face substantial potential liability or economic loss from remining, they will continue to
    focus on mining virgin areas and ignore abandoned mine lands that may contain significant coal
    resources. Based on information collected in support of this proposal, EPA believes that remining
    operations are environmentally preferable to ignoring the coal resources in abandoned mine lands.
    65 Fed. Reg. at 19,448.
    No. 02-3628              Citizens Coal Council, et al. v. EPA                                            Page 19
    Thus, contrary to the dissent’s position, the EPA determined that the pre-existing technology
    had resulted in little discharge reductions and therefore no environmental improvement, because the
    industry was not engaging in remining. Because BMP technologies had been proven to reduce
    pollutant loadings at remining sites, the EPA deemed it preferable to allow the use of BMPs rather
    than no action. This reading is entirely consistent with the CWA’s goal of eliminating water
    pollution. The EPA’s decision to set the effluent reduction attainable at zero was not arbitrary or
    capricious.
    B. Western Alkaline Coal Mining Subcategory
    The Western Alkaline Coal Mining Subcategory adopts alternative sediment control
    technologies for reclamation areas22 at coal mining sites in the arid West. Petitioners argued the
    EPA lacked authority to eliminate effluent limitations and replace them with BMPs, and, even if it
    has that authority, the EPA failed to demonstrate that continued imposition of the current effluent
    limitations for settleable solids and pH is infeasible. Petitioners further contend that the subcategory
    is overbroad, irrational, and inconsistent with the CWA and the SMCRA.
    1. Authority
    Petitioners contend that allowing BMPs instead of numeric effluent limitations in NPDES
    permits for Western Alkaline Coal Mine reclamation sites is contrary to the CWA. Again, for the
    reasons discussed elsewhere in this opinion, this argument is rejected.
    2. Infeasibility
    Petitioners also contend that the EPA failed to demonstrate that continued use of current
    effluent limitations for settleable solids and pH is infeasible, and that replacing them with BMPs was
    not supported by the record and was arbitrary and capricious.
    The EPA states that the administrative record supports its determination that numeric limits
    are infeasible because, in order to comply with these limits, the operator of a reclaimed mine site
    must use a sedimentation pond, and these ponds have serious non-water quality impacts in the arid
    West that may harm the environment. In the proposed rule the EPA stated that although
    sedimentation ponds are proven to be effective at reducing sediment discharge, in some instances
    they promoted negative environmental impacts in the arid and semiarid western regions. 65 Fed.
    Reg. at 19,453. Sedimentation ponds in reclamation areas are designed to capture and store water
    from a precipitation event and then slowly release the water in a continuous, low-velocity discharge.
    The EPA remarked that the slow release of water containing low amounts of sediment has caused
    negative environmental impacts in arid regions, including disruption of the natural hydrologic and
    sediment balance, stream channel instability, and water loss due to evaporation. 
    Id. The EPA
    explained that, in arid and semiarid western coal mine regions, the natural
    vegetative cover is sparse and rainfall typically occurs during localized, high-intensity, short-
    duration storms. These conditions contribute to flash floods and turbulent flows that transport large
    amounts of sediment that may contain up to several hundred thousand milligrams per liter of total
    suspended solids. Furthermore, fluvial areas and receiving channels in the arid west have developed
    according to these natural conditions. The receiving channels consist primarily of ephemeral
    arroyos that transport large volumes of flow and sediment, and may be affected by the alteration of
    sediment concentration and flow volume when sedimentation ponds are used. The EPA opined that
    “[d]ischarge of sediment-free water from a sedimentation pond may actually accelerate channel
    22
    “Reclamation area” means the surface area of a coal mine which has been returned to required contour and
    on which revegetation (seeding or planting) work had commenced. 67 Fed. Reg. at 3375.
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                    Page 20
    erosion because the sediment-free water will entrain sediment from the channel immediately below
    the pond.” 
    Id. If the
    sedimentation pond is later removed upon successful reclamation of the site,
    drainage from the reclaimed area would flow uninterrupted into the downstream watershed. 
    Id. This return
    to natural flow volumes and concentrations would “shock” the drainage channel,
    disrupting the fluvial and hydrologic balance that has developed based on the sedimentation pond
    discharge. This disruption in turn may cause severe channel reconfiguration, making the area more
    susceptible to instability and erosion than the pre-mining undisturbed conditions. 
    Id. Thus, the
    EPA
    determined that the most environmentally responsible goal was to reclaim the land such that the
    natural sediment loadings and hydrologic balance of undisturbed conditions is maintained at post-
    mined lands. 
    Id. Contrary to
    Petitioners’ assertions, the EPA’s conclusions are based on data in the
    administrative record. The administrative record reflects the EPA based its final rule on the results
    of the coal mining industry study conducted in the arid and semiarid region west of the 100th
    meridian. In that document, entitled, “Development Document–Proposed Western Alkaline Coal
    Mining Subcategory,” the EPA studied the problems that have been associated with the use of
    sedimentation ponds in arid and semiarid regions, examined modeling techniques that aid in BMP
    design and prediction effectiveness, studied the advantages provided by implementation of a site-
    specific sediment control BMP system, and looked at case studies showing both the use of mine
    models to determine sedimentation and the use of BMPs at existing mine sites in the arid and
    semiarid western coal region. These determinations, involving scientific and technical evaluations,
    are entitled to special deference, see BP 
    Exploration, 66 F.3d at 792
    , and Petitioners have not
    challenged the data itself or argued that specific data fails to support the EPA’s conclusions.
    Instead, Petitioners assert that the EPA acknowledges throughout the rulemaking record that
    sedimentation ponds are effective at reducing sediment discharge, and that meeting current numeric
    effluent limitations is feasible. However, this argument misses the point. As the EPA points out,
    feasibility, as used in the Act, concerns the establishment of numeric limitations; it does not pertain
    to whether a particular technology is the most appropriate technology. The EPA determined that
    sedimentation ponds, although an effective technology for controlling sediment, are not the most
    appropriate technology to meet the numeric standards for settleable solids in the arid and semiarid
    West because of the negative impacts in arid and semiarid environments. Further, the EPA in fact
    noted that numeric limits based on sedimentation ponds are infeasible in arid areas because
    precipitation is sporadic, intense, and isolated, thereby making it extremely difficult to evaluate
    overall performance of BMPs. 67 Fed. Reg. at 3381. This conclusion is based on data contained
    in the administrative record. See Phase I Report: Technical Information Package: Western Alkaline
    Mining Subcategory, dated January 21, 1999.
    Thus, the EPA’s determination that alternative controls were appropriate for the new
    subcategory of reclamation areas is based on several reasons supported in the administrative record,
    namely that sediment is a natural component of runoff in western watersheds, sediment is typically
    the only parameter of concern in runoff in western alkaline reclamation areas, BMPs are proven to
    be effective at controlling sediment, and computer modeling procedures can accurately predict
    sediment runoff conditions. These reasons amply supported the creation of the new subcategory,
    and the Administrator’s actions were not arbitrary and capricious. Finally, under the new rule,
    sediment pond technology is still available; at the same time, states now have the flexibility to
    develop alternative strategies through application of effluent guidelines in the Western Alkaline Coal
    Mining Subcategory.
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                                 Page 21
    3. Irrationality
    Petitioners claim that this subcategory is overbroad and irrational because the perceived
    problems associated with the use of sedimentation ponds to meet effluent limits do not justify the
    remedy of eliminating numeric effluent limits for all Western Alkaline mines in arid and semiarid
    regions. In support, Petitioners note that the EPA acknowledged in its rulemaking that the23problem
    is not universal and that sedimentation ponds are effective at reducing sediment discharge. Again,
    as a matter within its technical expertise, the EPA’s decision to establish the Western Subcategory
    was reasonable and entitled to deference. See BP 
    Exploration, 66 F.3d at 792
    , 804. As just
    discussed, the administrative record supports the EPA’s decision. Furthermore, as a practical
    matter, the use of sedimentation ponds is still an option. The new subcategory merely gives
    operators of reclamation sites and permitting authorities flexibility in designing BMPs to control
    sediment runoff.
    4. Statutory Intent
    Petitioners claim that the Administrator’s goal of approximating natural conditions is
    manifestly contrary to the intent of Congress, as expressed in the CWA and SMCRA, and that by
    redefining the goal for effluent reductions from mining operations in the west, the Administrator has
    replaced the CWA’s goal of “elimination of pollutants” with “not upsetting the natural hydrologic
    balance.” See 67 Fed. Reg. at 3404.
    As the EPA points out, this argument ignores the fact that the CWA gives the EPA wide
    latitude in fashioning effluent limitation guidelines. We addressed a similar argument in BP
    Exploration. There, the NRDC argued that the EPA illegally refused to require a zero discharge of
    produced waters through reinjection, because the record showed that reinjection was technologically
    and economically feasible. We stated:
    The NRDC mistakenly asserts that BAT must be based on the “best single performer
    in an industry.” To the contrary, the CWA’s requirement that EPA choose the “best”
    technology does not mean that the chosen technology must be the best pollutant
    removal. Obviously, BAT and NSPS must be acceptable on the basis of numerous
    factors, only one of which is pollution control.
    NRDC ignores the statutory language, which sets up a “‘limited’ balancing
    test.” 
    Weyerhaeuser, 590 F.2d at 1045
    (citing Senator Muskie’s remarks during
    debate on the CWA). In enacting the CWA, “Congress did not mandate any
    particular structure or weight for the many consideration factors. Rather, it left EPA
    with discretion to decide how to account for the consideration factors, and how much
    weight to give each factor.” 
    Id. BP Exploration,
    66 F.3d at 796. The CWA instructs the EPA to consider other factors besides
    pollution reduction, including costs, non-water quality environmental impacts, and (in the case of
    BPT and BAT) “such other factors as the Administrator deems appropriate.” 33 U.S.C.
    §§ 1314(b)(1), (b)(2), 1316(b)(1). Thus, contrary to Petitioners’ assertion, the subcategory is not
    inconsistent with the CWA.
    23
    Furthermore, “[a]n initial agency interpretation is not instantly carved in stone,” and “the agency, to engage
    in informed rulemaking, must consider varying interpretations and the wisdom of its policy on a continuing basis.”
    
    Chevron, 467 U.S. at 863
    . Thus, the fact that the EPA has redefined the goal for effluent limitations does not render its
    action arbitrary and capricious.
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                   Page 22
    Petitioners also cite SMCRA section 515(b)(10)(B)(i), which requires that surface mining
    be conducted “so as to prevent . . . additional contributions of suspended solids to streamflow.” 30
    U.S.C. § 1265(b)(10)(B)(i). Petitioners interpret “additional” in the statute to require the EPA to
    promulgate regulations designed to remove all pollutants from discharge. The EPA counters that
    the Western Alkaline Mining Subcategory guidelines do not conflict with this provision because
    “additional contributions” in the SMCRA means contributions in excess of pre-mining conditions,
    which is the same standard as the subcategory. The EPA argues that this interpretation is supported
    by SMCRA section 515(b)(10), which requires a mine operator to “minimize the disturbances to the
    prevailing hydrologic balance at the mine-site and in associated offsite areas.” 
    Id. § 1265(b)(10).
    The EPA asserts that “prevailing hydrological balance” means the balance prevailing before mining
    began, and that this is the balance that the challenged guideline seeks to restore. Thus, the EPA
    argues that the guideline is in conformity with the SMCRA. This is not an unreasonable
    interpretation, and we therefore defer to the EPA’s reading.
    Further, the administrative record reflects that the EPA worked with OSMRE to ensure that
    the EPA’s regulations coordinated with SMCRA requirements. The EPA explained that the
    SMCRA permit application process requires a coal mining operator to submit an extensive operation
    and reclamation plan, documentation, and analysis to OSMRE or the permitting authority for
    approval. The EPA noted that, similar to Western Subcategory guidelines, OSMRE requirements
    include a description of baseline ground and surface water characteristics under seasonal conditions,
    and an analysis of the hydrologic and geologic impacts caused by reclamation activity. 67 Fed. Reg.
    at 3382-83.        Specifically, the plan must include a “probable hydrologic consequences”
    determination of the impacts of the mining on existing hyrdrologic conditions and a hydrologic
    reclamation plan to show measures for reducing impact and to meet water quality laws and
    regulations. The coal mining regulatory authority is also required to conduct a similar analysis. 
    Id. at 3383.
    The EPA concluded that sediment control plans developed to comply with SMCRA
    requirements will usually fulfill the requirements of the Western Alkaline Coal Mining Subcategory.
    
    Id. Further, the
    guidelines’ requirement of using modeling techniques is also consistent with
    SMCRA permit requirements, because mining facilities already submit a watershed model as part
    of their SMCRA reclamation plan. Additionally, OSMRE observed in its comments to the proposed
    rule that, since November 1986, SMCRA regulations no longer require that all runoff from
    disturbed areas be treated by passing through a sedimentation pond. Instead, the rules require that
    appropriate sediment control measures be designed, constructed, and maintained using the best
    technology currently available to prevent additional contributions of sediment outside the permit
    area, to meet effluent limitations and to minimize erosion to the extent possible. (Id.)
    In short, Petitioners’ contention that the Western Alkaline Coal Mining Subcateory conflicts
    with the CWA and SMCRA is without merit.
    Thus, based upon the arguments properly presented in the initial appeal, we hold that the
    EPA did not violate Chevron or act arbitrarily or capriciously in promulgating the Final Rule.
    C. Panel Majority Decision
    The panel majority did not address any of Petitioners’ challenges to individual aspects of the
    Final Rule, discussed above. Instead, it sua sponte found that the Coal Remining and Western
    Alkaline regulations were promulgated “without procedure required by [§ 304(b) of the Clean Water
    Act],” failed to consider available technology, and failed to consider site-specific factors in
    establishing the guidelines. The EPA asserts that the en banc Court should not reach the merits of
    the issues upon which the panel majority previously ruled sua sponte against the EPA, and that if
    it does decide the merits, uphold the Final Rule.
    No. 02-3628                Citizens Coal Council, et al. v. EPA                                                   Page 23
    The EPA argues that we should not reach issues that were not previously raised to the EPA
    during the course of rulemaking24 or to the Court during the initial appeal.25
    The general rule of appellate procedure is that issues not presented in an appellant’s initial
    merits brief are waived. Brindley v. McCullent, 
    61 F.3d 507
    , 509 (6th Cir. 1995). Petitioners had
    a full and fair opportunity to make all of their arguments to the Court in the initial merits brief. That
    they chose not to raise such arguments should be definitive of the matter here. Although this rule
    is prudential and not jurisdictional, and may be waived “in exceptional cases or to avoid a
    miscarriage of justice,” Mayhew v. Allsup, 
    166 F.3d 821
    , 823 (6th Cir. 1999); see also Zhislin v.
    Reno, 
    179 F.3d 810
    , 813 n.1, 812-14 (6th Cir. 1999); Dorris v. Absher, 
    179 F.3d 420
    , 425-26 (6th
    Cir. 1999), this is not such a case. Rather, this case involves a rather typical challenge to an EPA
    rulemaking. The parties are all sophisticated entities, well-versed in both the subject matter and the
    administrative and legal process. It is telling that Petitioners did not raise these particular challenges
    either during notice and comment or in the initial appeal. It can be presumed that they did not make
    these arguments because they did not have a problem with the manner in which the Administrator
    promulgated the Final Rule, and only make the arguments now in order to preserve the panel
    majority’s conclusion in their favor. There are no exceptional circumstances in this case to excuse
    application of the general rule to the issues decided sua sponte by the panel majority. Cf. 
    Zhislin, 179 F.3d at 813
    n.1 (holding that failure to consider controlling Supreme Court decision issued after
    the appellant had filed his merits brief might result in a miscarriage of justice); 
    Dorris, 179 F.3d at 425-26
    (ruling that the court could consider the application of a statute favorable to the defendant’s
    position despite the defendant’s failure to address it at either the trial or on appeal; concluding that
    the district court had both incorrectly applied the statute in question against one of the appellants
    and awarded an excessive judgment based on an erroneous interpretation of the applicable statute
    and only a pure question of law was presented).
    In short, the panel majority erred in ruling on grounds not raised by the parties. Because it
    was improper for the panel majority to reach issues not briefed by the parties, especially in the
    context of the deferential Chevron analysis, we decline to reach those issues here. We granted en
    24
    In their original brief, Petitioners state that both “CCC and KRC participated in the administrative
    proceedings in which the Administrator proposed and then adopted the challenged regulation.” (Petr.’s Br. at 4.).
    25
    The EPA’s argument that we should not consider arguments not raised at the appropriate time in the
    underlying administrative proceedings implicates the rulemaking/adjudication dichotomy pervasive in administrative
    law. The waiver rule should not be applied freely in both areas, given the fundamental differences between the two
    endeavors. See, e.g., Am. Motorcyclist Ass’n v. Watt, 
    543 F. Supp. 789
    , 794-95 (C.D. Cal. 1982) (“The rule barring
    assertion of points not previously raised before the agency makes considerable sense where . . . the prior administrative
    proceeding was adjudicatory in nature, at least in the sense that identifiable parties were involved or application of a rule
    to specific individuals was considered. However, it would be inappropriate to apply this rule to notice and comment rule
    making of the type involved in the present case.”); see also David Douglas Spencer, Note, The Duty to Participate in
    Agency Rulemaking, 54 Geo. Wash. L. Rev. 628, 660 (1986) (“A sounder approach is taken by those courts that apply
    the doctrine while taking into account the differences between rulemaking and adjudication. Both the broad applicability
    and the future perspective of rulemaking decisions should make courts hesitant to apply the review preclusion doctrine
    to deny judicial review to parties challenging agency rules.”).
    There are cases involving environmental law determinations that fall on the rulemaking side of the
    rulemaking/adjudication dichotomy for certain purposes holding that a party challenging a rule can waive an issue by
    not making a comment on point during the comment period. See, e.g., Univ. Health Servs. Inc. v. Thompson, 
    363 F.3d 1013
    , 1019-20 (9th Cir. 2004) (and cases cited therein); Mich. Dep’t of Envtl. Quality v. Browner, 
    230 F.3d 181
    , 183
    n.1 (6th Cir. 2000); see also BP Exploration & Oil, Inc. v. EPA, 
    66 F.3d 784
    , 798 (6th Cir. 1996) (noting that industry
    petitioners had sufficient notice of issue during rulemaking, but concluding that even if the industry petitioners had the
    right to raise the issue, the EPA’s argument on the merits was dispositive). However, all of these cases nonetheless
    contain some characteristics of adjudications, and should not be applied broadly.
    In any event, we need not resolve this question in this case, because this case is decided on other grounds.
    No. 02-3628          Citizens Coal Council, et al. v. EPA                                  Page 24
    banc review, however, precisely to vacate the panel majority’s sua sponte determination of those
    unbriefed issues. Thus, we end our analysis here and uphold the Final Rule.
    IV. Conclusion
    The EPA was faced with a situation in which Congress enacted CWA section 301(p) to
    provide incentives for remining abandoned mine lands by exempting remining operations from
    certain BAT effluent limits, yet coal mining companies remained hesitant to pursue remining
    without formal EPA approval and guidelines. The EPA recognized that although available
    technologies theoretically might reduce effluents at abandoned mines, these reductions were not
    occurring because they were cost prohibitive. Thus, the Final Rule was a reasonable response to a
    real problem. Had the EPA done nothing in light of this problem, it would not have been acting to
    fulfill the mandate of Congress “to restore and maintain the chemical, physical, and biological
    integrity of the Nation’s waters.” 33 U.S.C. § 1251. Nothing in the plain text of the Rahall
    Amendment, section 304(b), or section 301 prohibits the EPA from promulgating the Final Rule,
    and the EPA’s determination of its authority to establish the new subcategories was not arbitrary or
    capricious. Finally, the Final Rule accomplishes what the Rahall Amendment sought, but failed, to
    achieve. For all these reasons, then, we AFFIRM the Final Rule for the Coal Remining
    Subcategory and the Western Alkaline Coal Mining Subcategory and DENY the petition.
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                               Page 25
    ________________
    DISSENT
    ________________
    BOYCE F. MARTIN, JR., Circuit Judge, with whom DAUGHTREY, MOORE, COLE, and
    CLAY, Circuit Judges, join, dissenting. Because the EPA regulations at issue represent a race to
    the bottom in conflict with the Clean Water Act’s goal of cleaner water, I respectfully dissent from
    the Court’s decision upholding the regulations. It is disheartening to see a regulatory agency adopt
    regulations that effectively undermine its own authority and sell out the environment for the sake
    of the coal industry. The regulation at issue in this case does just that. The Final Rule that the EPA
    adopted is inconsistent with Congress’s intent and reflects arbitrary and capricious interpretations
    of the Clean Water Act. I would invalidate the EPA’s Final Rule under the Administrative
    Procedure Act, 5 U.S.C. § 706, and therefore respectfully dissent from the majority’s decision to
    sustain it.
    I.
    The majority opinion does an admirable job attempting to sort through the complex facts of
    this case. Nonetheless, the majority has failed to highlight several important aspects of the
    background of this case. Thus, I will begin with my own short review of the facts in order to
    properly frame the issues. The original panel majority aptly noted that the Clean Water Act, located
    at 33 U.S.C. § 1251, et seq., is an enigmatic piece of legislation. See also Am. Petroleum Institute
    v. EPA, 
    540 F.2d 1023
    , 1027 (10th Cir. 1976) (“The Act is difficult to understand, construe and
    apply.”). Its breadth is expansive, its technical terms daunting, and its circuitous cross-references
    dizzying. It is, to say the least, complex.1 The Clean Water Act’s purpose, however, is simple. The
    Act seeks to “restore and maintain the chemical, physical, and biological integrity of the Nation’s
    waters.” 33 U.S.C. § 1251; see also BP Exploration & Oil v. EPA, 
    66 F.3d 784
    , 789 (6th Cir. 1995);
    E.I. du Pont de Nemours & Co. v. Train, 
    430 U.S. 112
    (1977); see EPA v. State Water Resources
    Control Bd., 
    426 U.S. 200
    , 202-09 (1976) (discussing the adoption of the Clean Water Act); Natural
    Res. Def. Council v. Train, 
    510 F.2d 692
    , 695-97 (D.C. Cir.1975) (same). “[T]he guiding star [of
    the Clean Water Act] is the intent of Congress to improve and preserve the quality of the Nation’s
    waters.” Am. Petroleum 
    Institute, 540 F.2d at 1028
    . Congress’s original goal in promulgating the
    Act was to eliminate the discharge of all pollutants into navigable waters by the year 1985. See 33
    U.S.C. § 1251(a)(1).
    To further the Act’s goal of restoring and maintaining clean water, Congress established both
    technology-driven limits and water-quality-based limits on pollution. See Natural Res. Def. Council
    v. EPA, 
    915 F.2d 1314
    , 1317 (9th Cir. 1990); BP Exploration & 
    Oil, 66 F.3d at 789
    (“The CWA
    directs EPA to formulate national effluent-limitation guidelines for those entities that discharge
    pollutants into the navigable waters of the United States.”). Technology-based limits are designed
    to prevent pollution by requiring polluters to install and use various forms of technology designed
    to reduce the pollution discharged into the nation’s waters. See Texas Oil & Gas Ass’n v. EPA, 
    161 F.3d 923
    , 927 (5th Cir. 1998). Water-quality regulations are triggered when a given body of water’s
    pollution level exceeds the level that a state deems acceptable for the body of water’s intended use
    or function. Natural Res. Def. 
    Council, 915 F.2d at 1317
    (citing 33 U.S.C. §§ 1312, 1313).
    Essentially, when the technology-based regulations are insufficient to contain the aggregate impact
    1
    The Tenth Circuit has noted that “[p]opular demand for legislative action to control water pollution is shown
    by the fact that on the votes to override the presidential veto [of the Clean Water Act], only 12 senators and 23
    representatives voted to sustain the veto. Perhaps the pressure on Congress to do something was a major cause of the
    unsatisfactory legislation.” American Petroleum 
    Institute, 540 F.2d at 1027
    . The court also noted that “[t]he two
    volume, 1766 page, legislative history does not help us much.” 
    Id. No. 02-3628
                  Citizens Coal Council, et al. v. EPA                                                 Page 26
    of polluters’ activities despite the polluters’ compliance with the regulations, the water-quality
    regulations kick in. And when they kick in, they tighten the noose on polluters — that is, the
    pollution controls become stricter and require greater pollution control from the individual polluters.
    
    Id. The technology-driven
    regulations act in terms of “effluent-limitation guidelines.” An
    “effluent-limitation” is “any restriction [including schedules of compliance] established . . . on the
    quantities, rates, and concentrations of chemical, physical, biological, and other constituents which
    are discharged from point sources into . . . water.” 33 U.S.C. § 1362(11). The EPA’s role in all of
    this, the Clean Water Act instructs, is to issue effluent-limitation guidelines for various types of
    “point sources.” A “point source” is “any discernible, confined and discrete conveyance . . . [such
    as a pipe, ditch, or channel] from which pollutants are or may be discharged.” 
    Id. at §
    1362(14).
    In short, a “point source” spews pollutants into the environment. Thus, the EPA’s effluent-limitation
    guidelines are technology based regulations requiring polluters to adopt certain technologies
    designed to reduce pollution dispensed into the environment from a point source.
    Section 304 of the Clean Water Act charges the EPA with duties related to setting effluent-
    limitation guidelines for existing sources of pollution. The Clean Water Act “directs [the] EPA to
    institute progressively more stringent effluent discharge guidelines in stages.” BP Exploration &
    
    Oil, 66 F.3d at 789
    (emphasis added). Under section 304(b)(3), the EPA must identify the specific
    control measures and practices available to the various categories and classes of point sources. 33
    U.S.C. § 1314(b)(3). The EPA is required to then take these existing point sources of pollution and
    identify, in terms of the amount of pollutants, the pollution reduction (or “effluent reduction”)
    attainable through application of the three different sources of technology — best practicable
    control technology (“BPT”), 
    id. at §
    1314(b)(1)(A), best available technology economically
    achievable (“BAT”), 
    id. at §
    1314(b)(2)(A), and best conventional pollutant control technology
    (“BCT”), 
    id. at §
    1314(b)(4)(A).2 That is, “[a]t the first stage of pollutant reduction, [the] EPA is
    to determine the level of effluent reduction achievable within an industry with the implementation
    of” BPT. See BP Exploration & 
    Oil, 66 F.3d at 789
    (citing 33 U.S.C. § 1314(b)(1)(A)). In general,
    BPT is the “average of the best existing performances by [the various types of polluters] of various
    sizes, ages, and unit processes within the point source category or subcategory.” 
    Id. At the
    second
    stage of pollutant reduction, the EPA is “to set generally more stringent standards for toxic and
    conventional pollutants.” 
    Id. at 790.
    For toxic pollutants, the EPA is required to set the standard
    for the BAT. 
    Id. “BAT represents,
    at a minimum, the best economically achievable performance
    in the industrial category or subcategory.” 
    Id. (citing Natural
    Res. Def. Council, Inc. v. EPA, 
    863 F.2d 1420
    , 1426 (9th Cir. 1988) (citing EPA v. Natural Crushed Stone Ass’n, 
    449 U.S. 64
    , 74
    (1980))). Compared to BPT, which is required at the first stage of pollutant reduction, “BAT calls
    for more stringent control technology that is both technically available and economically
    achievable.” 
    Id. (emphasis added).
    In contrast to the toxic pollutants, conventional pollutants are
    governed by BCT. “This standard is designed to control conventional pollutants about which much
    is known but for which stringent BAT standards might require unnecessary treatment.” BP
    Exploration & 
    Oil, 66 F.3d at 790
    . BCT was intended by Congress to “prevent the implementation
    of technology for technology’s sake.” 
    Id. Thus, BCT
    is not a third level of pollutant control, but
    rather replaces BAT for conventional pollutants. 
    Id. Section 301
    of the Clean Water Act requires that the effluent-limitations attainable by the
    various levels of technology “shall be achieved” by various dates, all of which were are least fifteen
    2
    Congress intended for the EPA to consider numerous factors in achieving the goal of pollution reduction: “The
    Committee believes that there must be a reasonable relationship between costs and benefits if there is to be an effective
    and workable program.” Clean Water Act of 1972, Pub. L. No. 92-500, 1972 U.S.C.C.A.N. (86 Stat.) 3713.
    No. 02-3628              Citizens Coal Council, et al. v. EPA                                               Page 27
    years past. 
    Id. at §
    1311(b).3 Polluters discharging toxic and nonconventional pollutants must apply
    the best available technology economically achievable (BAT) to meet the BAT effluent-limitations.
    
    Id. at §
    1311(b)(2)(A). Point sources discharging conventional pollutants must apply the best
    conventional pollutant control technology (BCT) to meet the BCT effluent-limitations. 
    Id. at §
    1311(b)(2)(E).
    The effluent-limitations, however, are not self-executing — that is, they create binding
    obligations on polluters only through the National Pollutant Discharge Elimination System
    (“NPDES”) permits. See, e.g., Am. Paper. Indust., Inc. v. EPA, 
    996 F.2d 346
    , 350 (D.C. Cir. 1993).
    The Clean Water Act provides that: “Except as in compliance with this section and sections 1312,
    1316, 1317, 1328, 1342, and 1344 of this title, the discharge of any pollutant by any person shall be
    unlawful.” 33 U.S.C. § 1311(a). Stated another way, nobody may pollute without a NPDES permit.
    Further, “[t]hrough the Clean Water Act, Congress has directed the EPA to incorporate into the
    permits increasingly stringent technology-based effluent-limitations.” Rybachek v. EPA, 
    904 F.2d 1276
    , 1283 (9th Cir. 1990).
    The EPA Administrator may issue permits to individual polluters only if the permitted
    discharges comply with the requirements for effluent-limitations set out in section 301 of the Clean
    Water Act. 
    Id. at §
    1342(a). When issuing permits, the EPA determines the polluter’s place within
    the appropriate point source category and subcategory. The EPA then assigns the appropriate level
    of technology to the polluter. Finally, the EPA sets limits for various pollutants on how much the
    polluter may discharge. In setting the limits for individual polluters, the EPA imports standards
    from the effluent-limitation guidelines issued under section 304(b) of the Clean Water Act. The
    EPA then applies the numeric limitations to the polluter via the NPDES permit. This process,
    therefore, implements and executes the effluent-limitations attainable under the various applicable
    technologies.
    To set the effluent-limitation guidelines under section 304(b) of the Clean Water Act, which
    are then imported into setting the standards for individual polluters pursuant to the NPDES permit
    process, the EPA broke down the various point sources by the different industries. One of those
    industries that the EPA established was the Coal Mining Point Source Category. See 40 C.F.R.
    § 434. Then, in 1985, the EPA amended the effluent-limitation guidelines and requirements
    concerning the Coal Mining Point Source Category by creating four subcategories: (1) Coal
    Preparation Plants and Coal Preparation Plant Associated Areas; (2) Acid or Ferruginous Mine
    Draining; (3) Alkaline Mine Draining; and (4) Post Mining Areas. See 50 Fed. Reg. 41296 (1985).
    Thus, post-1985, any coal mining operator seeking an NPDES permit was placed into one
    of the four subcategories based on the type of mining operation being conducted. And, from that
    point on, whenever a coal mining operation applied for a NPDES permit, the EPA incorporated the
    effluent-limitation guidelines corresponding to the appropriate subcategory for that operation.
    Under the 1985 regulations, there was no category that encompassed coal mining operations that
    sought to remine previously mined, but later abandoned, lands. This had heretofore not been much
    of a problem because it was both technologically and economically infeasible to remine such lands.
    Advances in technology, however, had recently made it both feasible and potentially profitable to
    mine such abandoned lands.
    From the perspective of the industry there was one problem. The potential for profit from
    remining was made less likely by the fact that the same regulations for mining virgin lands applied
    3
    Polluters with point sources requiring best practicable control technology were to achieve the goals by July
    1, 1977. 
    Id. at 1311(b)(1)(A).
    March 31, 1989 was the date for point sources requiring the best conventional pollutant
    control technology, 
    id. at §
    1311(b)(2)(E), and best available technology economically achievable, 
    id. At §§
    1311(b)(2)(C)-(D), (F).
    No. 02-3628              Citizens Coal Council, et al. v. EPA                                              Page 28
    also to coal reminers. First time mining of virgin lands is, understandably, less costly, because it
    is naturally easier to keep virgin lands from exceeding pollution controls than it is for already mined
    land essentially leaking pollutants to be brought into compliance with the same regulations. Thus,
    reminers were faced with major and often prohibitive up-front costs in order to bring remining
    operations into compliance with the same standards for virgin lands. Weighing these costs led many
    potential reminers to leave abandoned mines to remain fallow, often with resulting untreated
    pollution from “pre-existing discharges” that could otherwise have been treated and improved.
    This often prohibitive up-front cost of remining got Congress’s attention. In 1987,
    responding to the disincentives for remining, Congress passed several amendments to the Clean
    Water Act. For our purposes, the relevant amendment was proposed by West Virginia
    Representative Nick Rahall. The “Rahall Amendment” became section 301(p) of the Clean Water
    Act. 33 U.S.C. § 1311(p); see also Stephen G. Allen, Special Legal Problems with Other
    Environmental Laws Under SMCRA, 7 J. Mineral & L. Pol. 129 (1991/1992) (calling the Rahall
    Amendment one of the “more fiercely debated provisions of the 1987 Water Quality Amendments”).
    The purpose of the Rahall Amendment was to stimulate the remining of abandoned coal mines by
    exempting certain remining operations from the otherwise applicable national effluent-limitations.
    In short, the Rahall Amendment recognized the disincentive to remining provided by employing the
    high national standard for effluent-limitations to remining operations. Thus, the Rahall Amendment
    lowered the bar to a more permissive effluent-limitation standard that Congress determined would
    provide sufficient incentive to remine the lands in order to restore and maintain the integrity of the
    nation’s water system.
    Under the Rahall Amendment, a remining operation permitted to take advantage of the more
    permissive effluent-limitation standard is one that began remining after February 4, 1987, at a site
    where coal mining had ceased prior to August 3, 1977. See 33 U.S.C. § 1311(p)(3)(A). Rather than
    requiring these remining operations to meet the national effluent-limitation guidelines mandated by
    the 1985 amendments to the Clean Water Act, the Rahall Amendment allowed4the NPDES permit
    writer — that is, the EPA Administrator or corresponding state administrator — to waive these
    requirements in favor of a modified permit.
    These modified permits incorporate pollution site-specific numerical limits for pre-existing
    discharges of iron, manganese, and pH based upon the Administrator’s “best professional
    judgment.” The Rahall Amendment, however, forbade the EPA or relevant state agency from
    issuing a permit if remining would cause discharges that exceeded the pre-existing levels of
    pollutants found in the surrounding waterways. See 33 U.S.C. § 1311(p)(2). The Amendment also
    required that applicants for these permits provide evidence that the remining operations would
    potentially improve the quality of water in the area, and that the permits comply with applicable state
    water quality standards. Thus, in sum, the Rahall Amendment endeavored to make remining
    economically feasible, while at the same time furthering the purposes of the Clean Water Act.
    Congress made the determination that remining would improve the water quality of abandoned
    mines, and that it would be an improvement that would not have occurred under the more stringent
    guidelines that applied to all mined lands.
    The EPA’s Final Rule
    The EPA apparently determined that the progress under the Rahall Amendment was
    insufficient and began the informal rulemaking process by publishing proposed amendments to
    effluent-limitation guidelines on April 11, 2000. 65 Fed. Reg. 19440 (April 11, 2000). Following
    4
    In jurisdictions where the EPA has authorized a state agency to administer the NPDES program, state agencies
    may issue NPDES permits. See 33 U.S.C. § 1342(a)-(d).
    No. 02-3628                Citizens Coal Council, et al. v. EPA                                                  Page 29
    notice and comment procedures, the EPA promulgated the Final Rule on January 23, 2002. 67 Fed.
    Reg. 3370 (Jan. 23, 2002).
    The Final Rule creates two new subcategories under the Coal Mining Point Source Category.
    The EPA had already created the original four subcategories of point sources following the 1985
    Amendments to the Clean Water Act in order to set the standards for individual polluters pursuant
    to the NPDES permit process. The two new subcategories are: (1) the Coal Remining Subcategory,
    and (2) the Western Alkaline Coal Mining Subcategory.
    The Coal Remining Subcategory
    The regulations for the Coal Remining Subcategory apply to all “pre-existing discharges”
    at “coal remining operations.” Pre-existing discharges are defined as any “discharge resulting from
    mining activities that have been abandoned prior to the time of a remining permit application.” See
    40 C.F.R. § 434.70(b) (2004). A coal remining operation is defined as “a coal mining operation at
    a site on which coal mining was previously conducted and where the site has been abandoned or the
    performance bond has been forfeited.” 
    Id. at §
    434.70(a).
    The regulations further require that every remining operator create and submit to the EPA
    or to the state agency with authority to issue NPDES permits a site-specific Pollution Abatement
    Plan. 
    Id. at §
    434.72(a). Remining operators must include within their Pollution Abatement Plans
    “best management practices,” and they must design the plans “to reduce the pollution load from pre-
    existing discharges.” 
    Id. Further, each
    Pollution Abatement Plan “must describe the design
    specifications, construction specifications, maintenance schedules, criteria for monitoring and
    inspection, and expected performance of [best management practices].” 
    Id. Under the
    Final Rule,
    these Pollution Abatement Plans qualify as “best practicable control technology currently available”
    or BPT, 
    id. at §
    434.72, the “best available technology economically achievable” or BAT, 
    id. at §
    434.73, and the “best conventional pollutant control technology” or BCT, 
    id. at §
    434.74, for
    remining operators. Stated another way, by creating and submitting these Plans, the Final Rule
    allows polluters to meet the Clean Water Act’s requirements in section 304(b)(3), which requires
    the EPA to identify the effluent reduction attainable under BPT, BAT, and BCT. See 33 U.S.C.
    § 1314(b)(1)(A)-(b)(4)(A) (also requiring the EPA to identify the facts it will consider when
    deciding which control measures and practices apply to each of the various point source classes and
    categories).
    The Final Rule also sets the effluent-limitations for four pollutants: (1) total iron; (2) total
    manganese; (3) net acidity; and (4) total suspended solids. For each of these four pollutants, the
    discharger’s effluent-limitation is that it “[m]ay not exceed baseline loadings.” See 40 C.F.R.
    § 434.72(b)(1). Baseline loadings are defined as background conditions or the status quo — that is,
    those conditions that exist when remining commences. The Final Rule prescribes the procedures
    for determining what these background conditions are. 
    Id. at §
    434 App. B. There is, however, an
    exception to the requirement of determining these background conditions. If the EPA or state
    permitting authority  concludes that collecting samples to measure the baseline pollutant loadings
    is “infeasible”5 and that remining “will result in significant improvement that would not otherwise
    occur,” then no measurable effluent-limitations apply. 
    Id. at §
    434.72(b)(2). Stated another way,
    if the EPA determines that it cannot measure how bad the pollution currently is at a remining site,
    5
    The Final Rule provides several examples of when such measurements would be infeasible: Pre-existing
    discharges for which it is infeasible to collect samples for determination of baseline pollutant levels include, but are not
    limited to, discharges that exist as a diffuse groundwater flow that cannot be assessed via sample collection; a base flow
    to a receiving stream that cannot be monitored separate from the receiving stream; a discharge on a steep or hazardous
    slope that is inaccessible for sample collection; or, a number of pre-existing discharges so extensive that monitoring of
    individual discharges is infeasible. See 40 C.F.R. § 434.72(b)(2).
    No. 02-3628                Citizens Coal Council, et al. v. EPA                                                  Page 30
    and makes the determination that remining will improve the site, then no measurable effluent-
    limitations apply to the polluter. There appear to be no guidelines upon which the EPA’s
    determinations here must be based; rather, it appears to be within the complete discretion of the EPA
    or state agency. These regulations apply to all three levels of technology. 
    Id. at §
    § 434.72, 434.73,
    434.74.
    Western Alkaline Coal Mining Subcategory
    The EPA’s Final Rule also creates and establishes effluent-limitations for the Western
    Alkaline Coal Mining Subcategory. The rule regulates certain    types of draining in certain areas of
    “western coal mining operations.” See 40 C.F.R. § 434.81.6 “Western coal mining operations” are
    defined as “surface or underground coal mining operation[s] located in the interior western United
    States, west of the 100th meridian west longitude, in an arid or semiarid environment with an
    average annual precipitation of 26.0 inches or less.” 
    Id. at §
    434.80(f).
    There are no uniform standards for drainage under the Rule. Instead, western alkaline coal
    mining operators are required to create “a site-specific Sediment Control Plan . . . designed to
    prevent an increase in the average annual sediment yield from pre-mined undisturbed conditions.”
    
    Id. at §
    434.82(a) (emphasis added). Like the Pollution Abatement Plans required under the Coal
    Remining Subcategory, each Sediment Control Plan “must identify best management practices
    (BMPs) and also must describe design specifications, construction specifications, maintenance
    schedules, criteria for inspection, as well as expected performance and longevity of the best
    management practices.” 
    Id. The Rule
    does not require Sediment Plans to include numerical
    limitations on sediment. Instead, operators must demonstrate “[u]sing watershed models” that the
    Sediment Control Plans “will result in average annual sediment yields that will not be greater than
    the sediment yield levels from pre-mined, undisturbed conditions.” 
    Id. at §
    434.82(b). Essentially,
    the Rule is a do-no-harm rule. Beyond that, the operators’ only obligation is to implement and
    maintain the best management practices described in the Sediment Control Plan. 
    Id. at §
    434(c).
    Under the Final Rule, this obligation satisfies the Clean Water Act’s requirements of BPT, 
    id. at §
    434.82, BAT, 
    id. at §
    434.83, and BCT, 
    id. at §
    434.84.
    II.
    Standard of Review
    The majority speeds through the standard of review and its intricacies, but I believe that it
    is necessary to consider, in some detail, the function we serve here. There are two separate and
    distinct inquiries in this case: First, we determine whether the EPA’s interpretation of the Clean
    Water Act — that the EPA has authority to promulgate regulations with regard to coal remining —
    complies with Chevron, U.S.A., Inc. v. Natural Res. Def. Council, Inc., 
    467 U.S. 837
    (1984). The
    second inquiry, separate and distinct from Chevron, is under the Administrative Procedure Act
    where we determine whether the agency’s rule may stand. See United States v. Mead Corp., 
    533 U.S. 218
    , 227 (2001) (“When Congress has explicitly left a gap for an agency to fill, there is an
    express delegation of authority to the agency to elucidate a specific provision of the statute by
    regulation, 
    Chevron, 467 U.S., at 843
    -844, and any ensuing regulation is binding in the courts unless
    procedurally defective, arbitrary or capricious in substance, or manifestly contrary to the statute.”
    (citing APA, 5 U.S.C. §§ 706(2)(A), (D))).
    6
    The Rule’s regulations apply to “alkaline mine draining,” see 40 C.F.R. § 434.81(a), and to draining with:
    (1) pH equal to or greater than 6.0; (2) dissolved iron concentration less than 10 mg/L; and (3) net alkalinity greater than
    zero, 
    id. at §
    434.81(b). Further, the regulation applies only to reclamation areas, brushing and grubbing areas, topsoil
    stockpiling areas, and regraded areas. 
    Id. at §
    434.82; see also 
    id. at §
    434.80 (providing definitions of the regulated
    areas).
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                    Page 31
    When we review an agency’s interpretation of a congressional statute, we engage in a two-
    step progress — the Chevron-two-step. The Chevron analysis is designed to determine whether
    agency action under a statute, here the Clean Water Act, is based on a permissible interpretation of
    the statute. The first step of this two-step dance requires us to determine whether Congress has
    directly spoken with regard to the matter at issue. 
    Id. at 842;
    Nations Bank of N.C., N.A. v. Variable
    Annuity Life Ins. Co., 
    513 U.S. 251
    , 257 (1995) (asking “whether the intent of Congress is clear as
    to the precise question at issue”); Wachovia Bank, N.A. and Wachovia Mortgage Corp. v. Watters,
    
    431 F.3d 556
    , 560-61 (6th Cir. 2005). “If the intent of Congress is clear, that is the end of the
    matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent
    of Congress.” 
    Chevron, 467 U.S. at 842-43
    (footnote omitted). If we determine that Congress has
    not directly addressed the specific issue in dispute, we proceed to step two of the dance. At step
    two, we may not simply impose our own view as to our construction of the statute. 
    Id. at 843.
    “Rather, if the statute is silent or ambiguous with respect to the specific issue, the question for the
    court is whether the agency’s answer is based on a permissible construction of the statute.” 
    Id. (footnote omitted);
    Clarke v. Sec. Indus. Ass’n, 
    479 U.S. 388
    , 403 (1987) (instructing courts to “give
    great weight to any reasonable construction” by the agency); Wachovia 
    Bank, 431 F.3d at 562
    (noting that if the agency’s “interpretation is reasonable, we must defer to its construction of the
    statute”); Hospital Corp. of America & Subsidiaries v. Commissioner of Internal Revenue, 
    348 F.3d 146
    , 140 (6th Cir. 2003) (noting that “regulations are appropriately accorded Chevron deference
    when they constitute an exercise of implicitly delegated power to give content to ambiguous
    statutory terms”).
    Separate and distinct from Chevron, because the EPA promulgated the regulations at issue
    through informal rulemaking, the scope of review over the informal rulemaking process is generally
    governed by section 10(2)(e) of the Administrative Procedure Act, 5 U.S.C. § 706(2); see also
    Weyerhaeuser Co. v. Costle, 
    590 F.2d 1011
    , 1024 (D.C. Cir. 1978). The APA precisely states that
    a court must “hold unlawful and set aside agency action, findings, and conclusions found to be . .
    . arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law . . . .” 5
    U.S.C. § 706(2)(A); see also generally Citizens to Preserve Overton Park, Inc. v. Volpe, 
    401 U.S. 402
    , 415-17 (1971). In conducting our review, we require that
    the agency must examine the relevant data and articulate a satisfactory explanation
    for its action including a rational connection between the facts and the choice
    made. . . . Normally, an agency rule would be arbitrary and capricious if the agency
    has relied on factors which Congress has not intended it to consider, entirely failed
    to consider an important aspect of the problem, offered an explanation for its
    decision that runs counter to the evidence before the agency, or is so implausible that
    it could not be ascribed to a difference in view or the product of agency experience.
    Motor Vehicle Manuf. Ass’n v. State Farm Mut. Auto. Ins. Co., 
    463 U.S. 29
    , 43 (1983). Moreover,
    this Court has adopted the D.C. Circuit’s approach dividing the review process into three categories:
    “statutory, procedural, and substantive.” BP Exploration & 
    Oil, 66 F.3d at 792
    (citing
    
    Weyerhaeuser, 590 F.2d at 1024
    ). In conducting our review, we
    intervene not merely in case of procedural inadequacies, or bypassing of the mandate
    in the legislative charter, but more broadly if the court becomes aware, especially
    from a combination of danger signals, that the agency has not really taken a ‘hard
    look’ at the salient problems, and has not genuinely engaged in reasoned decision-
    making.
    Greater Boston Television Corp. v. FCC, 
    444 F.2d 841
    , 851-52 (D.C. Cir. 1970), cert. denied, 
    403 U.S. 923
    (1971); Thomas Jefferson Univ. v. Shalala, 
    512 U.S. 504
    (1994) (determining that
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                   Page 32
    reviewing courts must “hold unlawful and set aside” agency action that is “arbitrary, capricious, an
    abuse of discretion, or otherwise not in accordance with law”) (quoting 5 U.S.C. § 706(2)(A)).
    Analysis
    “The language of the [Clean Water] Act clearly manifests Congress’[s] intention that [the]
    EPA formulate BPT, BAT, [and] BCT [] within certain time deadlines and having considered
    various factors.” BP Exploration & 
    Oil, 66 F.3d at 791
    (citing 33 U.S.C. § 1311(b)). Further,
    Congress obviously intended for the Clean Water Act to reduce pollutants and clean up the nation’s
    waters. For the most part, I agree with the majority and the original panel that the EPA has
    reasonably interpreted the Clean Water Act with respect to its authority to regulate in the coal
    remining area. The opt-in measures created by the Rahall Amendment do not necessarily limit the
    EPA’s authority to create and regulate subcategories as it did here. Nevertheless, in doing so, those
    regulations must also satisfy the requirements of the Administrative Procedure Act. The majority,
    in error I believe, devotes its analysis nearly exclusively to Chevron. At the end of each section of
    the majority’s analysis, it does make the perfunctory conclusion that the EPA did not act arbitrarily
    or capriciously in adopting the provisions of the Final Rule. The sum total of its analysis, however,
    appears to conflate the review under Chevron and the APA, seemingly suggesting that as long as
    Chevron is not violated, then no problems exist under the APA. This view is legally erroneous and
    fails to take into account the myriad of additional requirements the APA places upon agency
    regulations not implicated under Chevron. As a result, the majority’s analysis under the APA is
    surely not the “searching and careful review” of agency action required by Citizens to Preserve
    Overton 
    Park, 401 U.S. at 416
    . For that reason and the reasons set forth below, I dissent from the
    majority’s opinion and would hold that the EPA’s Final Rule is arbitrary and capricious and not
    otherwise in accordance with law.
    III.
    The Final Rule With Respect to the Coal Remining Subcategory Violates the Administrative
    Procedure Act
    Prior to 1987, with the stringent national effluent-limitation standards applicable to virgin
    lands as well as abandoned lands available for remining, it became clear that the higher standards
    provided a disincentive for remining. The cost to remine and bring the abandoned lands into
    compliance with the stringent national effluent-limitations standards was essentially cost-prohibitive
    in most circumstances and led to the acknowledgment that operators were passing up the
    opportunities to remine abandoned lands. Thus, Congress sought to counter this disincentive with
    the passage of the 1987 Amendments to the Clean Water Act — specifically, the Rahall
    Amendment. As described above, the Rahall Amendment incentivized remining abandoned lands
    by exempting certain remining operations and allowing them to operate under a more permissive
    standard. Nevertheless, the goal under the Clean Water Act, including the Rahall Amendment,
    remained improvement in the nation’s waters over that which existed prior to remining. See 133
    Cong. Rec. H 168 (daily ed. Jan. 8, 1987) (statement of Rep. Rahall) (“The end result of this effort
    will be the reclamation of the site and as such, as [sic] improvement in water quality over that which
    existed at the site prior to remining.”).
    I agree with the majority that the Rahall Amendment sets up a permissive opt-in scheme.
    The Amendment states that the EPA “may issue” modified permits under the Amendment, 33 U.S.C.
    § 1311(p), so long as the more permissive limitations “result in reasonable further progress toward
    the national goal of eliminating the discharge of all pollutants,” 
    Id. at §
    1311(b)(2)(A). The more
    permissive limitations under the Rahall Amendment are optional; nothing in the Amendment
    requires remining operations to take advantage of the more lenient pollution controls and nothing
    prevents remining operations from seeking permits under the otherwise generally applicable
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                   Page 33
    effluent-limitations. In sum, all that the Amendment does is permit the EPA to modify individual
    NPDES permits on a case-by-case basis to allow certain remining operations to take advantage of
    a more permissive effluent-limitations structure instead of abiding by the otherwise applicable
    national effluent-limitations. In other words, all coal mining polluters are required to comply with
    the more stringent effluent-limitation standards unless they qualify under the Rahall Amendment,
    which provides a more permissive structure for those meeting the qualifications set out by Congress.
    Thus, it is true that there is nothing in the Rahall Amendment that either grants the EPA
    additional authority to promulgate regulations for coal remining or restricts that authority. The
    Rahall Amendment merely presents another option for coal remining effluent-limitations. Thus, the
    EPA has the authority to issue regulations for the remining industry. That the EPA has the power
    to issue rules in this area does not ipso facto mean that any rule that it does issue is in compliance
    with the APA. This is where we must determine whether the Final Rule complies with the APA.
    In 1992, the EPA adopted the Final Rule. The Final Rule, as the majority notes, differs from
    the Rahall Amendment in four ways: First, the EPA’s rule includes a much broader definition of
    “coal mining operation” than the Rahall Amendment. As a result, the rule permits a greater number
    of operations to take advantage of the more permissive effluent-limitation standards. Second, the
    Final Rule adds “suspended solids” as an additional pollutant eligible for more permissive discharge
    treatment. Third, the Final Rule allows the EPA to issue discharge permits without setting numeric
    effluent-limitations. Fourth, the Final Rule applies a more stringent standard to commingled
    discharges.
    The EPA’s Decision to Set Effluent Reduction Attainable at Zero Is Arbitrary & Capricious
    Congress adopted the Clean Water Act to “restore and maintain the chemical, physical, and
    biological integrity of the Nation’s waters.” 33 U.S.C. § 1251; see also BP Exploration & Oil v.
    EPA, 
    66 F.3d 784
    , 789 (6th Cir. 1995); E.I. du Pont de Nemours & Co. v. Train, 
    430 U.S. 112
    (1977). “[T]he guiding star [of the Act] is the intent of Congress to improve and preserve the quality
    of the Nation’s waters.” Am. Petroleum 
    Institute, 540 F.2d at 1028
    . Congress’s original goal in
    promulgating the Act was to eliminate the discharge of all pollutants into navigable waters by the
    year 1985. See 33 U.S.C. § 1251(a)(1). Congress recognized, however, that the goal would not be
    achieved overnight. 
    Id. at 1031.
    Nevertheless, it adopted a system of regulation wherein
    increasingly stringent pollution controls would be imposed on polluters. BP Exploration & 
    Oil, 66 F.3d at 789
    (noting that the Clean Water Act “directs [the] EPA to institute progressively more
    stringent effluent discharge guidelines in stages”) (emphasis added); 
    Rybachek, 904 F.2d at 1283
    (“Through the Clean Water Act, Congress has directed the EPA to incorporate into the permits
    increasingly stringent technology-based effluent-limitations.”) (emphasis added).
    The EPA’s decision to set effluent reduction attainable at zero for remining areas, if not in
    direct conflict with congressional intent, at the very least raises a red flag. The Final Rule, while
    ensuring that remined sites are not further polluted, does little to further the Clean Water Act’s
    guiding star to improve the quality of our nation’s waterways. It also shirks the EPA’s duty to
    determine the degree of effluent reduction attainable. In fact, there is no evidence that the EPA ever
    explored the prospect of accomplishing more than preserving the status-quo in order to make
    “reasonable further progress toward the national goal of eliminating the discharge of all pollutants.”
    33 U.S.C. § 1311(b)(2)(A). By requiring remining operators to preserve only the status quo, the
    EPA undermines the potential for reduction that is available and required under the Clean Water
    Act. See Am. Petroleum 
    Institute, 540 F.2d at 1030
    (“The Act is ineffective unless somebody fixes
    effluent-limitations.”).
    No. 02-3628               Citizens Coal Council, et al. v. EPA                                                 Page 34
    As discussed above, we must be convinced that the EPA “examine[d] the relevant data and
    articulate[d] a satisfactory explanation for its action including a rational connection between the
    facts and the choices made.” State Farm Mut. Auto. Ins. 
    Co., 463 U.S. at 43
    . In my opinion, there
    is no evidence that the EPA considered the relevant technological data or sought data on the types
    of technology that would result in a reduction of effluent-limitations beyond that which existed
    when the land was abandoned. Further, the EPA has failed to articulate a satisfactory reason for why
    effluent reduction is not attainable. See also Natural Res. Def. Council, Inc. v. Costle, 
    568 F.2d 1369
    , 1380 (D.C. Cir. 1977) (noting that if numerical effluent-limitations are infeasible, the EPA
    may opt for “a gross reduction in pollutant discharge rather than the fine-tuning suggested by
    numerical limitations”). In fact, the D.C. Circuit rejected the EPA’s argument that, where it was
    infeasible to establish numerical effluent-limitations, the EPA could exclude certain point sources
    from the NPDES program, because the Clean Water Act gives the EPA a fair amount of flexibility
    in framing the permits to achieve its desired reduction of pollutant discharges. 
    Id. The EPA
    ,
    however, has not put forth any evidence as to why its desired reduction of pollutant discharges in
    this context is zero. But, to avoid being arbitrary and capricious in this context, an EPA rule must
    seek some sort of reduction in effluent-limitations or adequately explain why such a limitation is not
    attainable.
    We must hold that an agency rule is arbitrary and capricious if the agency “entirely failed
    to consider an important aspect of the problem.” State Farm Mut. Auto. Ins. 
    Co., 463 U.S. at 43
    .
    Part of the major problem that led Congress to enact the Rahall Amendment to encourage remining
    was the fact that abandoned mines were emitting pollutants at an unacceptable rate. The goal was
    to make remining economically feasible — thus, the environment benefits from pollution reduction
    and industry benefits from profitable remining. The EPA’s Final Rule, however, effectively subverts
    the goal of pollution reduction by demanding only the status quo and disregards the reason Congress
    initially got involved in the problem and delegated regulatory authority to the EPA.7 These
    regulations preserve   the status quo and, therefore, effectively freeze the lands at the pre-remining
    pollution rates.8
    In my opinion, these “leave it the way you found it” remining regulations also provide a
    disincentive to employ newer and better technologies. Many of these abandoned lands, specifically
    those that qualified under the Rahall Amendment, were abandoned prior to the mid-1970s. Thus,
    at the time they were mined, these operations utilized technology that was at least thirty years old
    at the time the EPA’s Final Rule was adopted. The Final Rule effectively tells remining operators
    not to make it any worse than it already is. Technology, however, has significantly advanced in the
    past thirty years and is certainly capable of outperforming the thirty-year-old technology and is
    certainly capable of reducing effluent-limitations. For whatever reason (because the EPA has not
    articulated one), the EPA has required remining operations only to utilize BMP such that the
    polluter’s effluent-limitations “[m]ay not exceed baseline loadings.” 40 C.F.R. § 432.72(b)(1).
    And, there is even an exception to this toothless standard if the EPA determines that measuring
    baseline loadings is “infeasible.” For these reasons, I believe that the EPA’s decision not to consider
    7
    According to the majority, “the EPA determined that pre-existing technology had resulted in little discharge
    reductions and therefore no environmental improvement.” Maj. Op. at 19. The EPA, however, stated that “[m]ore than
    ten years of remining under the Rahall Amendment have demonstrated success in improving abandoned mine land and
    acid mine damage.” 65 Fed. Reg. at 19445. Unfortunately, the EPA, based on that situation, decided the best remedy
    was to lower the environmental standards even further to allow more remining than the Rahall Amendment permitted
    without adhering to the purpose of the Clean Water Act. To that effect, this Final Rule fails.
    8
    The EPA suggests that remining by itself improves pollution rates at remining sites, however, there is no
    evidence in the record to support this claim. Moreover, if this fact is true then it begs the question why the EPA did not
    set the effluent reduction for remining sites at a numerical requirement higher than zero.
    No. 02-3628                Citizens Coal Council, et al. v. EPA                                                  Page 35
    the effluent-limitation reduction attainable is arbitrary and capricious, and I would invalidate the
    Final Rule.9
    The EPA’s Argument That Non-Numeric Effluent-Limitations are Infeasible is Without Merit and,
    therefore, the Limitations are Arbitrary and Capricious
    An “effluent-limitation” is “any restriction established . . . on the quantities, rates, and
    concentrations of chemical, physical, biological, and other constituents which are discharged from
    point source into . . . water.” 33 U.S.C. § 1362(11). The Federal Register states that the “EPA
    interprets the definition of ‘effluent-limitation’ in section 502 of the CWA to include non-numeric
    effluent-limitations where it is not feasible to establish numeric effluent-limitations.” 67 Fed. Reg.
    3378. Therefore, the EPA acknowledges that the Clean Water Act requires numerical effluent-
    limitations, unless it is not feasible to establish numeric effluent-limitations. This requirement is
    based on Congress’s findings that numerical effluent-limitations are the best way to quantifiably
    ensure industry compliance and to make reasonable further progress toward the goal of eliminating
    pollutants into the nation’s waters. Unnecessarily removing the numerical burden for remining areas
    makes it exceedingly difficult to determine whether reminers are complying with the Clean Water
    Act and naively places faith in the ability of the mining community to self-regulate.
    Assuming the correctness of the EPA’s determination that numeric effluent-limitations are
    required unless it is not “feasible” to establish such numeric limits requires us to determine what the
    term “feasible” actually means. Feasible, according to Black’s Dictionary means “capable of being
    done.” BLACK’S LAW DICTIONARY 609 (6th ed. 1990). The EPA has put forth no evidence in the
    record to substantiate a determination that establishing numerical effluent-limitations for remining
    is not capable of being done. I can sympathize with the suggestion that setting numerical limits may
    be difficult. Perhaps the EPA would argue that setting numerical guidelines for remining is more
    difficult than for virgin lands because there are more variations in lands that have been mined and
    abandoned. But, it is always “difficult” to determine the best numerical effluent-limitation, even
    for virgin sites and for other industries. The EPA, however, has not put forth any evidence that
    would distinguish remining from the other categories and subcategories of point sources that the
    EPA regulates, leading to the conclusion that setting numeric limits for remining is not capable of
    being done. This is not to say that the EPA is required to employ the same rigid numerical limitation
    for every remining site. The EPA could require a percentage effluent reduction to be shown in order
    to approve a NDPES permit. The EPA does not put forth any evidence why, even with variations
    site-by-site, it is inconsistent with the Clean Water Act to require certain percentage reductions in
    order to make reasonable further progress toward the elimination of discharges into the nation’s
    waters. Percentage reductions would also have the benefit of allowing some site-by-site deviations,
    thus maintaining some flexibility to the regulatory scheme. See Am. Petroleum 
    Institute, 540 F.2d at 1030
    (“The balance of general rule and narrow exceptions assures all possible uniformity without
    sacrifice of the flexibility needed to adjust for disparate plans in dissimilar circumstances.”).
    Likewise, the D.C. Circuit has stressed that “when numerical effluent-limitations are infeasible, [the]
    EPA may issue permits with conditions designed to reduce the level of effluent discharges to
    acceptable levels. This may well mean opting for a gross reduction in pollutant discharge rather
    than the fine-tuning suggested by numerical limitations.” Natural Res. Def. Council, Inc., 
    568 F.2d 9
               The majority, in section III-A-2-e, states that we have “overlook[ed] the whole point of not only the Final Rule,
    but the Rahall Amendment itself,” when in fact it is the majority who is committing the very same mistake as the EPA
    in this case. Maj. Op. at 18. The majority believes the EPA followed the Rahall Amendment by setting the baseline of
    pollution reduction at zero. The Amendment states that “in no event shall such a permit allow the pH level of any
    discharge, and in no event shall such a permit allow the discharges of iron and manganese, to exceed the levels being
    discharged from the remined area before the coal mining operation begins.” 33 U.S.C. § 1311(p)(2). This is the absolute
    minimum that the Amendment requires of reminers, but the EPA interpreted this to mean the maximum attainable
    pollution levels for remining operations were the levels at which they found the site. In sum, the Rahall Amendment
    gave an inch and the EPA took a mile. That is arbitrary and capricious.
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                     Page 36
    at 1380 (emphasis added). Thus, courts have recognized that where it is truly infeasible to set
    numeric limits, other types of limitations designed to achieve gross reductions are sufficient. Here,
    however, the EPA has not demonstrated either that fine-tuned numerical limits are infeasible or that
    some other type of gross reduction is not feasible either. Instead, the EPA has simply adopted a rule
    designed to maintain the status-quo.
    Thus, I would hold that the EPA acted arbitrarily and capriciously in failing to set numeric
    effluent-limitations or put forth evidence that numeric limitations are not capable of being imposed.
    Had the agency put forth sufficient evidence in the record to demonstrate that numeric limitations
    are infeasible, the Clean Water Act still requires some type of limit designed to achieve a gross
    reduction in effluent-limitations. The EPA’s decision to set non-numeric limitations that do nothing
    but preserve the status quo is arbitrary and capricious.
    The EPA’s Decision to Set Non-Numerical Effluent-Limitations is Arbitrary and Capricious
    The EPA’s Final Rule effectively substitutes best management practices or BMPs for
    numerical effluent-limitations. This substitution is arbitrary and capricious for two reasons:
    (1) under the Clean Water Act and the EPA’s own regulations, BMPs may be promulgated as
    supplemental to effluent-limitations, but may rarely if ever take the place of effluent-limitations; and
    (2) the Clean Water Act requires the EPA to consider numerous factors when establishing control
    measures that become BPT, BAT, and BCT, but there is no evidence in the record to support the
    claim that the EPA actually considered these factors in reaching its conclusion that BMPs were
    sufficient to satisfy BPT, BAT, and BCT.
    First, I agree with the original panel majority’s conclusion that the language of the Clean
    Water Act strongly suggests that the drafters of the Clean Water Act did not envision BMPs
    replacing numerical effluent-limitations. Specifically, section 304(e) of the Clean Water Act permits
    the EPA to publish best management practices that are “supplemental to any effluent-limitations.”
    33 U.S.C. § 1314(e). Black’s defines “supplemental” as “that which is added to a thing or act to
    complete it.” BLACK’S LAW DICTIONARY 1438 (6th ed. 1991). Thus, this provision of the Clean
    Water Act strongly indicates that BMPs are different from effluent-limitations, and although BMPs
    may be used to add to effluent-limitations, Congress did not intend for them to replace effluent-
    limitations. Congress was likely aware that BMPs are insufficient to result in reasonable further
    reduction in pollutant discharges.
    Furthermore, the EPA has interpreted the NPDES permit regulations to allow BMPs to be
    used in specific delineated circumstances. The majority opinion suggests that“[t]he EPA’s NPDES
    permit regulations reflect the EPA’s longstanding interpretation of the CWA as allowing BMPs to
    take the place of numeric effluent limitations under certain circumstances.” Maj. Op. at 16, n.18.
    This is an erroneous interpretation of the NPDES permit regulations. Rather, the regulations do
    reflect the EPA’s policy of allowing BMPs to supplement effluent-limitations but the regulations
    allow BMPs to replace effluent-limitations only under rare circumstances. Contrary to the majority’s
    assertions, these rare circumstances are not present here.
    According to 40 C.F.R. § 122(k), permits may include BMPs to control or abate the
    discharge of pollutants when: (1) “[a]uthorized under section 304(e) of the CWA for the control of
    toxic pollutants and hazardous substances from ancillary industrial activities”; (2) “[a]uthorized
    under section 402(p) of the CWA for the control of storm water discharges”; (3) “[n]umeric effluent-
    limitations are infeasible”; or (4) “[t]he practices are reasonable necessary to achieve effluent-
    limitations and standards or to carry out the purposes and intent of the CWA.” Only one of these
    circumstances, 40 C.F.R. §122(k)(3), allows BMPs to replace numeric effluent-limitations.
    Moreover, none of the facts here justify replacing effluent-limitations with BMPs under this
    regulation. The majority opinion suggests that the EPA’s “longstanding interpretation” is that BMPs
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                    Page 37
    may “take the place of numeric effluent-limitations under certain circumstances.” Even if this were
    true (and I dispute that it is), none of the circumstances in which this is allegedly permissible under
    section 122.44(k) are present here. The majority opinion’s suggestion misreads section 122.44(k),
    and its reading is not supported by the text of section 122.44(k). The text of three of the four
    subsections of 122.44(k) contemplates use of BMPs to supplement, not replace as the majority
    alleges, effluent-limitations. First, subsection 122.44(k)(1) suggests that BMPs may be used where
    section 304(e) authorizes such use. Section 304(e) states, of course, that BMPs, are permitted as
    “supplemental to any effluent-limitations.” Second, the majority opinion’s reading gains no support
    from subsection two. We are not dealing here with storm water draining. And, subsection four
    states that “[t]he [best management] practices are reasonably necessary to achieve effluent-
    limitations,” which clearly implies that effluent-limitations themselves exist and BMPs are
    reasonably necessary to help achieve those limitations. There is, in fact, only one circumstance
    under section 122.44(k) that contemplates a replacement of effluent-limitations with BMPs — that
    is where “[n]umeric effluent-limitations are infeasible.” As the preceding argument demonstrates,
    however, there is no evidence whatsoever to support the conclusion that numeric effluent-limitations
    are infeasible for coal remining. Even if the Clean Water Act authorizes BMPs to be used in place
    of effluent-limitations under certain rare circumstances, the EPA’s own regulation, section
    122.44(k), belies the suggestion that it is appropriate here. Thus, the EPA’s Final Rule is arbitrary
    and capricious for allowing BMPs to replace numeric effluent-limitations.
    The EPA’s Final Rule is also arbitrary and capricious because the Clean Water Act requires
    the EPA to consider numerous factors when establishing control measures that become BPT, BAT,
    and BCT, but there is no evidence in the record to support the claim that the EPA actually
    considered these factors in this case. Subsections 304(b)(1)(B), (b)(2)(B), and (b)(4)(B) require the
    EPA to consider a list of factors when establishing which control measures shall become BPT, BAT,
    and BCT. 33 U.S.C. § 1314(b)(1)(B), (b)(2)(B), (b)(4)(B). The factors in Subsections 304(b)(1)(B)
    and (b)(4)(B) include the age of the equipment and facilities involved, the process employed, the
    engineering aspects of the application of various types of control techniques, process changes, non-
    water quality environmental impact (including energy requirements), and other such factors as the
    Administrator deems appropriate. 33 U.S.C. § 1314(b)(1)(B), (b)(4)(B). Subsection 304(b)(2)(B)
    contains all of the above, and also includes an additional factor — “the cost of achieving such
    effluent reduction” after “process changes.” 33 U.S.C. § 1314(b)(2)(B). Although the EPA has
    broad discretion in determining the amount of weight it assigns to each of the factors, see Texas Oil
    & Gas Ass’n v. EPA, 
    161 F.3d 923
    , 928 (5th Cir. 1998) (citing Natural Res. Def. Council v. EPA,
    
    863 F.2d 1420
    , 1426 (9th Cir. 1988)), the EPA may not ignore a factor altogether, 
    id. at 934
    (holding
    that an agency abuses its discretion when it ignores a statutory factor).
    The record in this case, however, contains no evidence that the EPA analyzed these factors
    when concluding that BMPs statisfy BPT, BAT and BCT for the two new subcategories. The EPA’s
    Proposed Rule, located at 65 Fed. Reg. 19440 (April 11, 2000), lists the factors in its discussion of
    BPT for the Coal Remining Subcategory, 
    id. at 19450-51,
    but it contains no evidence in the form
    of scientific data or otherwise that the EPA actually considered the factors it listed. For instance,
    the Proposed Rule contains no discussion or even information about the age of the equipment or
    facilities or of “the engineering aspects of the application of various types of control techniques.”
    The Rule’s discussion of BCT and BAT are worse — they do not even bother to list the factors. 
    Id. at 19452.
    The Final Rule fares no better. See 67 Fed. Reg. 3370-3410 (Jan. 23, 2002).
    It cannot be disputed that the statute requires the EPA to include a “consideration” of the
    factors listed. The statute uses the language “shall include.” It cannot reasonably be contended that
    “shall” vests the EPA with discretion as to whether it actually considers the factors. The “assertion
    that ‘shall’ does not create a mandatory command simply flies in the face of standard interpretation.”
    United States v. Ostrander, 
    411 F.3d 684
    , 688 (6th Cir. 2005) (citing Lopez v. Davis, 
    531 U.S. 230
    ,
    241 (2001) ( “a mandatory ‘shall’ . . . impose[s] discretionless obligations”)); see also Lexecon Inc.
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                    Page 38
    v. Milberg Weiss Bershad Hynes & Lerach, 
    523 U.S. 26
    , 35 (1998) (concluding that “the mandatory
    ‘shall’ . . . normally creates an obligation impervious to judicial discretion”); Hewitt v. Helms, 
    459 U.S. 460
    , 471 (1983) (calling shall “language of an unmistakably mandatory character”).
    The nature of the rulemaking process requires that rules are based on sufficient evidence in
    the record, including scientific evidence. We will “defer in large part to [the] EPA’s scientific
    findings,” BP Exploration & 
    Oil, 66 F.3d at 792
    ; Baltimore Gas & Elec. Co. v. NRDC, Inc., 
    462 U.S. 87
    , 103 (1983) (recognizing that a reviewing court should be at its most deferential in
    reviewing an agency’s scientific conclusions in an area within the agency’s expertise), but will not
    do so when there are no scientific findings to defer to.
    Therefore, on the record before us, I cannot conclude that the EPA followed Congress’s
    statutory directives in determining the appropriate pollution controls to assign to point sources in
    the two new subcategories. Thus, the EPA’s Final Rule is arbitrary and capricious and not supported
    by substantial evidence within the meaning of the APA.
    The EPA’s Decision to Use Background Conditions As Effluent-Limitations Is Arbitrary and
    Capricious
    The Final Rule utilizes background conditions — those conditions that exist prior to the start
    of remining — as effluent-limitations. At the outset, this procedure is inherently suspect and in
    tension with Congress’s declaration that any remining is intended to improve the quality of water
    via reduced effluent-limitations. The use of background conditions is problematic for the additional
    reason that it results in different effluent-limitations for each site. This do-no-harm approach fails
    to approximate needed uniformity in the regulations without any persuasive reasoning from the
    EPA. The Clean Water Act does not mandate strict uniformity. But it also does not contemplate
    the type of site-by-site regulations adopted by the EPA in this case. See Am. Petroleum 
    Institute, 540 F.2d at 1030
    (“The balance of general rule and narrow exceptions assures all possible
    uniformity without sacrifice of the flexibility needed to adjust for disparate plans in dissimilar
    circumstances.”). The Tenth Circuit has recognized that: “The intent of Congress was to clean up
    the Nation’s waters. This cannot be done overnight. On the road to attainment of the no discharge
    objective some flexibility is needed. The extent of that flexibility may be determined by Congress
    through precise amendments to the Act .” 
    Id. at 1031-32.
    Further, in Texas Oil & Gas Ass’n v.
    EPA, 
    161 F.3d 923
    , 937-38 (5th Cir. 1998), the Fifth Circuit held that the Clean Water Act does not
    require absolute uniformity among effluent-limitations within a single category or subcategory of
    sources. The court also recognized “the textual mandate of the CWA that [effluent-limitation
    guidelines] be established for ‘categories and classes’ rather than individual point sources” and
    “agree[d] that Congress intended to foreclose plant-by-plant evaluation of facilities within a
    subcategory.” 
    Id. at 938-39.
    Here, the EPA’s decision to set effluent-limitations through sole
    reliance on site-specific factors runs afoul of this policy against site-specific regulations.
    I also find it arbitrary and capricious because the EPA has set forth no limiting principle to
    its rationale for relying on site-specific factors. There is nothing inherently different about coal
    remining that makes uniformity infeasible or undesirable. In fact, even with virgin lands, site-
    specific factors could be used because there will always be some difference in soil content and other
    factors. Congress’s policy, while not demanding absolute uniformity, certainly seeks uniformity in
    this context and the EPA has put forth no reasonable rationale for why site-specific factors should
    be determinative here (but not in other areas of regulation). In my opinion, the EPA has not
    “articulate[d] a satisfactory explanation for its action” and has “entirely failed to consider an
    important aspect of the problem.” State Farm Mut. Auto. Ins. 
    Co., 463 U.S. at 43
    . Thus, the EPA’s
    decision is arbitrary and capricious.
    No. 02-3628           Citizens Coal Council, et al. v. EPA                                    Page 39
    IV.
    The Western Alkaline Coal Mining Subcategory
    I would apply the same reasoning to invalidate the Final Rule with respect to the Western
    Alkaline Coal Mining Subcategory. These regulations again contain no numeric effluent-limitations
    for total suspended solids and no effluent-limitations at all for settleable solids and pH. Under the
    regulations, each polluter “must submit a site-specific Sediment Control Plan . . . designed to prevent
    an increase in the average annual sediment yield from pre-mined, undisturbed conditions.” 40
    C.F.R. § 434.82(a). Like the Pollution Abatement Plans required under the Coal Remining
    Subcategory, these Sediment Plans feature BMPs. 
    Id. Additionally, the
    operators are required to
    demonstrate the plans’ effectiveness “[u]sing watershed models.” 
    Id. at §
    434.82(b). These
    Sediment Control Plans, submitted by the operators, constitute the only effluent-limitation required
    for all three levels of technology, and they contain no numeric limits on sediment whatsoever. See
    
    id. at §
    434.82 (BPT); § 434.83 (BAT); § 434.84 (BCT). Beyond the initial models, there are no
    measurable checks on sediment levels at all.
    Thus, I would apply the same reasoning from the Coal Remining Subcategory to hold invalid
    the Rules here. The EPA’s decision to adopt non-numeric effluent-limitations is arbitrary and
    capricious as failing to consider an important aspect of the problem, State Farm Mut. Auto. Ins. 
    Co., 463 U.S. at 43
    — i.e., Congress’s goal of pollution reduction — and by failing to determine the
    degree of effluent reduction attainable. Additionally, the EPA failed to consider the factors required
    of it by the Clean Water Act in deciding which control measures would become BPT, BAT, and
    BCT. For these reasons, I would invalidate the Rule and respectfully dissent from the majority
    opinion’s decision to sustain it.
    V.
    In conclusion, the EPA’s new remining regulations violate the APA in that they are arbitrary
    and capricious and also not in accordance with the law. The regulations irrationally violate the spirit
    of the Clean Water Act in that they are not designed to improve the quality of our nation’s waters.
    This conclusion is borne out in the EPA’s abdication of its regulatory responsibility by eliminating
    numerical effluents limitations in favor of weakened requirements. I can only hope that this
    regulation represents an isolated regulatory decision by the EPA and not a sign of things to come
    in future environmental regulations. As political winds shift, so too sometimes do administrative
    agency’s priorities. Here, however, with respect to the Clean Water Act, Congress’s intent has
    remained fixed and clear. See Am. Petroleum 
    Institute, 540 F.2d at 1028
    (“[T]he guiding star [of
    the Clean Water Act] is the intent of Congress to improve and preserve the quality of the Nation’s
    waters.”). The EPA’s Final Rule, which lessens the controls on polluting industries and inflicts a
    self-imposed wound on the regulatory authority of the EPA, is arbitrary and capricious under the
    Administrative Procedure Act. I would invalidate the Rule and therefore respectfully dissent from
    the majority’s misguided decision to uphold it.