Full opinion text
Opinion PER CURIAM. Concurring opinion filed by Circuit Judge RANDOLPH. PER CURIAM: These consolidated petitions present a multifarious challenge to Environmental Protection Agency (“EPA”) regulations promulgated under the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (“CERCLA”), 42 U.S.C. §§ 9601-9675, as amended by the Superfund Amendments and Reauthorization Act of 1986 (“SARA”), Pub.L. No. 99-499, 100 Stat. 1613. The regulations under review are portions of the National Oil and Hazardous Substances Pollution Contingency Plan, 40 C.F.R. Part 300, commonly known as the “NCF.” Glossary of Acronyms ARAR Applicable or Relevant and Appropriate Requirements CERCLA. Comprehensive Environmental Response, Compensation, and Liability Act of 1980 EPA Environmental Protection Agency FS Feasibility Study J.D.A. Joint Deferred Appendix MCL Maximum Contaminant Level MCLG Maximum Contaminant Level Goal MOCO Missouri Coalition for the Environment NCP National Contingency Plan NIH National Institutes of Health OMB Office of Management and Budget O & M Operations and Maintenance PRP Potentially Responsible Party RI Remedial Investigation ROD Record of Decision SARA Superfund Amendments and Reau-thorization Act of 1986 SDWA Safe Drinking Water Act SMOA Superfund Memorandum of Agreement I Before Congress created the Environmental Protection Agency (“EPA” or “the Agency”), and long before Congress enacted the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (“CERCLA”), 42 U.S.C. §§ 9601-9675, there was a National Contingency Plan (“NCP”). In 1968, a group of federal agencies developed the first NCP, which was a multi-agen-cy strategy for dealing with environmental disasters. See Freedman, Proposed Amendments to the National Contingency Plan: Explanation and Analysis, 19 Envtl.L.Rep. 10,103, 10,105-06 (1989). In 1970, Congress incorporated the NCP into the Federal Water Pollution Control Act, 33 U.S.C. §§ 1251— 1376, and pursuant to its directive, the President issued the first published NCP. Water and Environmental Quality Improvement Act of 1970, Pub.L. No. 91-224, 84 Stat. 91, § 102 (1970); 35 Fed.Reg. 8508 (1970). The NCP, which acquired its current name — the National Oil and Hazardous Substances Pollution Contingency Plan, 36 Fed.Reg. 16,215 (1971) — in 1971, was revised a number of times throughout the 1970s. See 37 Fed. Reg. 2808 (1972); 38 Fed.Reg. 21,888 (1973); 45 Fed.Reg. 17,832 (1980). By 1980, a comprehensive NCP was in place, although it applied only to discharges into waters regulated by the Clean Water Act. Id. “It did not apply to releases to groundwater or soil, and it did not provide authority or funding for long-term' federal response to chronic hazards.” Freedman, supra, 19 Envtl. L.Rep. at 10107. CERCLA came next. Enacted in 1980, CERCLA provided “for liability, compensation, cleanup, and emergency response for hazardous substances released into the environment and the cleanup of inactive waste disposal sites.” Pub.L. No. 96-510, 94 Stat. 2767, 2767. We have summarized its general scheme in previous decisions. See, e.g., Ohio v. United States Dep’t of Interior, 880 F.2d 432, 438-40 (D.C.Cir.), reh’g denied, 897 F.2d 1151 (1989) (en banc); Ohio v. EPA 838 F.2d 1325, 1327-29 (D.C.Cir.1988). Of particular importance to this case is the prominent role of the NCP under CERCLA Section 104(a)(1) of CERCLA authorizes the President “to act, consistent with the national contingency plan, to remove or arrange for the removal of, and provide for remedial action relating to such hazardous substance, pollutant, or contaminant at any time ..., or take any other response measure consistent with the national contingency plan which the President deems necessary to protect the public health or welfare or the environment.” 42 U.S.C. § 9604(a)(1). The NCP thus “provide[s] the organizational structure and procedures” for responding to hazardous waste threats. 40 C.F.R. § 300.1. It is the means by which EPA implements CERCLA. When Congress enacted CERCLA in 1980, it directed the President to revise and republish the NCP in light of the new law. 42 U.S.C. § 9605(a). Pursuant to section 115 of CERCLA, the President assigned EPA the responsibility of amending the NCP. See 42 U.S.C. § 9615; Exec. Order No. 12,316, 46 Fed.Reg. 42,237 (1981); Exec. Order No. 12,-580, 52 Fed.Reg. 2923 (1987). In 1982, EPA issued a new version of the NCP. 47 Fed. Reg. 31,180 (1982). EPA revised the NCP again in 1985. 50 Fed.Reg. 47,912 (1985). When Congress passed the Superfund Amendments and Reauthorization Act of 1986 (“SARA”), Pub.L. No. 99-499, 100 Stat. 1613, which significantly revised the statute, Congress directed the President to revise the NCP again to reflect the changes in CERC-LA. 42 U.S.C. § 9605(b). EPA issued these revisions to the NCP in 1990. 55 Fed.Reg. 8666 (1990). Petitioners, whom we shall call “the States,” include both states and private parties contending that EPA’s changes to the NCP in 1985 and 1990 are inconsistent with the requirements of CERCLA. The petitions for review challenge two general categories of NCP provisions. One category involves claims that the NCP unlawfully diminishes the level of environmental protectiveness in the remedy selection process and cleanup provisions of CERCLA (These claims are resolved in Parts II, III, and IV of the opinion.) The second category involves claims that the NCP improperly limits the States’ participation in the cleanup process while increasing their financial burden. (These claims are resolved in Part V of the opinion.) The specific provisions of CERC-LA and the NCP at issue in this case will be discussed in the portion of the opinion analyzing petitioners’ claims regarding those provisions. II The States first challenge several elements of the NCP definition of legally “applicable” or “relevant and appropriate” environmental standards, known as “ARARs.” CERCLA does not define ARARs, but the statute does require that remedial actions at Superfund sites result in a level of cleanup or standard of control that at least meets the legally applicable or otherwise relevant and appropriate federal (or stricter state) requirements. 42 U.S.C. § 9621(d)(2)(A). The NCP defines “applicable requirements” as follows: Applicable requirements means those cleanup standards, standards of control, and other substantive requirements, criteria, or limitations promulgated under federal environmental or state environmental or facility siting laws that specifically address a hazardous substance, pollutant, contaminant, remedial action, location, or other circumstance found at a CERCLA site. Only those state standards that are identified by a state in a timely manner and that are more stringent than federal requirements may be applicable. 40 C.F.R. § 300.5. “Relevant and appropriate requirements” are those substantive requirements that, while not “applicable,” nonetheless “address problems or situations sufficiently similar to those encountered at the CERCLA site that their use is well suited to the particular site.” Id. A Does the NCP definition of ARARs as “substantive” requirements violate CERCLA? The States claim that the NCP definition of ARARs is contrary to CERCLA because it excludes “procedural” requirements, such as recordkeeping and reporting to the government, by inserting the word “substantive” into the definition. The States argue that limiting ARARs to substantive requirements is contrary to the plain language of CERCLA because the statute itself does not distinguish between substantive and procedural requirements. They also contend that the definition is inconsistent with congressional intent because the SARA legislative history gives no indication that Congress intended for ARARs to be limited to substantive requirements. The States argue in the alternative that EPA’s distinction between substantive and procedural requirements is irrational. The States are correct that CERCLA does not explicitly draw a line between substantive and procedural requirements, but neither does the statutory language clearly forbid the NCP distinction. In fact, as the following discussion indicates, an application of traditional tools of statutory construction, see NLRB v. United Food & Commercial Workers Union, Local 23, 484 U.S. 112, 123, 108 S.Ct. 413, 416, 98 L.Ed.2d 429 (1987); Natu ral Resources Defense Council v. Reilly, 983 F.2d 259, 266 (D.C.Cir.1993), strongly suggests that CERCLA is concerned only with substantive environmental requirements. In any case, the NCP limitation of ARARs to substantive standards certainly represents a reasonable and permissible construction of the statute. See Chevron v. Natural Resources Defense Council, 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). We do not dwell in our analysis on the question of which of Chevron’s two prongs best resolves this issue. In limiting ARARs to procedural requirements, EPA reasonably interprets CERC-LA’s reference to “a level or standard of control” to be directed at those environmental laws governing “how clean is clean” — that is, the level or degree of cleanup required to remedy various types of toxic contamination. The CERCLA section at issue, section 121(d), is titled “Degree of cleanup,” and it talks of standards that apply “to any hazardous substance, pollutant or contaminant,” 42 U.S.C. § 9621(d)(2)(A), not of standards that apply more generally to a site or a party executing a cleanup. Moreover, the only specific requirements explicitly set out in the statute are substantive standards such as Maximum Contaminant Levels established in the Safe Drinking Water Act and Federal Water Quality Criteria established in the Clean Water Act. Finally, contrary to the States’ claim, the SARA Conference Report explicitly states that “[n]ew section 121(d) establishes the substantive standards that remedial actions ... must meet.” H.R.Conf. Rep. No. 962, 99th Cong., 1st Sess. (1985), U.S. Code Cong. & Admin. News 1986, pp. 2835, 3339. The States are surely correct that the procedural requirements of various environmental statutes are intended to ensure that the substantive contaminant levels are met. However, this does not compel EPA to impose these requirements under CERCLA The language and structure of section 121(d) strongly support, if not compel, the EPA interpretation. The NCP represents at the very least a permissible construction of CERCLA within the dictates of Chevron. B. Does the NCP improperly restrict the meaning of state ARARs to standards that are generally applicable and legally enforceable? The States also claim that the NCP construction of the statutory term “promulgated” is inconsistent with CERCLA As noted supra p. 1526, CERCLA requires that Superfund remedial actions result in a level of cleanup that at least meets federal, or stricter state, ARARs. 42 U.S.C. § 9621(d)(2)(A). The statute contains an additional requirement with regard to state standards: they must be “promulgated ... under a State environmental or facility siting law” in order to be considered as possible ARARs. 42 U.S.C. § 9621(d)(2)(A)(ii). CERCLA does not define “promulgated,” but the NCP interprets the term to mean “standards [that] are of general applicability and are legally enforceable.” 40 C.F.R. § 300.400(g)(4). None of the States’ arguments establishes that EPA’s definition is an impermissible construction of this admittedly undefined term. Under Chevron, EPA need not establish that the statute compels its regulation. Where congressional intent on the precise question at issue is unclear, it is enough that the Agency’s construction is reasonable. Chevron v. Natural Resources Defense Council, 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). EPA’s definition of “promulgated” clearly meets this standard. The States claim that the ordinary meaning of the term “promulgated” precludes the NCP’s narrow definition. However, the dictionary definitions that the States cite— which include notions such as “official announcement” and “to make ... obligatory”— are perfectly consistent with the NCP requirements of general applicability and legal enforceability. Neither the absence of clear legislative history, nor the fact that the word sometimes has a broader meaning, demonstrates that the NCP definition is impermissible. The States also argue that another CERC-LA provision, allowing the President to waive ARARs that “the State has not consistently applied,” 42 U.S.C. § 9621(d)(4)(E), indicates that EPA carries the burden of proving inconsistent application by the State if it decides to waive an ARAR. The NCP definition of “promulgated,” the States argue, shifts the burden to the States to prove the general applicability of a state standard before it will be adopted as an ARAR. This argument is unavailing because the NCP definition and the cited CERCLA provision are perfectly consistent. Under the NCP definition, a standard must be generally applicable on its face, and if so, the standard is a potential ARAR. However, if such generally applicable standard is not applied consistently, then the standard may be waived under section 9621(d)(4)(E). The States’ remaining arguments on this point merely suggest alternative reasonable interpretations of the statute. The States suggest different language that Congress might have used to indicate clearly its authorization of EPA’s approach. However, just as the statute does not compel EPA’s interpretation, neither does the absence of clear language render the Agency’s approach impermissible. Furthermore, the inclusion in CERCLA of the terms “standards,” “criteria,” and “limitations” in addition to “requirements” does not, as the States suggest, necessarily indicate a broader class of state rules than those generally applicable and legally enforceable. Finally, the States’ attack on EPA’s allegedly inconsistent uses of the term must be rejected. EPA’s definition of “promulgate” is limited to the specific context of state requirements, and the Agency is defining an ambiguous term inserted in the statute by Congress. See 40 C.F.R. § 300.-400(g)(4). EPA is not acting inconsistently by using the term differently from its use in other contexts. C. Does the NCP improperly restrict the meaning of federal ARARs to those “promulgated” under federal environmental laws? The States also object to the NCP definition of ARARs insofar as it is limited to requirements “promulgated under federal ” environmental laws. 40 C.F.R. § 300.5 (emphasis added). The States argue that in setting out possible ARARs, CERCLA includes the word “promulgated” in reference to state standards, but not federal standards. Compare 42 U.S.C. § 9621(d)(2)(A)® (“any standard ... under any Federal environmental law”) and 42 U.S.C. § 9621(d)(2)(A)(ii) (“any promulgated standard ... under a State environmental or facility siting law”) (emphasis added). Thus, argue the States, the NCP is contrary to CERCLA insofar as it requires that federal standards must be promulgated to be considered as possible ARARs. We do not reach the merits of this argument because the States waived the claim by failing to raise it during rulemaking proceedings before the Agency. Linemaster Switch Corp. v. EPA 938 F.2d 1299, 1308 (D.C.Cir.1991); Washington Ass’n for Television & Children v. FCC, 712 F.2d 677, 680 (D.C.Cir.1983). The States argue that the court should exercise its discretion to consider this issue despite the States’ failure to raise it below because the policies behind the waiver rule would not be frustrated if the court were to address the merits in this case. We disagree. The States point to some of the purposes of the waiver doctrine — to allow an administrative agency to make a factual record and exercise its discretion or apply its expertise, see McKart v. United States, 395 U.S. 185, 193-94, 89 S.Ct. 1657, 1662-63, 23 L.Ed.2d 194 (1969) — and argue that these concerns are not implicated here because the States raise a purely legal challenge to the NCP. However, with the possible exception of developing a factual record, these concerns are relevant to an agency’s legal interpretation of a statute which it is implementing. The notion of deference to agency interpretations of law embodied in Chevron is founded on just such concerns. See Chevron, 467 U.S. at 843-45, 104 S.Ct. at 2782-83. Furthermore, the waiver doctrine is also concerned with notions of agency autonomy and judicial efficiency. The doctrine promotes agency autonomy by according the agency an opportunity to discover and correct its own errors before judicial review occurs. Judicial efficiency is served because issues that are raised before the agency might be resolved without the need for judicial intervention. McKart, 395 U.S. at 195, 89 S.Ct. at 1663. The efficiency concern is especially germane to this challenge to the NCP, involving an extremely complex rule-making in which a multitude of issues might be raised for the first time before this court in the absence of the waiver doctrine. The States also point out that this court has “excused the exhaustion requirements for a particular issue when the agency has in fact considered the issue,” Natural Resources Defense Council v. EPA, 824 F.2d 1146, 1151 (D.C.Cir.1987), but they offer no evidence that EPA actually considered an objection to the limitation of ARARs to “promulgated” federal standards. Neither the States nor any other party raised an objection to the use of the word “promulgated” with respect to federal environmental standards, and EPA therefore had no opportunity to consider the issue. Finally, the States argue that this issue presents a matter of great public importance worthy of allowing an exception to the waiver doctrine. See Foundation on Economic Trends v. Heckler, 756 F.2d 143, 156 (D.C.Cir.1985). In Foundation, this court decided the level of environmental review required of the National Institutes of Health (“NIH”) before it approved the first deliberate release of genetically engineered, recombinant-DNA-containing organisms into the open environment. Although the plaintiffs had failed to raise their objections to the release during the period of NIH review, the court nonetheless upheld the district court’s decision to address the claims because of the grave public importance of insuring appropriate environmental review “of a new technology with unknown environmental consequences.” Id. Of course, the public health that CERCLA and the NCP are aimed at protecting is also an extremely important concern. But the choice between two alternative readings of the CERCLA provision at issue here is not so critical to the overall scheme. The States present no convincing argument that limiting ARARs to promulgated federal standards will compromise CERCLA’s health protection goals or is otherwise of such gravity as to warrant departure from settled waiver principles. D. Does the NCP improperly fail to apply zero-level Maximum Contaminant Level Goals (“MCLGs”) as ARARs? The States challenge EPA’s decision that Maximum Contaminant Level Goals (“MCLGs) established under the Safe Drinking Water Act (“SDWA”), 42 U.S.C. §§ 300f to 300j-26, do not have to be attained for contaminants whose MCLG has been set at a level of zero. 40 C.F.R. § 300.430(e)(2)(i)(C). The States contend that EPA lacks authority to depart from a statutory requirement to achieve MCLGs, and in the alternative, that even if EPA possesses this authority, it has failed to provide a reasoned basis for its departure. The SDWA is specifically referenced in section 121(d)(2)(A) of CERCLA as one of the federal laws containing ARARs for Superfund cleanups. 42 U.S.C. § 9621(d)(2)(A). The SDWA identifies two standards for exposure to contaminants. The first, Maximum Contaminant Level Goals (“MCLGs”), are generally unenforceable goals that reflect the level for a given contaminant at which “no known or anticipated adverse effects on the health of persons occur and which allows an adequate margin of safety.” 42 U.S.C. § 300g-l(b)(4). Many MCLGs for carcinogens are set at zero. 55 Fed.Reg. 8750 (1990). The second type of standards, Maximum Contaminant Levels (“MCLs”) — the actual maximum permissible concentration levels under the SDWA — must be set as close as “feasible” to their corresponding MCLGs, taking into account available technology and cost. 42 U.S.C. § 300g-l(b)(4)-(5). While MCLGs are unenforceable under the SDWA, section 121 of CERCLA converts them into enforceable - goals, providing: Such remedial action shall require a level or standard of control which at least attains Maximum Contaminant Level Goals established under the Safe Drinking Water Act ... where such goals or criteria are relevant and appropriate under the circumstances of the release or threatened release. 42 U.S.C. § 9621(d)(2)(A). Consistent with this requirement, the NCP generally requires the attainment of MCLGs. 40 C.F.R. § 300.430(e)(2)(i)(B). When the MCLG for a contaminant has been set at a level of zero, however, the NCP requires only that the MCL be attained. In essence, EPA has made a categorical determination that MCLGs set at a level of zero are never “relevant and appropriate under the circumstances” of a release. This determination was based on EPA’s conclusion “that it is impossible to detect whether ‘true’ zero has actually been attained.” 55 Fed.Reg. 8752 (1990). During rulemaking to promulgate MCLGs under the SDWA, EPA “emphasized that ... zero is not a measurable level in scientific terms.” 50 Fed.Reg. 46,884, 46,896 (1985). “Due to limitations in analytical techniques, it will always be impossible to say with certainty that the substance is not present. In theory, RMCLs [Recommended Maximum Contaminant Levels] at zero will always be unachievable (or at least not demonstrable).” 49 Fed. Reg. 24,330, 24,347 (1984). The States contend that EPA’s decision concerning zero-level MCLGs is inconsistent with CERCLA’s mandate that all remedial actions attain MCLGs. This argument ignores the full language of the section, which imposes the requirement “where such goals ... are relevant and appropriate under the circumstances of the release or threatened release.” 42 U.S.C. § 9621(d)(2)(A). This language leaves EPA with discretion to determine when MCLGs are relevant and appropriate. The States contend, though, that such discretion cannot be exercised in a categorical manner, but instead must be based on a case-specific determination at individual sites. Hence, there is no reason for EPA to make an individualized determination of what they have concluded can never be relevant and appropriate. The States also contend that even if EPA has discretion to conclude that zero-level MCLGs are never relevant and appropriate, it has not justified the decision to do so in this case. But EPA articulated a number of justifications, see 55 Fed.Reg. 8750-52 (1990), and we find its reliance on the fact that true zero levels can never be detected to provide adequate support for the Agency’s decision. As we understand EPA’s scientific analysis, one can never prove a true zero level. If the measuring device indicates zero, this shows only that the device is not sufficiently sensitive to detect the presence of any contaminants. It does not show the total absence of the contaminants. In other words, if one asserts that zero contaminants are present, this can be falsified by showing the presence of some detectable level, but it can never be shown to be true. EPA chose to set MCLGs for carcinogens at zero under the SDWA because they “are goals which may or may not be practically achievable and the practicality of these goals should be factored into the MCLs,” not the MCLGs. 50 Fed.Reg. 46,896 (1985). In contrast, EPA concluded that “ARARs must be measurable and attainable since their purpose is to set a standard that an actual remedy will attain.” 55 Fed.Reg. 8752 (1990). The States do not contest EPA’s scientific conclusion that zero-level MCLGs are not achievable. Instead, they argue that EPA could select a method of measurement approximating zero by setting “a goal of achieving the analytical detection limits for specific carcinogens.” Final Amended Joint Brief of Petitioning States at 68. That EPA could do this, however, does not mean it is required to do so. Section 121 requires the selection of MCLs where MCLGs are unattainable. That is what the NCP does. That conclusion is reasonable given EPA’s discretion to determine when ARARs are relevant and appropriate. Ill The next set of challenges by the States addresses a variety of issues concerning remedy selection: the role of cost-benefit analysis in remedy selection; the requirement that selected remedies are permanent to the maximum extent practicable; the use of a cancer risk range in remedy selection; and the requirement of five-year review of certain remedial actions. A. Does the NCP establish an improper cost-benefit analysis in the remedy selection process? Section 121 of CERCLA, added by SARA, requires the selection of remedial actions “at a minimum which assures protection of human health and the environment,” 42 U.S.C. § 9621(d)(1). A though a different provision of section 121 requires the selection of remedial actions that are also cost-effective, 42 U.S.C. § 9621(b)(1), the States interpret section 121(d)(1) to prohibit EPA from considering the cost of a remedial action when it determines the level of protectiveness to be achieved by that remedial action. EPA is in full agreement with the States’ interpretation of § 121(d)(1). See 55 Fed. Reg. 8726 (1990). The States contend, however, that two provisions in the NCP implicitly authorize the use of cost-benefit analysis, thereby permitting cost to be considered in determining the level of protectiveness to be achieved by a remedial action. In making this argument, the States distort the language of the NCP, which is carefully structured so “that protection of human health and the environment will not be compromised by other selection factors, such as cost.” Id. The States first point to a provision in the NCP authorizing EPA to balance nine different criteria, including both protection of human health and cost, in selecting a remedy. 40 C.F.R. § 300.430(f)(l)(i)(A). But while the NCP identifies nine criteria to be used in selecting a remedy, all of the criteria are not given equal weight. Instead, they are divided into three classifications: threshold criteria, primary balancing criteria, and modifying criteria. Under this structure, “[o]verall protection of human health and the environment and compliance with ARARs (unless a specific ARAR is waived) are threshold requirements that each alternative must meet in order to be eligible for selection.” 40 C.F.R. § 300.430(f)(l)(i)(A). EPA explained in the preamble to the NCP that remedial alternatives “must be demonstrated to be protective ... in order to be eligible for consideration in the balancing process by which the remedy is selected.” 55 Fed.Reg. 8726 (1990). The identification of threshold criteria therefore undermines the States’ claim that by listing nine criteria, the NCP permits the level of protectiveness to be affected by cost. The States also point us to the NCP’s definition of “cost-effectiveness,” which states that “[a] remedy shall be cost-effective if its costs are proportional to its overall effectiveness.” 40 C.F.R. § 300.430(f)( 1)(ii)(D). The States contend that this language actually authorizes the use of cost benefit analysis. In making this argument, though, the States ignore the first sentence of the same section of the NCP that they are challenging. It states: “Each remedial action shall be cost-effective, provided that it first satisfies the threshold criteria set forth in § 300.-430(f)(l)(ii)(A) and (B).” Id; see also 55 Fed.Reg. 8727 (1990). Thus, consistent with the creation of threshold criteria, the NCP explicitly prohibits consideration of costs in the manner complained of by the States. B. Does the NCP improperly fail to require the selection of permanent remedies to the maximum extent practicable? The States next argue that the NCP is inconsistent with section 121(b)(l)’s requirement that the President select remedial actions “that utilize[] permanent solutions ... to the maximum extent practicable.” 42 U.S.C. § 9621(b)(1). The NCP classifies permanence as one of the five primary balancing criteria, along with reduction of toxicity, mobility, or volume; short-term effectiveness; implementability; and cost. 40 C.F.R. § 300.430(f)(l)(i)(B). The States reason that because the selection of permanent remedies “is one of the overarching statutory principles of remedy selection under CERCLA,” Final Amended Joint Brief of Petitioning States at 27, the other balancing .criteria, particularly cost, should play no role in EPA’s determination whether a permanent remedy is to be selected. In essence, the States would like permanence.to be treated as an additional threshold criterion that must be evaluated independently of cost. The flaw in the States’ argument is in the premise that permanence is an overarching statutory principle. This premise is not supported by the statutory language. Section 121(b)(1), which the States rely upon, requires the President to “select a remedial action that is protective of human health and the environment, that is cost effective, and that utilizes permanent solutions and alternative treatment technologies or resource recovery technologies to the maximum extent practicable.” 42 U.S.C. § 9621(b)(1). The statutory language places as much emphasis on the selection of cost-effective remedies as it does on the selection of permanent remedies. Although the NCP elevates protection of human health and the environment to a threshold criterion, a different provision in section 121 provides the basis for that treatment. 42 U.S.C. § 9621(d)(1); see supra p. 1531. But there is nothing in section 121 to suggest that selecting permanent remedies is more important than selecting cost-effective remedies. The States offer two responses. The first is a decision defining “practicable” as “ ‘possible to practice or perform’ or ‘capable of being put into practice, done, or accomplished.’” Ashton v. Pierce, 541 F.Supp. 635, 641 (D.D.C.1982) (quoting Webster’s Third New International Dictionary (1963)), aff'd, 716 F.2d 56 (D.C.Cir.1983); cf. American Textile Mfrs. Inst. v. Donovan, 452 U.S. 490, 508-09, 101 S.Ct. 2478, 2490-91, 69 L.Ed.2d 185 (1981). The Ashton court had before it a statute requiring a single goal to be achieved to the extent practicable. A 1973 amendment to the Lead-Based Paint Poisoning Prevention Act required the Secretary of Housing and Urban Development to “establish procedures to eliminate as far as practicable the hazards of lead paint poisoning with respect to any existing housing which may present such hazards and which is covered by an application for mortgage insurance or housing assistance payments under a program administered by the Secretary.” 42 U.S.C. § 4822. The regulations at issue in Ashton authorized the use of cost-benefit analysis in determining the appropriate remedy and the court found no basis for this approach in the statute. In contrast, section 121(b)(1) of CERCLA mandates the achievement of multiple goals. If EPA were to require the selection of permanent remedies whenever possible, it would be ignoring the statutory mandate to select cost-effective remedies. The States’ second response relies on comments made from the floor of Congress. We have frequently cautioned against placing much weight on such statements. See, e.g., Colorado v. United States Dep’t of Interior, 880 F.2d 481, 490 (D.C.Cir.1989); International Bhd. of Elec. Workers, Local Union No. 474 v. NLRB, 814 F.2d 697, 717 (D.C.Cir.1987); Northern Colorado Water Conservancy Dist. v. Federal Energy Regulatory Commission, 730 F.2d 1509, 1519 (D.C.Cir.1984). That caution is certainly warranted here. For every set of comments supporting the States’ position, there is another set of comments supporting the opposite position. See, e.g., 132 Cong.Rec. 29,-719-20 (1986) (statement of Rep. Lent); id. at 29,743 (statement of Rep. Eckart, Chairman of Conference Committee). The States argue in the alternative that even if permanence is not treated as a threshold criterion, the NCP should at least place special emphasis on the selection of permanent remedies. But the NCP does exactly that. It requires that “[t]he balancing [of alternative remedies] shall emphasize long-term effectiveness and reduction of toxicity, mobility, or volume through treatment.” 40 C.F.R. § 300.430(f)(l)(i)(E). In the preamble, EPA explained that “[t]hese two criteria are given primary consideration in the rule and preamble when analyzing the relative merits of the alternatives. These criteria will be the most important, decisive factors in remedy selection when the alternatives perform similarly with respect to the other balancing criteria.” 55 Fed.Reg. 8725 (1990). Given the statutory requirement to achieve a number of competing goals, EPA’s decision concerning how much emphasis to place on the selection of permanent remedies is a reasonable one. C. Does the NCP cancer risk range improperly fail to protect human health and the environment without regard to cost? The States next challenge EPA’s use of a cancer risk range between 10“6 and 10“4 in the NCP, arguing that an exposure level greater than 10“6 is never appropriate. A 10“4 risk subjects the surrounding population to an increased lifetime cancer risk of 1 in 10,000. A 10“6 risk subjects the surrounding population to an increased lifetime cancer risk of 1 in 1,000,000. When EPA develops objectives for a remedial action at a site, it selects a remediation goal that “establish[es] acceptable exposure levels that are protective of human health.” 40 C.F.R. § 300.430(e)(2)(i). EPA attempts to use health-based ARARs to set the goal, but if ARARs are nonexistent or unsuitable for use, EPA establishes the goal based on criteria in the NCP. 55 Fed.Reg. 8712 (1990). “For known or suspected carcinogens, acceptable exposure levels are generally concentration levels that represent an excess upper bound lifetime cancer risk to an individual of between 10“6 and 10“4....” 40 C.F.R. § 300.-430(e)(2)(i)(A)(2). The NCP expresses a preference for remedial actions that achieve a level of 10“6 however, the ultimate decision depends on a balancing of nine criteria, including cost. Id; 55 Fed.Reg. 8718 (1990). The States contend that by permitting cost to play a role in determining the level of exposure, the cancer risk range fails to meet the requirement in § 9621 that remedial actions be “protective of human health.” 42 U.S.C. § 9621(b)(1); see also 42 U.S.C. § 9621(d)(1). The States’ argument necessarily depends, though, on the notion that an exposure level greater than 10“6 is not protective of human health. CERCLA requires the selection of remedial actions “that are protective of human health,” not as protective as conceivably possible. A “risk range of 10“4 to 10“6 represents EPA’s opinion on what are generally acceptable levels.” 55 Fed.Reg. 8716 (1990). Although cost cannot be used to justify the selection of a remedy that is not protective of human health and the environment, it can be considered in se-leeting from options that are adequately protective. The States also argue that the actual risk range selected is not adequately protective. EPA concluded, though, that all levels of exposure within the risk range are protective of human health. Id. EPA has used 10 “4 as an upper bound for establishing risk levels in the past, see 53 Fed.Reg. 51,394, 51,426 (1988), and “[mjany ARARs, which Congress specifically intended be used as cleanup standards at Superfund sites, are set at risk levels less stringent than 10“6,” 55 Fed.Reg. 8717 (1990). The States offer no evidence challenging EPA’s position that 10“4 represents a safe level of exposure, and in any event, we give EPA’s findings on this point significant deference. See New York v. EPA, 852 F.2d 574, 580 (D.C.Cir.1988), cert. denied, 489 U.S. 1065, 109 S.Ct. 1338, 103 L.Ed.2d 809 (1989). The States also argue that EPA failed to justify the use of a range, instead of a single point. But EPA explained its decision to use a range. While “[t]he use of 10“6 expresses EPA’s preference for remedial actions that result in risks at the more protective end of the risk range,” 55 Fed.Reg. 8718 (1990), the Agency is also required to consider other factors in selecting an appropriate remedy. “Factors related to exposure, uncertainty and technical limitations may justify modifications of initial cleanup levels that are based on the 10“6 risk level.” Id. A flexible approach to developing remedial goals is justified by the multiple statutory mandates of CERCLA, so long as EPA meets the statutory requirement of protectiveness. The States’ final argument is that we should not defer to EPA’s judgment because of OMB’s role in developing the NCP. Executive Order No. 12,580 provides that “[a]ll revisions to the NCP, whether in proposed or final form, shall be subject to review and approval by the Director of the Office of Management and Budget.” 52 Fed.Reg. 2923, 2924 (1987). CERCLA, though, grants the President authority to revise the NCP, and OMB is part of the Executive Office. 42 U.S.C. §§ 9605, 9615. Perhaps for this reason, “[t]he States are not challenging the authority of OMB to review the NCP.” Final Amended Joint Brief of Petitioning States at 38. Instead, the States question whether deference is appropriate. The preamble reveals that EPA considered a number of comments from OMB, as well as from other interested parties, such as the States. EPA then settled on a final rule, and it alone claimed responsibility for the contents of the NCP. 55 Fed.Reg. 8813 (1990). Our review is based on EPA’s justification for changes in the NCP, and its response to comments from a number of parties. We are not reviewing, or deferring to, any justification offered by OMB. D. Has EPA improperly interpreted the CERCLA requirement of five-year review of certain remedial actions? The States next challenge EPA’s interpretation of the CERCLA requirement of five-year review of certain remedial actions. This claim must also be rejected. CERCLA provides for a five-year review of Superfund sites as follows: If the President selects a remedial action that results in any hazardous substances ... remaining at the site, the President shall review such remedial action no less often than each 5 years after the initiation of such remedial action to assure that human health and the environment are being protected by the remedial action being implemented. In addition, if upon such review ... action is appropriate at such site ... the President shall take or require such action. 42 U.S.C. § 9621(c). EPA, exercising power delegated from the President, is also required to supply Congress with a list of sites subject to review, the results of reviews, and any actions taken in light of the reviews. Id. EPA interprets this provision to require review only when remedial action “results in hazardous substances ... remaining at the site above levels that allow for unlimited use and unrestricted exposure.” 40 C.F.R. § 300.430(f)(4)(ii). A site is not designated for review when the initial remedial action renders the site safe, under the standards prevailing at the time of the determination, for all purposes and for an unlimited period of exposure through drinking water, air, or any other “exposure pathway.” The States attack this standard on two grounds. First, the States argue that EPA’s approach violates clear statutory language requiring a review when “any hazardous substances” remain at the site. 42 U.S.C. § 9621(c). The Agency responds that the regulation merely imposes a de minimis gloss on the CERCLA requirement in order to avoid an absurd result. EPA maintains that under the approach that the States suggest, the Agency would be required to conduct a review of every site, every five years, in perpetuity, because it is virtually impossible to prove that not a single molecule of hazardous material remains at a site. See supra p. 1530. The States do not dispute that their suggested approach would require review at all sites every five years and impose a mammoth monitoring burden on EPA. Rather, the States argue that-a de minimis exception is impermissible in this case under Public Citizen v. Young, 831 F.2d 1108 (D.C.Cir.1987), cert. denied, 485 U.S. 1006, 108 S.Ct. 1470, 99 L.Ed.2d 699 (1988). In Public Citizen, this court refused to allow a de minimis exception to the “Delaney Clause” in the Pure Food and Drug Act, which provided that a color additive will be deemed unsafe if appropriate tests reveal that it “induce[s] cancer in man or animal.” Public Citizen, 831 F.2d at 1112. The States seize in particular on the Public Citizen court’s admonition that the de minimis doctrine cannot “thwart a statutory command; it must be interpreted with a view to ‘implementing the legislative design.’” Id. at 1113 (quoting Alabama Power Co. v. Costle, 636 F.2d 323, 360-61 (D.C.Cir.1979)). The “legislative design” is not being flouted by EPA’s reading of the five-year review provision because the statutory command is not so clear as to rule out EPA’s application of a de minimis exception. The Public Citizen court relied heavily on the “almost inescapable” terms of the Delaney Clause and the substantial legislative history supporting an absolutist application of the language. See Public Citizen, 831 F.2d at 1112-17. The terms at issue here are not so rigid: the phrase “any hazardous substances” could easily mean “even one hazardous substance” as opposed to “any amount of any hazardous substance.” In addition, the legislative history provides no convincing support for the States’ position. The States point to the comment of a single Senator to bolster their position: The periodic review provision is intended to assure that Superfund cleanups keep pace with developing technologies and that remedial actions are upgraded to take advantage of such developing technologies. The ultimate goal of the Superfund program must be to implement permanent solutions at all national priorities list sites. One way to accomplish this goal is to require periodic review and to assure that sites are not removed from the ambit of the program until such solutions have been implemented. 132 Cong.Rec. 28,426 (1986) (statement of Sen. Mitchell). EPA’s interpretation is completely consistent with Senator Mitchell’s comments, which do not in any way suggest that a permanent solution has not been implemented within the meaning of the statute once a site is rendered safe for all purposes and for an unlimited period of exposure. Thus, EPA’s implementation of five-year review represents a permissible construction of the statute. Even assuming arguendo that the States’ reading of the statute were indeed the “literal” one, a de minimis exception might nonetheless be appropriate. The Public Citizen court noted that the literal meaning of a statute need not be followed where the precise terms lead to absurd or futile results, or where failure to allow a de minimis exception is contrary to the primary legislative goal. The States’ version of the statute would require that every CERCLA site be subject to five-year review because, as discussed supra p. 1530, EPA cannot detect whether “true” zero has been attained with respect to a particular hazardous substance. Section 9621(c) certainly does not appear to have been drafted to require perpetual five-year review at every Superfund site. EPA’s interpretation, which requires review only where a hazardous substance is present in an amount appreciable enough to present some possibility of harm, squares with the health-protective purpose of the statute. To go beyond that is to adjudge Congress incompetent to fashion a rational legislative design. The States also argue that under EPA’s approach, any five-year reviews that are conducted — at those sites where the initial cleanup action does not allow unlimited use and unrestricted exposure — will be rendered meaningless because EPA has stated that “the five-year review is not intended as an opportunity to consider an alternative to a protective remedy that was initially selected.” 55 Fed.Reg. 8730-31 (1990). The States argue that because all remedies must be “protective” as of implementation, the review will never provide an opportunity for new remedial action. EPA responds convincingly that new action will occur when the review reveals that the remedy is no longer protective — for example, where a remedial technology has failed, or where a newly promulgated standard indicates that the old standard is no longer protective. Thus, EPA’s construction does not render the five-year review provisions a nullity. The more substantial argument is that the Agency will not bring new toxicological information or new technologies to bear at those sites that initially fell within the Agency’s de minimis exception and are therefore not subject to five-year review. The States are correct that five-year review will not occur at sites deemed safe under the standards prevailing at the time of the determination, and that the latest information therefore will not automatically be brought to bear at these sites through the five-year review mechanism. However, this fact does not demonstrate that the Agency’s regulation is an impermissible interpretation of the statute. As long as the de minimis exception is permissible under the statute, as we hold that it is, the fact that new technologies and information will not be applied through the five-year review mechanism does not render EPA’s construction of the statute impermissible. We also hasten to note that a location initially deemed safe for all purposes and for an unlimited period of exposure would never be listed as a Superfund site in the first instance. Moreover, to say that new information will not be applied to a site via the five-year review mechanism is not to say that the new information will not be applied at all. If a site deemed safe for any use and any amount of exposure is later understood to be unsafe under new standards developed in light of new toxicological information, the site could again be eligible for Superfund treatment. Although five-year review of such sites might lead to greater protection of public health (at greater cost), we cannot say that omitting these sites from five-year review is an impermissible construction of the statute. IV The States make three additional challenges to the NCP remedy selection and cleanup provisions, none of which are ripe for judicial review. The ripeness doctrine requires us to “evaluate both the fitness of the issues for judicial decision and the hardship to the parties of withholding court consideration.” Abbott Laboratories v. Gardner, 387 U.S. 136, 149, 87 S.Ct. 1507, 1515, 18 L.Ed.2d 681 (1967). Both prongs of this test dictate delaying review of the States’ remaining claims. The claims are unfit for resolution because “judicial appraisal ... is likely to stand on a much surer footing in the context of a specific application of th[ese] regulation^] than could be the ease in the framework of the generalized challenge made here.” Toilet Goods Ass’n v. Gardner, 387 U.S. 158, 164, 87 S.Ct. 1520, 1524, 18 L.Ed.2d 697 (1967). “Where we believed the agency’s practical application of a statement would be important, we have found the issue not ripe.” Public Citizen v. Nuclear Regulatory Commission, 940 F.2d 679, 683 (D.C.Cir.1991). As to the second prong of the ripeness analysis, the States will not be prejudiced or suffer any other significant hardship by our decision to defer resolution of these issues until they are raised in the context of a site-specific challenge. See 42 U.S.C. §§ 9604(c)(3), 9621(f)(2), 9622(d), 9659, 9613(h)(4). We discuss each of the claims in turn. A. Does NCP remedy selection guidance concerning the use of engineering and institutional controls violate CERCLA’s remedy selection requirements? The States first argue that one of EPA’s “program expectations” violates CERCLA by authorizing the use of institutional controls (such as fences and deed restrictions) as a sole remedy at Superfund sites. The NCP provision regarding selection of an appropriate remedy provides in part as follows: (iii) Expectations. EPA generally shall consider the following expectations in developing appropriate remedial alternatives: (D) EPA expects to use institutional controls such as water use and deed restrictions to supplement engineering controls as appropriate for short- and long-term management to prevent or limit exposure to hazardous substances, pollutants, or contaminants.... The use of institutional controls shall not substitute for active response measures (e.g., treatment and/or containment of source material, restoration of ground waters to their beneficial uses) as the sole remedy unless such active measures are determined not to be practicable, based on the balancing of trade-offs among alternatives that is conducted during the selection of remedy. 40 C.F.R. § 300.430(a)(l)(iii). The States interpret this language to authorize EPA to choose, based on cost considerations, institutional controls as the sole remedy for cleaning up hazardous waste sites. As a result, they believe that this provision may allow EPA to use cost considerations to select a cleanup remedy that may not comply with the minimum human health and environmental protectiveness requirements of CERCLA, see 42 U.S.C. § 9621(b)(1), (d)(1), and to select a remedy in which there is no treatment or removal of contaminants. However, EPA explained in the Federal Register that the program expectations are not intended to displace the use of the nine criteria identified in 40 C.F.R. § 300.-430(e)(9)(iii): EPA has placed the expectations in the rule to inform the public of the types of remedies that EPA has achieved, and anticipates achieving, for certain types of sites. These expectations are not, however, binding requirements. Rather, the expectations are intended to share collected experience to guide those developing cleanup options.... However, the fact that a proposed remedy may be consistent with the expectations does not constitute sufficient grounds for the selection of that remedial alternative. All remedy selection decisions must be based on an analysis using the nine criteria. 55 Fed.Reg. 8702 (1990) (emphasis added); see also 40 C.F.R. § 300.430(f)(l)(i)(A) (“Overall protection of human health and the environment and compliance with ARARs ... are threshold requirements that each alternative must meet in order to be eligible for selection.”) Thus, any remedy relying on institutional controls must meet the threshold requirement of protectiveness. As the foregoing discussion amply demonstrates, this issue is unfit for judicial decision at this time because the States’ argument is premised on a hypothetical application of a nonbinding statement in the NCP. The States acknowledge that institutional controls can be utilized as a sole remedy where other remedies are not practicable, and they must concede that EPA might never implement institutional controls as a sole remedy in a manner that the States (or another party with standing) find objectionable. Furthermore, any appeal that is brought would necessarily have to be decided on the basis of the precise circumstances of the cleanup at issue and the alternative remedies available and practicable in that context. Thus, the issue is better resolved in the context of a specific application of the nonbinding statement. B. Do the NCP provisions concerning ground water restoration strategies and approaches improperly exempt certain contaminated groundwater resources ? The States next argue that the NCP provisions for dealing with contaminated ground water are inconsistent with the CERCLA mandate for protection of human health and the environment and for compliance with ARARs. See 42 U.S.C. § 9621(b)(1), (d)(1), (d)(2)(A). In the preamble to 40 C.F.R. § 300.430, EPA sets out the following program expectations: EPA expects to return usable ground waters to their beneficial uses wherever practicable, within a timeframe that is reasonable given the particular circumstances of the site. When restoration of ground water to beneficial uses is not practicable, EPA expects to prevent further migration of the plume, prevent exposure to the contaminated ground water, and evaluate further risk reduction. 55 Fed.Reg. 8846 (1990). The NCP also provides that the documentation of a remedy selection must “[¡Indicate, as appropriate, the remediation goals ... that the remedy is expected to achieve. Performance shall be measured at appropriate locations in the ground water, surface water, soils, air, and other affected environmental media.” 40 C.F.R. § 300.430(f)(5)(iii)(A). The States challenge the NCP approach to ground water contamination on four grounds. First, the States assert that EPA’s expectation of selecting “a timeframe that is reasonable given the particular circumstances of the site,” 55 Fed.Reg. 8846, permits significant delay in implementing remedies and thereby permits EPA to avoid making improvements in the environment and the level of protectiveness. The States claim that the NCP should require rapid implementation of remedies whenever possible. EPA points in response to language describing its general ground water policy and explaining that the Agency’s preference is for rapid restoration, when practicable, of Class I ground waters and contaminated ground waters that are currently, or likely in the near-term to be, the source of a drinking water supply. The most appropriate timeframe must, however, be determined through an analysis of alternatives.... More rapid restoration of ground water is favored in situations where a future demand for drinking water from ground water is likely and other potential sources are not sufficient. Rapid restoration may also be appropriate where the institutional controls to prevent the utilization of contaminated ground water for drinking water purposes are not clearly effective or reliable. 55 Fed.Reg. 8732 (1990). Thus, in a situation where health could be jeopardized, EPA intends to rapidly restore the water; in other situations, the timeframe may be longer. Second, the States argue that the NCP improperly permits a remedy to incorporate a point of compliance that is an unlimited distance away from the source of ground water contamination. The States point to the following language in the preamble to 40 C.F.R. § 300.430(f)(5)(iii)(A): EPA believes that remediation levels should generally be attained throughout the contaminated plume, or at and beyond the edge of the waste management area, when the waste is left in place. However, EPA acknowledges that an alternative point of compliance may also be protective of public health and the environment under site-specific circumstances. 65 Fed.Reg. 8753. The States emphasize the flexible nature of the preamble language. EPA notes in reply that the preamble expresses a clear preference for remediation throughout the plume and states that alternatives must in any ease be protective of public health and the environment. Third, the States argue that the EPA ground water policy permits EPA to ignore compliance with ARARs. The States assert that EPA achieves this result with respect to Class I and II ground water by establishing an exclusive federal ARAR. The States point to the following statement of EPA’s general ground water policy: For Class I and II ground waters, preliminary remediation goals are generally set at maximum containment levels, and non-zero MCLGs where relevant and appropriate, promulgated under the Safe Drinking Water Act or more stringent state standards .... 55 Fed.Reg. 8732. EPA responds that the NCP clearly requires compliance with all ARARs as a threshold requirement, and that the general statement on ground water policy does not affect the NCP requirement. As for Class III ground water, the States argue that EPA has determined improperly that Safe Drinking Water Act (“SDWA”) standards are not ARARs. The States note the following language: For Class III ground water (i.e., ground water that is unsuitable for human consumption — due to high salinity or widespread contamination that is not related to a specific contamination source — and that does not have the potential to affect drinkable or environmentally significant ground water), drinking water standards are not ARAR and will not be used to determine preliminary remediation goals. 55 Fed.Reg. 8732. EPA responds that standards from other statutes such as the SDWA only apply where “legally applicable.” 42 U.S.C. § 9621(d)(2)(A). Thus, EPA argues, it has properly concluded that where the ground water does not come within the scope of the SDWA, the Agency is not obligated to apply those standards. EPA acknowledges that it must apply the standards in any case if it determines that they are otherwise “relevant and appropriate under the circumstances” of the specific site in question. Id. The NCP sets out the procedure for making the “relevant and appropriate” determination. See 40 C.F.R. § 300.400(g)(2). Fourth, and finally, the States assert that a variety of additional preamble statements, regarding general ground water policy and specific NCP regulations, permit remedies that are inconsistent with the CERCLA mandate for remedies that protect human health and the environment and are permanent to the maximum extent practicable. See 42 U.S.C. § 9621(b)(1), (d)(1). EPA again responds that the nine criteria set out in 40 C.F.R. § 300.430(e)(9)(iii) — the first of which is protection of human health and the environment, see id. § 300.430(e)(9)(iii)(A), and the third of which is permanence, see id. § 300.430(e)(9)(iii)(C) — must always be used in selecting a remedy. EPA points out that the nine criteria will be balanced on a site-specific basis, but that overall protection of the environment is a threshold requirement that each alternative must meet in order to be considered. 40 C.F.R. § 300.-430(f)(l)(i)(A). The States must make site-specific challenges to press each of its four ground water contamination claims — that the NCP permits remedy implementation timeframes that are unreasonably long, that the NCP permits remedies to incorporate unreasonably remote points of compliance, that the NCP permits EPA to ignore compliance with ARARs, and that the NCP permits remedies that are inconsistent with the CERCLA mandates of protection of human health and the environment and permanence. EPA argues with regard to each claim that the States have simply misapprehended the import of the various statements that form the basis of their arguments. Because the claims are premised on hypothetical applications of nonbinding statements in the NCP, we conclude that they should be addressed in site-specific challenges in which the reviewing court can consider “the agency’s practical application” of its statements. See Public Citizen, 940 F.2d at 683. C. Does the NCP improperly fail to apply Federal Water Quality Criteria (“FWQC”) as ARARs? The States’ final set of unripe claims involves EPA’s decision to use MCLs and non-zero MCLGs in place of the federal water quality criteria established under the Clean Water Act (“CWA”). CERCLA requires that remedial actions attain these federal water quality criteria (“FWQC”) wherever “relevant and appropriate under the circumstances of the release or threatened release.” 42 U.S.C. § 9621(d)(2)(A). Like MCLGs, FWQC do not have any independent regulatory impact. See supra pp. 1529-30. Rather,