Edison Electric Institute v. United States Environmental Protection Agency

Opinion PER CURIAM. PER CURIAM: This consolidated appeal presents numerous petitions for review of EPA’s Toxicity Characteristic (“TC”) rule, promulgated pursuant to the Resource Conservation and Recovery Act of 1976. The TC seeks to identify waste which, if mismanaged, may release hazardous amounts of toxic materials into the environment. In order to accomplish this goal, the TC includes a toxicity test designed to simulate the actual leaching of wastes that might occur in a municipal solid waste landfill. The petitioners challenge several aspects of this test. We hold that EPA has failed to explain adequately the application of the TC toxicity test to certain mineral processing and electric utility wastes, and therefore remand to EPA for further proceedings consistent with this opinion. The petitions for review are denied in all other respects.' Glossary of Acronyms AMC American Mining Congress API American Paper Institute DAF Dilution and Attenuation Factor EEI Edison Electric Institute EP Extraction Procedure EPA Environmental Protection Agency HSWA Hazardous and Solid Waste Amendments of 1984 ISRI Institute of Scrap Recycling Industries MSW Municipal Solid Waste NFPA National Forest Products Association NIPDWS National Interim Primary Drinking Water Standards RCRA Resource Conservation and Recovery Act of 1976 TC Toxicity Characteristic TCLP Toxicity Characteristic Leaching Procedure tpy tons of waste per year UST Underground Storage Tank I. BacKGRound A. The Resource Conservation and Recovery Act of 1976 and the 1980 Toxicity Characteristic Congress enacted the Resource Conservation and Recovery Act of 1976 (“RCRA”), 42 U.S.C. §§ 6921-6982, to establish a comprehensive program to regulate the handling of solid wastes. Environmental Defense Fund v. EPA, 852 F.2d 1309, 1310 (D.C.Cir.1988) (“EDF I”). Subtitle C of RCRA, 42 U.S.C. §§ 6921-6939b, establishes “a ‘cradle-to-grave’ regulatory structure overseeing the safe treatment, storage and disposal of hazardous waste.” United Technologies Corp. v. EPA, 821 F.2d 714, 716 (D.C.Cir.1987). Solid wastes that are not identified as hazardous are subject to the less stringent requirements of RCRA Subtitle D, 42 U.S.C. §§ 6941-6949a. EDF I, 852 F.2d at 1310. *441 RCRA itself does not include a list of hazardous wastes nor a specific method for determining whether a waste is hazardous. Instead, the statute defines “hazardous waste” generally as a solid waste, or combination of solid wastes, which because of its quantity, concentration, or physical, chemical, or infectious characteristics may— (A) cause, or significantly contribute to an increase in mortality or an increase in serious irreversible, or incapacitating reversible, illness; or (B) pose a substantial present or potential hazard to human health or the environment when improperly treated, stored, transported, or disposed of, or otherwise managed. 42 U.S.C. § 6903(5). Having set out this general definition, Congress delegated to the Environmental Protection Agency (“EPA” or “the Agency”) the duty to “promulgate regulations identifying the characteristics of hazardous waste, and listing particular hazardous wastes ... which shall be subject to the provisions of [Subtitle C].” 42 U.S.C. § 6921(b)(1). Thus, Congress directed the Agency to identify hazardous wastes in two ways: (1) identify certain characteristics which would render a solid waste hazardous, and (2) list specific solid wastes that are, so to speak, per se hazardous. This appeal concerns only the former category of solid wastes, those deemed hazardous by virtue of possessing certain general characteristics. More specifically, the petitioners present a multifarious challenge to EPA’s final rule revising the Toxicity Characteristic (“TC”) — one of the four characteristics (the other three are ignitability, corrosivity, and reactivity) set out in EPA regulations for the purpose of identifying hazardous solid wastes. See 40 C.F.R. § 261.24; 55 Fed.Reg. 11,798 (1990). The TC seeks to “identify waste which, if improperly disposed of, may release toxic materials in sufficient amounts to pose a substantial hazard to human health or the environment.” 43 Fed.Reg. 58,952 (1978). In 1980, EPA established a “protocol” for determining the TC of solid wastes, which it dubbed the “Extraction Procedure” (“EP”). See 45 Fed.Reg. 33,110-12 (1980). The EP toxicity test is based on a particular mismanagement scenario — “co-disposal of toxic wastes in an actively decomposing municipal landfill which overlies a groundwater aquifer,” id at 33,110 — and is intended to simulate the actual leaching of wastes that might occur in a municipal solid waste (“MSW”) landfill. The test requires a waste generator to mix a representative sample of its waste with an acidic leaching medium for 24 hours, and then to test the resulting liquid waste to see if it contains unsafe levels of any of 14 toxic contaminants identified in the National Interim Primary Drinking Water Standards (“NIPDWS”) promulgated pursuant to 42 U.S.C. § 300g-l. Id. In order to duplicate the attenuation in concentration expected to occur between the point of leachate generation and the point of human or environmental exposure, the EP applies a dilution and attenuation factor (“DAF”) of 100 to the concentration of toxic contaminants observed in the test extract. 45 Fed.Reg. 33,111. Thus, a waste would be considered hazardous, and subject to RCRA Subtitle C regulation, if the results of the EP toxicity test revealed the presence of any listed contaminant at a level of at least 100 times the applicable NIPDWS. B. The Revised Toxicity Characteristic The initial challenge to the 1980 TC and other RCRA regulations was filed shortly after their promulgation, but this Court suspended briefing in light of ongoing settlement negotiations. Shell Oil v. EPA, No. 80-1532 (D.C.Cir. May 21, 1982). After negotiations failed to produce a settlement, the challenge to the TC rule was reactivated, id. (D.C.Cir. June 22, 1988), but subsequently severed from the other issues in the case and held in abeyance pending completion of a rulemaking commenced in 1986 to revise the TC. Id. (D.C.Cir. March 7, 1989); id. (D.C.Cir. May 17, 1989). The challenges to the 1980 TC rule were eventually consolidated in this challenge. Edison Electric Institute v. EPA, No. 89-1320 (D.C.Cir. August 8, 1990). *442 The 1986 rulemaking to revise the TC was necessitated by two intervening pieces of legislation. First, in 1980, Congress passed the Bevill Amendment, codified at 42 U.S.C. § 6921(b)(3)(A), as part of the Solid Waste Disposal Act Amendments of 1980, Pub.L. No. 96-482, 94 Stat. 2334. The Bevill Amendment exempted from Subtitle C regulation certain waste produced by fossil fuel combustion and mineral processing and directed EPA to study the environmental effects of such wastes and to determine whether special regulations were necessary to govern their disposal. See Solite Corp. v. EPA, 962 F.2d 473, 478 (D.C.Cir.1991). In litigation to compel EPA to meet the statutory deadlines for implementation, this Court found that Congress intended to exempt “only those wastes from processing ores or minerals that [are] ‘high volume, low hazard’ wastes.” Environmental Defense Fund v. EPA, 852 F.2d 1316, 1328-29 (D.C.Cir.1988) (“EDF II"), cert. denied, 489 U.S. 1011, 109 S.Ct. 1120, 103 L.Ed.2d 183 (1989). In a subsequent rulemaking, EPA decided to exempt only those high volume, low hazard wastes that fell within the top 6% of the largest individual waste streams managed by Subtitle C facilities. Solite, 952 F.2d at 488. The second piece of intervening legislation was the Hazardous and Solid Waste Amendments of 1984 (“HSWA”), Pub.L. No. 98-616, 98 Stat. 3221. In one provision of HSWA, Congress expressed concern with the TC and the EP toxicity test and directed the Agency to reevaluate: Not later than twenty-eight months after November 8, 1984, the Administrator shall examine the deficiencies of the extraction procedure toxicity characteristic as a predictor of the leaching potential of wastes and make changes in the extraction procedure toxicity characteristic, including changes in the leaching media, as are necessary to insure that it accurately predicts the leaching potential of wastes which pose a threat to human health and the environment when mismanaged. 42 U.S.C. § 6921(g). The legislative history indicates that Congress believed that EPA’s test was deficient because it was underinclusive in identifying hazardous wastes. H.R.Conf.Rep. No. 1133, 98th Cong., 2d Sess. 105 (1984) (“the Administrator is directed to change the [EP] to predict the leaching potential of wastes upon exposure to more aggressive leaching media than the media presently used”) (emphasis added), reprinted in 1984 U.S.C.C.A.N. 5576, 5649, 5676. S.Rep. No. 284, 98th Cong., 1st Sess. 35 (1983) (same). In response to HSWA, EPA revised its regulations and adopted a new testing procedure in place of the EP, known as the Toxicity Characteristic Leaching Procedure (“TCLP”). See 40 C.F.R. pt. 261, app. II. After considering several alternatives, the Agency left the EP’s mismanagement scenario essentially unchanged for purposes of the TCLP. 55 Fed.Reg. 11,806 (1990). However, the TCLP differs from the EP in at least two significant respects. First, the TCLP requires the use of one of two leaching fluids depending upon the alkalinity of the waste being tested. 51 Fed.Reg. 21,655-56 (1986). Second, the TCLP eliminates a prior exception to the general rule that large solid pieces of waste must be reduced to particles prior to testing. 55 Fed.Reg. 11,823. In addition to developing the TCLP, EPA proposed to expand the list of constituents regulated under the TC from 14 to 52, and to adopt compound-specific dilution and attenuation factors for the 38 newly regulated compounds and for six of those previously covered. The levels for the remaining eight substances (all of the regulated metals) continued to be based on a “generic” DAF of 100. However, because of technical problems in calculating some of the dilution factors, the number of new constituents covered by the 1990 TC was eventually reduced from 38 to 25, 55 Fed.Reg. 11,798, and a single DAF of 100 was used for all regulated constituents. Id. at 11,825-26. The petitions present four separate challenges by industry and environmental groups to the TC rule. Representatives of the mining and electric utility industries maintain that the application of the TCLP generic mismanagement scenario to certain mineral processing wastes and electric utility wastes is unreasonable and unsupported by the record. Representatives of the pulp and paper *443 industry claim that the TC’s regulatory level for chloroform fails properly to consider biodegradation. Representatives of the scrap recycling industry complain that EPA failed to provide adequate notice and opportunity to comment on the application of the TCLP to metals and other inorganic constituents. Finally, an environmental group and representatives of the waste treatment industry assert that EPA violated the RCRA by temporarily deferring application of the TC rule to petroleum-contaminated soils resulting from clean-ups of leaking underground storage tanks. II. APPLICATION OF THE TCLP MISMANAGEMENT SCENARIO to Mineral Prooessing AND ELECTRIC UTILITY WASTES The American Mining Congress (“AMC”) and the Edison Electric Institute (“EEI”) challenge EPA’s application of the TCLP generic mismanagement scenario to mineral processing wastes and manufactured gas plant wastes. The parties agree that EPA treats these wastes similarly, and they are hereinafter referred to collectively as “mineral wastes.” They do not fall under the Bevill Amendment’s exemption from Subtitle C regulation, as implemented in EPA’s regulations, because individual waste generators do not produce them in sufficient quantities to qualify for the high volume, low hazard exception. See Solite, 952 F.2d at 488. (EEI also challenges the mismanagement scenario as applied to coal combustion wastes. However, as EEI concedes, EPA is currently considering whether to place these wastes under Subtitle C regulation, and they are temporarily exempt from Subtitle C regulation under 42 U.S.C. § 6921(b)(3)(A)(i). Thus, EEI has yet to suffer an injury and lacks standing to challenge the application of the TC to coal combustion wastes. See Lujan v. Defenders of Wildlife, — U.S. -, -, 112 S.Ct. 2130, 2136, 119 L.Ed.2d 351 (1992).) AMC and EEI claim that EPA’s approach violates the congressional directive to insure that the TCLP “accurately predicts the leaching potential of wastes ... when mismanaged,” 42 U.S.C. § 6921(g), because the generic scenario is based on factual assumptions that do not apply to mineral wastes. The petitioners also argue that EPA has failed to articulate a rational connection between the TCLP generic mismanagement scenario and the mineral wastes in question. We agree with EPA that the adoption of a generic mismanagement scenario, and its application to mineral wastes, does not violate the RCRA command that the TCLP accurately predict the leaching potential of mismanaged wastes. Nothing in the statute requires EPA to tailor the TCLP to the conditions to which mineral wastes are typically exposed. However, the TC rule must at least bear some rational relationship to mineral wastes in order for the Agency to justify its application to those wastes. Motor Vehicle Mfrs. Ass’n v. State Farm, 463 U.S. 29, 43, 103 S.Ct. 2856, 2866, 77 L.Ed.2d 443 (1983). We hold that EPA has failed to demonstrate any such relationship on the record, and we therefore remand to the Agency for further proceedings consistent with this opinion. A. Statutory Requirement of “Accuracy” AMC and EEI claim that the application of the TCLP to mineral wastes violates the statutory mandate for a more accurate TCLP, see 42 U.S.C. § 6921(g), because the generic mismanagement scenario is based on factual assumptions that do not apply to mineral wastes. They maintain that it is extremely unlikely that mineral wastes will ever be disposed of in municipal solid waste (“MSW”) landfills in light of the extremely large volumes of waste that are generated and the relatively small capacity of MSW landfills. In support of this claim, the petitioners point to EPA findings that two-thirds of all MSW landfills handle less than 8,000 tons of waste per year (“tpy”), U.S. Environmental Protection Agency, Report to Congress: Solid Waste Disposal in the United States, vol. II, at 4-10 (October 1988), and that at least four types of non-Bevill solid mineral wastes are generated in average volumes greater than 8,000 tpy per facility. See 56 Fed.Reg. 55,160, 56,184-85 (1991). AMC and EEI cite several additional factors that they claim render disposal of mineral wastes in an MSW landfill extremely un *444 likely. They point to regulatory barriers against such disposal, and also assert that the wastes in question “exhibit physical and/or chemical properties that make management in an MSW landfill impractical.” In addition, the petitioners note that some of the wastes in question contain significant reserves of valuable metals that would be lost if they were disposed of in an MSW landfill rather than stockpiled for future re-mining. Finally, AMC and EEI maintain that the wastes are typically generated at facilities with their own on-site landfill capacity, far from any municipal site. The petitioners argue in the alternative that even if the mineral wastes were disposed of in an MSW landfill, the large volume of waste would comprise most of the content of the landfill, and thus the chances that the waste would be subject to the acidic leaching processes prevalent in normal MSW landfills would be greatly reduced. Finally, AMC and EEI assert that the inapplicability of the TCLP generic mismanagement scenario to the mining context renders three specific TCLP elements — the aggressiveness of the leaching medium, the particle size reduction requirement, and the assumed DAF of 100 — far too severe in predicting the hazardousness of these types of wastés. As to the first element, AMC and EEI contend that mineral wastes are not likely to be exposed to the severe acidic conditions simulated by the TCLP acetic acid leaching medium. They point to EPA’s acknowledgment that mineral processing wastes are unlikely to be managed in environments that contain or are capable of generating organic acids, such as the acetic acid formed by decaying garbage; mineral processing wastes, with very few exception, do not contain appreciable quantities of organic matter. 54 Fed.Reg. 36,601 (1989). Thus, argue the petitioners, the TCLP leaching medium overstates the leaching of mineral wastes that can be ejected in the real world. The petitioners also claim that the TCLP particle size reduction requirement, like the leaching medium, mimics conditions in MSW landfills that are not appropriate for mineral wastes. The particle size reduction requirement, which calls for the grinding of wastes prior to extraction, is meant to simulate three forces in MSW landfills: (1) the weight of trucks and other heavy machinery that might compress the waste; (2) freeze-thaw cycles that might cause contraction and expansion of the waste; and (3) wet-dry cycles that might also cause contraction and expansion. See 55 Fed.Reg. 11,826 (1990). The petitioners contend that EPA has failed to demonstrate that mineral wastes are exposed to these forces. AMC and EEI claim that mineral wastes are not subject to as much compression as wastes disposed of in MSW landfills because they do not require the same daily application of cover material and maintenance involving heavy machinery. Furthermore, they argue that mineral wastes are generally not exposed to regular wet-dry cycles because most are processed in the arid west. Finally, with respect to EPA’s dilution and attenuation assumptions, the petitioners argue that the DAF employed in the TCLP also overestimates the hazardousness of mineral wastes. The hypothetical TCLP DAF of 100 is based on the assumption that a drinking water well is located 500 feet down gradient from the MSW landfill site. 45 Fed.Reg. 33,111 (1980). AMC and EEI argue that this assumption is unrealistic with regard to mineral wastes because most mineral processing facilities are located in remote areas of the west, often miles from the nearest drinking well. The petitioners also claim that the unsaturated soil layer in the arid west is generally very thick, which would result in considerably greater attenuation than at most MSW landfills during migration from the landfill to the aquifer. EPA maintains in response that it reasonably established a single toxicity test for making a threshold determination of hazard-ousness based on reasonable assumptions about how industrial wastes might plausibly be mismanaged in the absence of RCRA Subtitle C controls. The Agency argues that the statutory command to improve the accuracy of the TCLP, see 42 U.S.C. § 6921(g), does not require EPA to tailor carefully the *445 test to the conditions to which mineral wastes, or any other particular type of industrial wastes, are typically exposed. The Agency also maintains that the particular generic scenario on which it settled is permissible under RCRA. We agree with both contentions. Congress gave EPA the authority to “promulgate regulations identifying the characteristics of hazardous waste,” 42 U.S.C. § 6921(b)(1); see also id. § 6921(a), and defined hazardous waste as those solid wastes which may “pose a substantial present or potential hazard to human health or the environment when improperly treated, stored, transported, or disposed of, or otherwise managed.” 42 U.S.C. § 6903(5)(B). RCRA does not define improper management or specify any particular type of mismanagement scenario. Thus, because congressional intent on the precise question at issue is unclear, EPA’s adoption of the generic mismanagement scenario should be upheld if it constitutes a reasonable construction of RCRA. Chevron v. Natural Resources Defense Council, 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). EPA decided in developing the TC to make assumptions about plausible worst-case mismanagement practices in order to determine whether a waste would pose a threat to human health and the environment if mismanaged. 45 Fed.Reg. 33,110. The Agency settled on a particular, hypothetical, worst-case mismanagement scenario: co-disposal of toxic wastes in an actively decomposing MSW landfill which overlies a groundwater aquifer. Id. EPA adequately justified its approach both in the initial rulemaking promulgating the TC and the EP, and in the later proceedings that produced the revised TC and the TCLP. EPA settled on a single mismanagement scenario and rejected a “management-based” approach to identifying hazardous wastes, which would require a separate toxicity test for each category of waste that is typically managed in a particular way. In keeping with the RCRA directive to “promulgate regulations identifying the characteristics of hazardous waste,” 42 U.S.C. § 6921(b)(1) (emphasis added), EPA decided that “the most effective and appropriate approach [to implementing RCRA] is to ... identify! ] properties of wastes that would pose a threat to human health and the environment if improperly managed.” 55 Fed.Reg. 11,807 (1990) (emphasis added). The Agency concluded that a management-based approach would raise complex enforcement problems because of the difficulty in determining beforehand how any particular solid waste will eventually be managed. Id. As we have already stated, nothing in the RCRA mandate of a more accurate TC dictates a management-based approach, and the use of a generic mismanagement scenario is a reasonable interpretation of the statutory language. EPA selected the specific MSW landfill mismanagement scenario because contamination of groundwater through the leaching of land-disposed wastes is a prevalent environmental hazard that is well-documented in EPA damage files and with which Congress was especially concerned in passing RCRA. 45 Fed.Reg. 33,110. In response to comments that industrial solid wastes are not often disposed of in MSW landfills, EPA pointed out that states impose few restrictions on the types of non-hazardous wastes accepted as MSW landfills, and that a substantial quantity of the wastes actually received at MSW landfills are industrial wastes. 55 Fed.Reg. 11,806. EPA recognized that MSW landfills generate a more aggressive leachate media than other landfills, but chose to adopt a particularly conservative scenario “in view of the statutory mandate to protect human health and the environment, the broad statutory definition of hazardous waste[,] and also because the phenomenon of long term leaching is only incompletely understood.” 45 Fed.Reg. 33,112. These choices represent a reasonable interpretation of RCRA. The HSWA requirement that EPA revisit the EP test to insure its accuracy does not dictate a different result. Congress directed EPA to examine the deficiencies of the extraction procedure toxicity characteristic as a predictor of the leaching potential of wastes and make changes in the extraction procedure toxicity characteristic, including *446 changes in the leaching media, as are necessary to insure that it accurately predicts the leaching potential of wastes which pose a threat to human health and the environment when mismanaged. 42 U.S.C. § 6921(g). EPA had already developed its first TC protocol, using the same mismanagement scenario adopted in the TCLP, at the time Congress passed HSWA. Whereas Congress chose to mention explicitly the extraction procedure, and focus even more specifically on the choice of leaching medium, the provision makes no mention of the mismanagement scenario itself. Although EPA was certainly free to consider a new mismanagement scenario or a management-based approach, nothing in the HSWA requirement of improved accuracy required EPA to take either of these directions. EPA reasonably responded to the congressional call for improved accuracy by changing the TCLP to include two leaching fluids and by eliminating an exception to the rule that large solid pieces of waste must be reduced to particles prior to testing. Thus, EPA’s decision to adopt a generic mismanagement scenario involving co-disposal in an MSW landfill represents a permissible interpretation of RCRA. AMC and EEI also claim that EPA is obligated to include a variance mechanism that would cure the alleged “overbreadth” of the TC. EPA properly responded that such a variance is tantamount to a management-based approach, which EPA reasonably rejected. See supra pp. 444-446; 55 Fed.Reg. 11,806. The Agency regulation allowing any party to petition for a regulatory amendment if they can suggest a superior toxicity test, 40 C.F.R. § 260.21(a), is sufficient to address the petitioners’ concerns. B. Reasonableness of Applying the TCLP Mismanagement Scenario to Mineral Wastes The inquiry is not completed by our conclusion that the TCLP mismanagement scenario represents a permissible construction of RCRA under Chevron. In addition, to pass muster under the APA, the TCLP must bear some rational relationship to mineral wastes in order for the Agency to justify the application of the toxicity test to those wastes. See Motor Vehicle Mfrs. Ass’n v. State Farm, 463 U.S. 29, 43, 103 S.Ct. 2856, 2866, 77 L.Ed.2d 443 (1983). We hold that EPA has failed to demonstrate any such relationship on the record, and therefore remand to the Agency for further proceedings consistent with this opinion. We have already stated in considerable detail the petitioners’ objections to the application of the TCLP mismanagement scenario to mineral wastes. See supra pp. 443-444. We agree with AMC and EEI that EPA’s justification on the record for applying the TCLP to mineral wastes consists of speculative factual assertions. EPA stated as follows in the final TC rule: Although large volume wastes from the extraction, beneficiation and processing of ores and minerals are currently exempt from subtitle C regulation and will not be affected by the TC rule, small volume mineral processing wastes which may be subject to subtitle C regulation can plausibly be disposed in municipal landfills. 55 Fed.Reg. 11,807. In support of this conclusion, EPA noted two studies demonstrating that “[although fewer types of industrial wastes are being disposed in municipal landfills now as compared to a few years ago, EPA’s information confirms the continued appropriateness of this scenario.” Id. at 11,-806. The record evidence on which EPA relies does not demonstrate that low volume mineral wastes have ever been disposed of in MSW landfills. EPA did note that a substantial quantity of industrial wastes are deposited in MSW landfills, but the Agency did not point to any evidence that any of these wastes were generated by the mineral processing or electric utility industries. EPA need not demonstrate that mineral wastes are typically or commonly deposited in MSW landfills, but the Agency must at least provide some factual support for its conclusion that such a mismanagement scenario is plausible. Even in the absence of evidence that at least some mineral wastes have actually been disposed in MSW landfills, EPA’s application of the TCLP to mineral wastes would none *447 theless pass muster if there were evidence on the record that mineral wastes were exposed to conditions similar to those simulated by the TCLP. EPA concluded that “even those [wastes] which are unlikely to be disposed of in a municipal landfill[] are likely to come into contact with some form of acidic leaching media during their management histories or could otherwise encounter environments which could cause them to leach comparable levels of toxic constituents.” 45 Fed.Reg. 33,112 (1980). Again, however, there is no evidence or explanation on the record to justify a conclusion that mineral wastes ever come into contact with any form of acidic leaching medium. Of course, we do not conclude or suggest that mineral wastes are never disposed of in MSW landfills, exposed to acidic leaching conditions, or otherwise managed in a manner that makes application of the TCLP reasonable. However, the Court “may not supply a reasoned basis for the agency’s action that the agency itself has not given.” State Farm, 463 U.S. at 43, 103 S.Ct. at 2867. We therefore remand to allow the Agency to provide a fuller and more reasoned explanation for its decision to apply the TCLP to mineral wastes. III. TREATMENT OF CHLOROFORM AND Biodegradation in the ToxiCity CHARACTERISTIC The American Paper Institute (“API”) and the National Forest Products Association (“NFPA”) contend that the regulatory level for chloroform set in the TC rule is unlawful and arbitrary and capricious. Chloroform is one of the 25 constituents considered under the TC for purposes of determining whether a solid waste should be regulated under RCRA Subtitle C. Thus, if chloroform is present in the liquid waste produced by the TCLP at a level of at least 100 times the appropriate NIPDWS, the tested waste is deemed hazardous and subject to Subtitle C regulation on account of its toxicity. See supra p. 441. The petitioners’ challenge to the regulatory level for chloroform rests on the undisputed fact that the DAF of 100, which is designed to account for the attenuation in concentration expected to occur between the point of leachate generation and the point of human or environmental exposure, see supra p. 441, does not consider the effect of biode-gradation on the subsurface migration of chloroform. API and NFPA argue that the failure to account for biodegradation is unlawful because it violates the congressional mandate that the Agency improve the accuracy of its leaching procedure. The petitioners also contend that the decision is arbitrary and capricious because it is not supported by the record and is inconsistent with various Agency policies. We disagree. A. “Accuracy” of the Toxicity Characteristic as Applied to Chloroform Taking the same tack as AMC and EEI in their challenge to the TCLP’s application to mineral wastes, API and NFPA contend that EPA’s failure to consider biodegradation violates the congressional mandate “to insure that [the TCLP] accurately predicts the leaching potential of wastes which pose a threat to human health and the environment when mismanaged.” 42 U.S.C. § 6921(g). They point to EPA’s acknowledgement that “biodegradation as a mechanism was neglected, i.e., the biodegradation coefficient was set to zero” in the TC for chloroform. EPA, Background Document for EPA’s Composite Model for Landfills § 6.2.4, at 79. API and NFPA cannot prevail under Chevron for reasons similar to those already discussed with respect to the statutory challenge of AMC and EEI, supra pp. 444-446. The statutory mandate for improved accuracy focuses explicitly on the extraction procedure and the choice of leaching medium, but makes absolutely no mention of the DAF or particular attenuation mechanisms. See 42 U.S.C. § 6921(g). The petitioners cannot plausibly maintain 'that this statutory directive evinces a clear congressional intent to consider biodegradation as an attenuating mechanism in determining the regulatory level for chloroform under the TC. Thus, because congressional intent on the precise question at issue is unclear, EPA’s approach survives Chevron scrutiny if it constitutes a reasonable construction of the statute. *448 Chevron, 467 U.S. 837, 104 S.Ct. 2778. As we have already concluded, supra pp. 444-446, EPA reasonably responded to the congressional call for improved accuracy by changing the TCLP to include two leaching fluids and eliminating an exception to the rule that large solid pieces of waste must be reduced to particles prior to testing. B. Reasonableness of Excluding Biodegra-dation in Setting the TC Regulatory Level for Chloroform The petitioners argue that EPA’s decision not to account for biodegradation of chloroform in its subsurface transport model is arbitrary and capricious because the decision is unsupported by the administrative record, inconsistent with other EPA action, and in violation of EPA policy regarding verification of model results. The petitioners also claim that EPA failed to consider regulatory alternatives before deciding not to account for biodegradation. We reject each of these claims. EPA decided not to take account of biode-gradation in its revised TC because the Agency did “not have adequate information at this time to include biodegradation in the model.” EPA, Responses to Public Comments on Subsurface Fate and Transport Model Used in the Toxicity Characteristic Rule, at 2-33 to 2-37 (“EPA Responses ”); see also 55 Fed.Reg. 11,823-24 (1990). The petitioners claim that EPA improperly failed to consider the extensive data they submitted on biodegradation rates — see Comments of API and NFPA (September 26, 1986), at VII-12 to VII-16; id. at Attachment 10 (citing articles submitted by API and NFPA); Comments of API and NFPA (June 5, 1988), at 10-14 — because similar data were not available for other organic constituents. EPA did reject the petitioners’ data in part because similar data were not available for most of the constituents regulated under the TC. See EPA Responses at 2-33 (“[o]nly a few of the chemicals affected by this rule-making were included [in the biodegradation data submitted by the petitioners], and most of the data was for aerobic conditions”). However, EPA gave another reason for deciding not to factor biodegradation into its model: Biodegradation rates in the subsurface are known to vary considerably across the United States, but there are insufficient data to quantify this variability.... At present, the Agency does not have sufficient data correlating biodegradation rates with subsurface conditions (e.g., pH, temperature) to accurately model anaerobic degradation in the subsurface. Id. API and NFPA have never contested, before the Agency or this Court, EPA’s finding that the submitted data failed to account for pH and temperature variations in different subsurface environments throughout the United States. In light of the substantial deference due to an Agency’s expert scientific judgment, see Natural Resources Defense Council v. Thomas, 824 F.2d 1211, 1216 (D.C.Cir.1987), we hold that this deficiency in the data warranted its rejection by the Agency. API and NFPA also claim that EPA failed to consider the following three alternative approaches to biodegradation: (1) adoption of a single, conservative biodegradation rate based on evidence submitted by API and NFPA; (2) use of that evidence to calculate a reasonable range of biodegradation rates for chloroform by using a “Monte-Carlo” mathematical technique; and (3) deferral of regulation of chloroform until EPA has developed the data it believed necessary to account for biodegradation. The failure to consider legitimate alternatives may render an agency decision arbitrary and capricious. See Yakima Valley Cablevision v. FCC, 794 F.2d 737, 746 (D.C.Cir.1986). However, EPA adequately considered and rejected each of the alternatives that the petitioners raised below. Thus, the Agency did not act in an arbitrary or capricious manner by failing to adopt one of the suggested alternatives. API and NFPA first suggest that EPA should have adopted a single, conservative biodegradation rate from within the range suggested by the petitioners’ studies. But as we have already discussed, EPA reasonably concluded that it could not rely upon these studies. Thus, the Agency was justified in setting the biodegradation rate to the single, *449 conservative rate of zero. See EPA, Background Document for EPA’s Composite Model for Landfills § 6.2.4, at 79 (noting that under the revised TC “the biodegradation coefficient was set to zero”). The second alternative, the use of a mathematical model known as the “Monte Carlo” technique, was not raised by the petitioners’ in any of its comments to the Agency. Thus, EPA was not obligated to address this alternative. 42 U.S.C. § 7607(d)(7)(B); Northside Sanitary Landfill v. Thomas, 849 F.2d 1516 (D.C.Cir.1988), cert. denied, 489 U.S. 1078, 109 S.Ct. 1528, 103 L.Ed.2d 833 (1989). In any case, EPA did explain that the Monte Carlo technique requires reliable nationwide data, which were lacking for biodegradation. See supra p. 448. The third and final alternative that API and NFPA suggested was for EPA to defer setting a regulatory level for chloroform. EPA chose to defer regulation of certain constituents based on reliable studies suggesting that the constituents had an unusual sensitivity to hydrolysis, one of the processes that affects subsurface attenuation. 55 Fed.Reg. 11,822-23 (1990). However, EPA found that there were insufficient reliable data showing an unusual sensitivity of chloroform, or any other constituent, to biodegradation rates in an anaerobic environment. See EPA Response at 2-33. Thus, EPA reasonably decided that deferral was not warranted based on the information submitted by the petitioners. The petitioners also maintain that EPA acted in an inconsistent manner by deferring regulation of some constituents, but declining to defer regulation of chloroform. But the distinction in the record discussed immediately above explains why EPA decided to treat differently chloroform and other constituents. The Agency had data demonstrating that certain constituents were highly sensitive to a particular model parameter or assumption, but it did not have sufficient data to remedy the inaccuracy resulting from that sensitivity. In other words, EPA had data indicating that there was a special problem and that more data were necessary to address it. In contrast, EPA had no reliable data indicating that the regulatory level for chloroform would be greatly affected by the development of a biodegradation parameter. Thus, EPA reasonably decided that there was no need to defer in anticipation of additional data regarding the biodegradation of chloroform. Finally, the petitioners contend that EPA’s failure to use the biodegradation data they submitted violated the Agency’s own policy that “model results must be compared to available measurements, and any significant discrepancies should be discussed.” 51 Fed.Reg. 34,048 (1986). But EPA’s decision to reject the petitioners’ biodegradation studies is not inconsistent with the announced policy of comparing the groundwater transport model to available empirical measurements. As we have noted, supra p. 448, EPA did consider the petitioners’ data but found it lacking on a number of levels — most prominently, its failure to account for variations in subsurface pH and temperature. Having reasonably rejected these studies, the agency was not obligated under its policy to compare the model to the results reported in the studies. Thus, based on the record evidence available at the time of the decision, EPA did not act in an arbitrary or capricious manner by deciding not to account for biodegradation as an attenuating mechanism in establishing the regulatory level for chloroform. The petitioners may seek to have EPA revise the rule if and when appropriate new data are available. 5 U.S.C. § 553(e); 42 U.S.C. § 6974. IV. APPLICATION OF THE TC TO METALLIC SUBSTANCES The petitioners contend that EPA acted unlawfully in making the TCLP more sensitive to lead and other inorganic constituents than was the Agency’s earlier leaching procedure, the EP. The Agency failed, they claim, (1) to give adequate notice of the scope of the TCLP; (2) to reconcile its conflicting views about the scope of the TCLP; (3) to provide an adequate statement of the basis and purpose of the TCLP rule; and (4) to act in accordance with its congressional mandate. Our review of these contentions indicates *450 that EPA’s promulgation of the TCLP was reasonable and proper. A Notice The petitioners contend that EPA failed to provide adequate public notice, as required under 5 U.S.C. § 553(b)(3), that the TCLP would be more sensitive than the EP to lead and other inorganic constituents. The petitioners rely most heavily upon a passage in the Regulatory Impact Statement that accompanied the proposed TCLP: The existing and proposed regulations do not differ in their treatment of metals. Thus, any impact of the proposed regulation on the municipal sector would be due solely to the additional organic lie., non-metallie] compounds. 51 Fed.Reg. 21,661. On initial inspection, this passage might indicate that the new TCLP would not differ from the EP in its sensitivity to lead and other metals. EPA, however, presents a number of persuasive reasons why this reading would be inaccurate. First, in context the passage is seen to refer only to the number of metallic constituents covered by the TCLP, not to the sensitivity of the test with respect to those constituents. It makes sense to view the passage this way because the TCLP added 25 new organic constituents not previously regulated under the EP, but added no new metals. Second, even if the passage were read to refer to the relative sensitivity levels of the old and new leaching procedures, the preceding paragraph indicates that any such comparison is limited to the effect of the new TCLP on sewage sludge generated by the municipal sector and does not extend to inorganic constituents generally. The petitioners also cite several passages indicating that EPA’s primary purpose in developing the TCLP was to improve the regulation of organic constituents only. For example, in initially proposing the TCLP, EPA explained that [t]he acetic acid models primarily the leaching of metals from an industrial waste. The impetus behind development of the [TCLP] was the need also to address the leaching of organic compounds. ■ EPA’s intent, then, was to develop an improved leaching test method suitable for use in evaluating wastes containing organic toxicants. 51 Fed.Reg. 21,653. As EPA notes, however, this passage hardly indicates that inorganic constituents would be immune from more stringent regulation. In fact, on the same page of the Federal Register, EPA notes that among the “[o]ther objectives [of the TCLP] were ... that it also model the mobility of inorganic species.” Id. Four pages earlier, EPA states that the TCLP “has been developed to address the mobility of both organic and inorganic com-pounds_” Id. at 21,649. In addition, EPA notes that it gave express notice of the terms of the TCLP as ultimately adopted. The petitioners concede that “EPA adopted the TCLP test in substantially the same terms as proposed,” but contend that the “substance” of the proposal “changed dramatically” and in a manner inconsistent with the Agency’s description of the test. In essence, the petitioners seem to argue that the Agency statements discussed above somehow misled them into thinking that the test would not regulate metals more stringently — despite the petitioners’ actual knowledge of the terms of the proposed TCLP. As noted above, however, the Agency’s statements did not indicate that the new test would treat metals in the same way as had the EP. Finally, EPA notes that a number of industry groups submitted comments criticizing the increased sensitivity of the TCLP to inorganic constituents. Such submissions are not dispositive on the question of notice, see Shell Oil Co. v. EPA 950 F.2d 741, 750-51 (D.C.Cir.1991), but they are at least probative evidence that the notice given was adequate. Id. at 757. Taken together with the portions of the TCLP proposal quoted above, it appears that the Agency gave adequate public notice of the character and scope of the TCLP. B. Reasonableness Second, the petitioners contend that the adoption of the TCLP was arbitrary and *451 capricious because EPA incorrectly assumed that the TCLP would yield the same results as had the EP with respect to metals, such as lead. For its part, EPA need only repeat that it never assumed that the TCLP and the EP would yield the same results for inorganic constituents. Again, as noted above, EPA made it clear when it initially proposed the TCLP that the new test would affect the regulation of both organic and inorganic constituents. The petitioners also contend that adoption of the TCLP is inconsistent with EPA’s later decision, see 55 Fed.Reg. 22,660 (1990), to retain the EP for the purpose of evaluating lead and other metals following treatment. EPA responds that it retained the EP for inorganic constituents following treatment because “data used to develop the treatment standards for these wastes were based on EP toxicity leachate data” whereas similar data based upon the TCLP were not yet available. Id. This explanation for EPA’s limited retention of the EP certainly seems to be reasonable. In an area implicating the Agency’s scientific expertise, as does this, we require no more of it. See Thomas, 824 F.2d at 1216 (“Happily, it is not for the judicial branch to undertake comparative evaluations of conflicting scientific evidence. Our review aims only to discern whether the Agency’s evaluation was rational”). C. Statement of Basis and Purpose Next, the petitioners argue that EPA failed coherently to state the “basis and purpose” of the rule adopting the TCLP, as required under 5 U.S.C. § 553(c). They describe EPA’s representations concerning the TCLP as “confused” because the initial TCLP proposal stated that for inorganic constituents it would predict the same leaching potential as the EP, while the Agency later stated that the TCLP was more sensitive than the EP. This claim is merely a recharacterization of the point raised above. And, as noted twice before, EPA’s initial proposal never stated that the TCLP would produce the same results as the EP for inorganic constituents. D. Congressional Mandate Finally, the petitioners contend that EPA’s decision to adopt a more “aggressive” leaching procedure is inconsistent with the Agency’s congressional mandate. The petitioners rely primarily upon the legislative history of § 3001(g) of RCRA, which directs the Agency to make such changes in the leaching procedure as are “necessary to insure that it accurately predicts the leaching potential of wastes....” 42 U.S.C. § 6921(g). They note that the original Senate version of this provision required EPA to develop “more aggressive” leaching procedures but the Congress deleted this phrase in the final version of the bill. From this sequence the petitioners urge us to infer that the “Congress considered and rejected the Senate’s mandate for a more stringent test,” so that “EPA’s adoption of a more sensitive test was contrary to Congressional intent.” Assuming that the statute in this case is sufficiently ambiguous in the first place to justify recourse to legislative history, we need only note that the deletion of a word or phrase in the throes of the legislative process does not ordinarily constitute, without more, evidence of a specific legislative intent. See American Trucking Assn. v. USA, 344 U.S. 298, 309-10, 73 S.Ct. 307, 314, 97 L.Ed. 337 (1953); Cheney R.R. Co. v. ICC, 902 F.2d 66, 69 (D.C.Cir.1990). In this instance, moreover, the Conference Committee specifically noted that the revision of the Senate bill was “not intended to alter the substance of the directive.” H.R.Conf.Rep. No. 1133, 98th Cong., 2nd Sess. 105-06, reprinted in 1984 U.S.C.C.A.N. 5649, 5676-77. V. DEFERRED REGULATION OF UST WASTE Petitioner Natural Resources Defense Council challenges EPA’s decision to defer application - of the Toxicity Characteristic (promulgated under Subchapter C of RCRA) to certain petroleum-contaminated waste that originates in underground storage tanks (“USTs”). The NRDC contends that EPA’s decision to defer Subchapter C regulation of UST waste violates the clear mandate of that Subchapter, which contemplates regulation of *452 “all hazardous wastes” and makes no express provision for deferrals, however temporary. EPA deferred regulation of UST waste under Subehapter C on the ground that such waste is already regulated under Subchapter I of RCRA, and concurrent imposition of Subchapter C requirements could, if the two regulatory regimes are not carefully coordinated, seriously interfere with ongoing hazardous waste monitoring, treatment, and cleanup. Upon considering the alternative requirements of the two regimes, we conclude that EPA’s approach is a “permissible construction of the statute,” which must be upheld under Chevron, 467 U.S. at 842, 104 S.Ct. at 2781. (Needless to say, the Congress did not evince in RCRA any specific intent concerning the way in which EPA is to harmonize the concurrent application of Sub-chapters C and I as applied to UST waste.) Initially, EPA contends that the plain language of RCRA causes Subchapter C and Subchapter I to overlap in the regulation of petroleum waste such as UST waste. Section 6991(2) of 42 U.S.C. defines a “regulated substance” under Subchapter I as (A) any substance defined in section 9601(14) of this title (but not including any substance regulated as a hazardous waste under Subchapter [C] of this chapter), and (B) petroleum. (Section 9601(14) in turn lists those substances regulated under the Comprehensive Environmental Response, Compensation, and Liability Act, see 42 U.S.C. § 9601(14), and expressly excludes petroleum.) Subsection 6991(2)(A) operates in such a way that, apart from petroleum, there can be no overlap in the substances regulated under Subchapters C and I: once a substance is defined as a hazardous waste it drops out from under Subchapter I regulation and into the domain of Subchapter C. A problem arises, however — unless EPA’s temporary deferral is lawful — because Subchapter C operates to define petroleum as a hazardous waste, and the kiekout provision of § 6991(2)(A) does not apply to petroleum. Therefore, petroleum remains under Subchapter I regulation even if it is also found to be a hazardous waste subject to Subchapter C. EPA contends further that temporary deferral of Subchapter C regulation of petroleum is necessary in order to “integrate” the provisions of Subchapters C and I, and that it is therefore authorized by 42 U.S.C. § 6905(b)(1), which provides that EPA shall integrate all provisions of [RCRA] for purposes of administration and enforcement and shall avoid duplication, to the maximum extent practicable, with the appropriate provisions of ... such other Acts of Congress as grant regulatory authority to the Administrator. According to EPA subjecting UST wastes immediately to Subchapter C as well as to Subehapter I could (1) “overwhelm the hazardous waste permitting program and the capacity of existing hazardous waste treatment, storage, and disposal facilities,” and (2) “delay cleanups significantly and severely discourage ... self monitoring and voluntary reporting.” 55 Fed.Reg. 11,836 (1990). The NRDC responds that Subchapter C and Subchapter I need not overlap at all in the present case, reasoning as follows: Sub-chapter I regulates “petroleum” only so long as it is in a UST (or otherwise in an easily retrievable liquid form). Once “petroleum” migrates from a UST and contaminates surrounding soil, however, it ceases to be “petroleum” and becomes a “hazardous waste” subject to Subehapter C and exempt from Subchapter I under § 6991(2)(A). Voila! There is no overlap between Subchapters C and I, and therefore neither need nor authority to defer regulation under Subchapter C. While the NRDC’s suggested interpretation of the statute is surely plausible, indeed elegant, EPA’s conclusion that Subchapters C and I do overlap must be upheld. The Congress did not clearly address the interaction of Subchapters C and I in the context of petroleum wastes, and EPA’s interpretation is (like the NRDC’s) a “permissible construction of the statute.” Chevron, 467 U.S. at 842, 104 S.Ct. at 2781. First impressions aside, moreover, EPA’s interpretation does not create a true conflict between the two Subchapters. Rather, they overlap with respect to petroleum wastes; hence the need for the Agency to coordinate regulation under the two regimes, and temporarily to de *453 fer the new layer of regulation while it sorts out the details. Alternatively, the NRDC argues that EPA’s integration strategy impermissibly “overrides” the requirements of Subchapter C, see Chemical Waste Management v. EPA, 976 F.2d 2, 22 (D.C.Cir.1992) (Agency may not “accommodate” two statutes by “overriding” the requirements of one), and that the Agency may not exempt a hazardous waste from regulation under that Subchapter without express legislative authorization. See Environmental Defense Fund v. EPA, 852 F.2d 1309, 1315 (D.C.Cir.1988) (“Subtitle C regulation of wastes found hazardous is mandatory”). The NRDC’s point would have more force behind it had EPA purported permanently to exempt UST waste from Subchapter C regulation. The Agency’s merely temporary deferral of such regulation hardly seems to “override” the requirements of that provision. Indeed, the temporary deferral is justified precisely and only because it is a waystation on the road to a permanent accommodation of the two Sub-chapters. The Agency’s deferral decision is permissible under Chevron, therefore: the Congress nowhere specifically proscribed deferring application of Subehapter C to UST waste, and substantial administrative difficulties would arise if the Agency could not do so. See generally Pennsylvania v. Lynn, 501 F.2d 848 (D.C.Cir.1974) (agency may suspend programs temporarily in order to determine whether, as appeared, they produced results contrary to Congress’s purpose). Furthermore, even during the deferral period Subchapter I assures that the petroleum wastes at issue will be regulated in a manner “necessary to protect human health and the environment,” 42 U.S.C. § 6991b(a)—the same general standard animating Subchapter C. See 42 U.S.C. § 6922. In particular, Subchapter I requires EPA to collect detailed data on USTs, see 42 U.S.C. § 6991a, and to implement standards for UST leak detection systems, leak reporting, and corrective actions including tank shutdowns and the relocation of threatened populations. See 42 U.S.C. §§ 6991b(c), 6991b(h)(5). Pursuant to these provisions, EPA has promulgated extensive regulations governing the identification and cleanup of soil and groundwater contaminated by UST waste. See 40 C.F.R. § 280.60-67. The continuing applicability of these Subchapter I regulations supports EPA’s claim that it is reasonable and proper for it temporarily to defer Subchapter C regulation. Conclusion EPA has failed to provide sufficient record justification for the application of the TCLP generic mismanagement scenario to mineral processing wastes and manufactured gas plant wastes. We therefore grant the petitions challenging such application and remand to the Agency for further proceedings consistent with this opinion. The petitions are denied in all other respects. So ordered.