Shell Oil Company v. Douglas M. Costle, as Administrator, U. S. Environmental Protection Agency Andthe United States Environmental Protection Agency

595 F.2d 224, 9 Envtl. L. Rep. (Envtl. Law Inst.) 20465, 13 ERC (BNA) 1219, 1979 U.S. App. LEXIS 14663, 13 ERC 1219

Opinion

WISDOM, Circuit Judge:

Shell Oil Company (Shell), the petitioner, operates a large petroleum refinery in Norco, Louisiana, about 30 miles north of New Orleans. In this appeal, Shell challenges the effluent limit allowed the refinery by the pollution permit issued to it by the Environmental Protection Agency (EPA). We find that the EPA’s calculation of the effluent limit is supported by substantial evidence in the record. We affirm, therefore, the decision of the EPA’s administrator, who refused to review the regional administrator’s determination that the EPA had correctly calculated the effluent limit,

I.

The Norco refinery has an intake capacity of about 300,000 barrels per day of crude oil and natural gas liquids. Shell uses a variety of petrochemical engineering processes at this refinery to produce gasoline as well as other products. 1 In one of the refinery’s many plants, Shell has two units designed primarily to produce ethylene and propylene, two petrochemicals that may be described in more general terms as olefins. We shall refer to these units as the “olefins plant”.

In 1974, pursuant to the 1972 amendments to the Federal Water Pollution Control Act, 33 U.S.C. § 1251 et seq., the EPA issued Shell a permit to discharge effluents from this refinery. Shell contested the terms of the permit, asserting that the EPA’s calculation of the refinery’s effluent limit failed to give proper credit for the effluents attributable to the olefins plant. Following an adjudicatory hearing before an administrative law judge, the regional administrator of the EPA held that the EPA’s calculations involving the olefins plant were correct. The administrator of the EPA refused to review that decision. In this appeal, Shell renews its argument that the EPA incorrectly calculated the effluent limit attributable to the olefins plant.

II.

In 1972, Congress extensively amended the Federal Water Pollution Control Act in an attempt “to restore and maintain the chemical, physical, and biological integrity of the Nation’s waters”. 33 U.S.C. § 1251(a). 2 These amendments established a new permit program, the National Pollutant Discharge Elimination System (NPDES). The NPDES permit establishes the conditions under which a refinery may discharge pollutants and states the types and quantities of pollutants that may be discharged. With certain exceptions not relevant to this case, no “point source” 3 may discharge any pollutant unless it has first obtained a valid NPDES permit from the administrator of the EPA. 33 U.S.C. §§ 1311, 1342. The administrator is authorized to issue a permit only if, among other things, the discharge of pollutants by a point source meets the standards of § 301 of the Act, which requires the achievement by July 1, 1977, of “effluent limitations” requiring the “best practicable control technology currently available”. 33 U.S.C. § 1311(b)(l)(A)(i). The amendments require that the administrator prescribe by regulation that high level of technology for each type of point source. 33 U.S.C. §§ 1311(b)(l)(A)(i), 1314(b); E. I. duPont de Nemours & Co. v. Train, 1977, 430 U.S. 112, 97 S.Ct. 965, 51 L.Ed.2d 204. In this scheme, then, “an NPDES permit serves to transform generally applicable effluent limitations and other standards . . . into the obligations (including a timetable for compliance) of the individual discharger”. EPA v. California ex rel. State Water Resources Control Board, 1976, 426 U.S. 200, 205, 96 S.Ct. 2022, 2025, 48 L.Ed.2d 578.

The EPA received the aid of the American Petroleum Institute (API) in developing regulations setting forth the amount of effluents to be allowed petroleum refineries. In 1972, the EPA and the API jointly sponsored a survey of a number of refineries in the United States. Shell’s Norco refinery was included in this survey. Each refinery surveyed reported what processes it used and the amount of waste it produced. To analyze these data, the EPA used a complex statistical technique known as regression analysis, which allowed the EPA to predict the effect of each refinery process on the total waste produced by the refinery. The EPA first determined that the 126 processes included in the EPA/API survey could be properly grouped into nine more general process categories:

[t]he term “point source” means any discernible, confined and discrete conveyance, including but not limited to any pipe, ditch, channel, tunnel, conduit, well, discrete fissure, container, rolling stock, concentrated animal feeding operation, or vessel or other floating craft, from which pollutants are or may be discharged.

1. crude processes;

2. cracking processes;

3. hydrocarbon processes;

4. lubes and greases;

5. coking processes;

6. treating and finishing processes;

7. first generation petrochemicals;

8. second generation petrochemicals;

9. asphalt production.

In the EPA’s regression calculations, six of the nine process categories proved significant in predicting the amount of refinery waste: crude processes, cracking processes, lubes and greases, coking processes, second generation petrochemicals, and asphalt production. The significance values of five of these six processes were positive numbers, which showed that their presence resulted in a relatively greater production of waste. The exception was second generation petrochemicals, which had a significance value of -6, indicating that its presence resulted in a relatively lesser production of waste. The EPA found that the remaining three processes — hydrocarbon processes, treating and finishing processes, and first generation petrochemicals — were not significant in predicting refinery waste. The EPA’s analysis is found in its Development Document for Effluent Limitations Guidelines and New Source Performance Standards for the Petroleum Refinery Point Source Category (1974). R„ 255-460.

Based on this analysis, the EPA promulgated regulations prescribing the effluent limitations for the petroleum refining industry. 40 C.F.R. Part 419. 4 The regulations divide the industry into five subcategories: topping, cracking, petrochemical, lube, and integrated. After the EPA has classified a refinery according to subcategory, the allowable amount of effluent for the refinery is calculated by multiplying three factors specified in the regulations: the effluent limitations factor, the size factor, and the process factor. The values for the effluent limitations factor and the size factor are obtainable simply by examining the appropriate tables in the regulations. 5 The value for the process factor, however, must be calculated.

The EPA designed the process factor to take into account the mix of processes found in a refinery. It is the direct result of the EPA’s regression analysis of the data gathered in the EPA/API survey on the effect of refinery processes on amount of effluents. To calculate the process factor, one must first determine the “weighting factor” for each process category found at the refinery. As set forth in 40 C.F.R. § 419.42(b), the weighting factors are as follows:

The weighting factor for a given process is equivalent to the significance value that process had in the EPA’s regression analysis. The three processes that had no significance value — hydrocarbon processes, treating and finishing processes, and first generation petrochemicals — receive a weighting factor of zero. Second generation petrochemicals, which had a -6 significance value, is also accorded a zero weighting factor.

The weighting factor for a process category is used to determine the “process configuration” for the process, according to the following equation:

40 C.F.R. § 419.42(b). The next quantity to be determined is the “refinery process configuration”, which is the sum of the process configurations for the several process categories present at the refinery. Id. Finally, the “process factor” for the refinery is determined by plugging the refinery process configuration into a table found in the regulations. 6

The EPA and Shell agree that the Norco refinery fits in the “cracking” subcategory of the regulations. 7 Counsel for Shell has offered us a graphic depiction of how the allowable amount of effluents for a cracking refinery is calculated. Because we find that illustration useful, we reprint it below.

Pet.Br. 15.

III.

Our discussion in Part II indicates that, in a somewhat roundabout way, the weighting factors accorded the processes of a refinery by 40 C.F.R. § 419.42(b) are important variables in the calculation of the refinery’s effluent limit. When the EPA calculated the effluent limitation and issued the NPDES permit for Shell’s Norco refinery, it treated the processes in use at the olefins plant as petrochemical processes, which under the regulations receive a weighting factor of zero. In this appeal, Shell contends that the olefins plant operations should be treated as thermal cracking processes, to which the regulations accord a weighting factor of +6. 8 We must uphold the EPA’s determination that the processes are petrochemical processes if it is supported by substantial evidence in the record. 5 U.S.C. § 706(2)(E); Marathon Oil Co. v. EPA, 9 Cir. 1977, 564 F.2d 1253, 1261 — 66. To resolve the dispute, we must examine in some detail the olefins plant processes and the effluent limitations regulations. 9

Shell and the EPA agree that the olefins plant uses a pyrolysis process, 10 in which feedstock is heated without a catalyst, to produce gasoline, liquid petroleum gas, fuel oil, and the petrochemicals ethylene and propylene, which both belong to the general, petrochemical category of olefins. In every day of operation, the olefins plant produces about 250,000 gallons of gasoline, 63,000 gallons of liquid petroleum gas, 17,000 gallons of fuel oil, 3.6 million pounds of ethylene, and one million pounds of propylene. According to the EPA, when all these products are measured in pounds, petrochemicals make up 64 percent of the daily output of the plant, with gasoline accounting for 29 percent, liquid petroleum gas five percent, and fuel oil three percent.

The regulations do not provide clear guidance for the resolution of this dispute. They do not define any of the processes identified in 40 C.F.R. § 419.42(b). Shell’s argument that the olefins plant processes should be considered to be thermal cracking is based on the general definition of thermal cracking provided in the Development Document underlying the regulations:

[H]eavy gas oil fractions . . . are broken down intoiower molecular weight fractions ... by heating, but without the use of a catalyst. Typical thermal cracking conditions are 480°-603 °C (900 o — 1100° F) and 41.6-69.1 atm (600-1000 psig). The high pressures result from the formation of light hydrocarbons in the cracking reaction (olefins, or unsaturated compounds, are always formed in this chemical conversion). There is also always a certain amount of heavy fuel oil and coke formed by polymerization and condensation reactions.

R. 291. Shell contends that, because the evidence in the record indicates that the olefins plant processes fit this general description, the EPA should have treated the processes as thermal cracking processes accorded a + 6 weighting factor in the calculation of the process factor.

The EPA counters that, because the ole-fins plant processes produce mostly olefins, the processes are “petrochemical processes” accorded a zero weighting factor by 40 C.F.R. § 419.42(b). The EPA argues that the regulations define petroleum processes by example. To support this contention, the EPA looks to subpart C of the regula tions, which defines the petrochemical subcategory of refineries. Subpart C applies to refineries that produce “petroleum products by the use of topping, cracking and petrochemical operations” and that do not include the processes specified in subparts D (lube subcategory) or E (integrated subcategory) of the regulations. 40 C.F.R. § 419.-30. Subpart C defines petrochemical operations as

the production of second generation petrochemicals (i. e. ■ alcohols, ketones, cumene, styrene, etc.) or first generation petrochemicals and isomerization products (i. e. BTX, olefins, cyclohexane, etc.) when 15 percent or more of refinery production is as first generation petrochemicals and isomerization products.

40 C.F.R. § 419.31(f). 11

The EPA argues that this provision indicates that olefins are petrochemicals. We do not dispute that characterization; indeed, neither does Shell. We do not agree, however, with the EPA’s further argument that this provision, by defining petrochemicals by example, also suggests that a petrochemical process is simply a process producing petrochemicals. As the definition of thermal cracking in the EPA’s own Development Document indicates, processes other than “petrochemical processes” also produce olefins.

The EPA is on much more solid ground when it seeks to support its position by reference to the method by which it determined the values of the weighting factors accorded various refining processes and the manner in which Shell has described the olefins plant processes. As discussed earlier, the EPA/API survey asked refinery operators to provide data on the processes in use at their refineries. There is substantial evidence in the record that, for purposes of this survey and for other purposes before the regulations became final, Shell considered the olefins plant processes to be petrochemical processes. Indeed, Shell criticized the proposed regulations because they gave a weighting factor of zero to such petrochemical processes. The EPA relied upon this characterization by Shell and other refiners that such processes are petrochemical processes when the agency used the data contained in the EPA/API survey to develop the weighting factors accorded each refinery process.

The EPA position is also bolstered by the testimony of two of its officials. Martin Halper, the EPA' official who had been project director for the Development Document and who drafted the regulations, stated that, in the development of the regulations, thermal cracking that produced mostly olefins was considered a petrochemical process accorded a weighting factor of zero. Kenneth Huffman, an EPA chemical engineer who drafted the NPDES permit for the Norco refinery, testified that he categorized the olefins plant processes as petrochemical processes because Shell’s own submissions to the EPA described the processes in question as petrochemical operations for ethylene production. According to Mr. Huffman, Shell described certain other operations at the Norco refinery as thermal cracking processes. Mr. Huffman also testified that the Development Document envisioned refinery processes as being defined in part by the products they produce.

Shell did not effectively rebut the EPA’s evidence. Its sole witness on this issue, J. V. Schauer, a senior staff engineer for Shell, did little more than attempt to find inconsistencies in the testimony of the EPA officials, 12 note that the regulations and the Development Document do not define petrochemical processes, and assert that the olefins plant processes fit the Development Document’s general definition of thermal cracking. Mr. Schauer acknowledged Shell’s earlier references to the olefins plant processes as petrochemical processes. He attempted to ameliorate the impact of that evidence by testifying that, at the time Shell made these references, no materials developed by the EPA explained the distinction between thermal cracking and petrochemical processes. This testimony does not help Shell’s cause, since to this day the regulations and Development Document do not clearly define these processes. Instead, we agree with the EPA that when it developed these regulations it relied upon the common usages of the refining industry to define refinery processes.

Upon reviewing the record, we find substantial evidence to support the EPA’s decision to categorize the olefins plant operations as petrochemical processes accorded a weighting factor of zero in the calculation of the process factor for the Norco refinery. We affirm, therefore, the agency’s decision refusing to increase the effluent limit for the refinery on this account.

AFFIRMED.

1 . The testimony of a Shell official revealed that the major processes in use at the refinery are crude desalting, crude distillation, catalytic cracking, hydrocracking, thermal cracking, thermal coking, catalytic reforming, hydrogen treating, alkylation, and asphalt manufacture. Shell uses these processes to produce gasoline, liquid petroleum gas, aviation turbine fuel, diesel fuel, furnace oil and heavy fuel oil, asphalt, coke, sulphur, and olefins (ethylene and propylene) for chemical manufacture.

2 . Congress also enacted major amendments to the Federal Water Pollution Control Act in 1977. Pub.L. No. 95-217, 91 Stat. 1566. These amendments have no bearing on the issue in this case.

3 . According to 33 U.S.C. § 1362(14),

4 . In American Petroleum Institute v. EPA, 10 Cir. 1976, 540 F.2d 1023, cert. denied, 430 U.S. 922, 97 S.Ct. 1340, 51 L.Ed.2d 601, the Court of Appeals for the Tenth Circuit upheld the validity of a portion of these regulations. The EPA argues that Shell’s position in this case is nothing more than an attempt to relitigate the validity of the regulations. We agree with Shell that it is seeking only judicial review of the application of the regulations to the Norco refinery.

5 . The values for these factors for a topping subcategory refinery are found in 40 C.F.R. § 419.12; for a cracking subcategory refinery, in 40 C.F.R. § 419.22; for a petrochemical subcategory refinery, in 40 C.F.R. § 419.32; for a lube subcategory refinery, in 40 C.F.R. § 419.-42; and for an integrated subcategory refinery, in 40 C.F.R. § 419.52.

6 . See regulations cited in note 4 supra.

7 . The regulations provide that a facility is a cracking refinery if it

produces petroleum products by the use of topping and cracking, whether or not the facility includes any process in addition to topping and cracking. The provisions of this subpart are not applicable however, to facilities which include the processes specified in Subparts C, D, or E of this part.

40 C.F.R. § 419.20.

8 . According to Shell, the Norco refinery effluent limit would be increased 16 percent if the olefins plant processes are considered to be thermal cracking accorded a weighting factor of + 6 rather than petrochemical processes accorded a weighting factor of zero.

9 . In so doing, we decline the EPA’s invitation to consider evidence attached to its brief that is not in the record. See Seacoast Anti-pollution League v. Costoe, 1 Cir. 1978, 572 F.2d 872, 878-82. cert. denied,-U.S.-, 99 S.Ct. 94, 58 L.Ed.2d 117; Roberts v. Morton, 10 Cir. 1976, 549 F.2d 158, 160, cert. denied, 434 U.S. 834, 98 S.Ct. 121, 54 L.Ed.2d 95; Tanners’ Council of America, Inc. v. Train, 4 Cir. 1976, 540 F.2d 1188, 1193 n. 13; Hooker Chemicals & Plastics Corp. v. Train, 2 Cir. 1976, 537 F.2d 620, 636.

10 . Pyrolysis may be defined as the “chemical decomposition or other chemical change brought about by the action of heat regardless of the temperature involved”. Webster’s Third New International Dictionary 1854 (1961).

11 . Shell argues that this definition of “petrochemical operations” should be applied to “petrochemical processes” as that term was used in determining the weighting factors found in 40 C.F.R. § 419.42(b). We disagree. Subpart C clearly indicates that the definition of “petrochemical operations” is to be used only for the subcategorization of refineries, not for the calculation of the process factor. 40 C.F.R. § 419.31.

12 . Shell argues vigorously that the following passage from the testimony of Mr. Halper, the draftsman of the regulations, is inconsistent with- the remainder of the EPA’s evidence:

Q. Would the thermal cracking for ethylene production normally also produce other products e. g. pyrolysis gasoline?

A. This question has no moment in the consideration of an effluent limitation, since the regulations are based on amount raw materi ai (crude oil or NGL) used and the processes employed. The regulation does not consider the products manufactured. In other words, the regulation is capacity and process oriented. This is unlike the regulations for many other industries which are product oriented, or set limitations based on amount of products produced.

R. 485, 490. Shell considers this testimony as supporting its position that the regulations look to the process involved, not the products produced. Of the nine general process categories used in the EPA’s regression analysis, however, only five have labels that appear to be oriented to processes: crude processes, cracking processes, hydrocarbon processing, coking processes, and treating and finishing processes. The other four — lubes and greases, first generation petrochemicals, second generation petrochemicals, and asphalt production — carry labels oriented to the products that are produced. It appears that the only way these four processes are defined is by the predominant products they produce. Thus, although a portion of Mr. Halper’s testimony is somewhat confusing, we believe that the context of his testimony indicates that he viewed the process involved at the olefins plant as a petrochemical process, even though it also produced some gasoline.