On October 27, 2020, the Tenth Circuit upheld a series of OSHA citations that were issued to an oil refining company following a fatal boiler explosion. In doing so, the Tenth Circuit upheld OSHA interpretations that expansively view the scope of a “process” that is subject to the Process Safety Management (PSM) standard. This decision has implications under both the OSHA PSM standard and the EPA Risk Management Program (RMP) standard, as the EPA RMP program uses a definition of process that generally adheres closely to the OSHA PSM definition.

Key Takeaways

• In a 2-1 decision in Scalia v. Wynnewood Refining Co., the Tenth Circuit held that the PSM standard unambiguously applied to a boiler that did not hold or process highly hazardous chemicals (HHCs). 

• In reaching this holding, the Tenth Circuit agreed with two significant OSHA interpretations:

  • the boundary of a process can extend beyond vessels and piping that contain HHCs, and can include vessels and piping that do not contain HHCs; and

  • the boundary of a process is not limited to vessels and piping that pose a risk of catastrophic release if there is interconnection.

• The Tenth Circuit did not address, however, what series of piping and other connections between vessels within a facility constitutes “interconnection” for the purposes of PSM.

• In light of this decision, employers that are subject to the PSM standard should consider whether they have defined their process boundaries consistent with the standard’s definition of “process” and OSHA’s interpretations.

Background

This case concerns a 2012 explosion of a steam boiler known as the Wickes boiler at a refinery in Oklahoma owned by Wynnewood LLC. The Wickes boiler produced steam that was used in the refinery’s fluid catalytic cracking unit (FCCU). The FCCU held above a threshold quantity of HHCs and was therefore a process subject to the OSHA PSM standard. The Wickes boiler, however, was not physically located in the FCCU. Through a series of pipe connections, the Wickes boiler delivered steam to the FCCU, and through a separate series of pipe connections, refinery gas created in the FCCU was piped to the Wickes boiler for use as fuel. The explosion occurred as a result of the overfeed of natural gas to the boiler during a start-up following a turnaround and resulted in two employee deaths. OSHA did not allege that the boiler itself held or processed HHCs.

After investigating the explosion, OSHA issued citations alleging PSM violations, which Wynnewood appealed. The Occupational Safety and Health Review Commission (OSHRC) upheld the citations, based on the conclusion that the Wickes boiler was part of a PSM process because of a series of interconnections between the vessels that managed HHCs and the boiler. OSHRC held that the use of the term “interconnected” in the PSM definition of “process” made it irrelevant whether the Wickes boiler was directly connected to, or involved with, the FCCU process. Instead, OSHRC determined, because steam was piped from the boiler to the FCCU, and refinery gas was piped from the FCCU to the boiler, the Wickes boiler and the FCCU were interconnected. Wynnewood appealed to the Tenth Circuit, arguing that the Wickes boiler was not part of a covered process.

Regulatory Interpretation: Kisor v. Wilkie

The issue of whether the Wickes boiler is part of a PSM process raises a question of how an agency’s interpretation of its own regulatory provisions should be construed. The Tenth Circuit’s analysis therefore analyzes this question using the framework known as Auer deference, based upon the Supreme Court’s decision in Auer v. Robbins, 519 U.S. 452 (1997), as recently modified by Kisor v. Wilkie, 139 S. Ct. 2400 (2019). Prior to Kisor, the Auer deference principle required courts to defer to an agency’s interpretation of its own ambiguous regulatory provision so long as the interpretation was based on some permissible construction of the rule. This gave significant discretion to an agency to fill in gaps where rule provisions were ambiguous. The Kisor decision significantly limited the deference that courts will apply to agency regulatory interpretations by establishing a three-part test to apply when questions of regulatory interpretation are raised:

1. Is the regulation “genuinely ambiguous,” even after considering all of the tools of statutory construction? If it is not, no deference is due to the agency’s interpretation.

2. If the regulation is ambiguous, does the agency’s reading fall within the bounds of “reasonable interpretation?”

3. If the agency’s interpretation is reasonable, does the character and context of the agency interpretation entitle it to controlling weight? If not, no deference will be due.

Kisor limits the circumstances under which judicial deference will be applied to agency regulatory interpretations and the Tenth Circuit Wynnewood decision exemplifies both the type of analysis that has become common in regulatory interpretation cases under Kisor, and the limitations that Kisor imposes on agency authority to fill in regulatory gaps.

Tenth Circuit Decision

Wynnewood argued before the Tenth Circuit that the Wickes boiler was not covered by PSM because it did not contain any HHCs and because it was not interconnected. Relying on Kisor v. Wilke, Wynnewood argued that the OSHA interpretation of the process boundary (found in the definition of the term “process”) was not entitled to deference. In a 2-1 decision, the majority of the Tenth Circuit panel disagreed, and using the three-step framework in Kisor found that because the definition of “process” is unambiguous on its face and the plain meaning of the term is clear, the agency was not entitled to any deference. The single judge in dissent found that the term is ambiguous and that judge would have applied all three steps in the Kisor framework.

The court then turned to the definition of “process” and applied it to the facts. In the PSM standard, the term “process” is defined as follows:

Process means any activity involving a highly hazardous chemical including any use, storage, manufacturing, handling, or the on-site movement of such chemicals, or combination of these activities. For purposes of this definition, any group of vessels which are interconnected and separate vessels which are located such that a highly hazardous chemical could be involved in a potential release shall be considered a single process.

While the parties agreed that the Wickes Boiler did not contain any HHCs, they disagreed on the impact of that fact. To resolve that dispute, the court read literally the two sentences in the definition. The first sentence says the term “process” means any activity involving an HHC, and the second sentence says interconnected vessels and separate vessels which are located such that an HHC could be involved in a release shall be considered a single process. The majority found on this point that “the definition of process unambiguously includes vessels which do not contain a highly hazardous chemical.”

The parties next disagreed on the extent to which interconnection could be applied to vessels that are well downstream of vessels containing HHCs. Wynnewood argued that a vessel could not be part of a process simply because it is interconnected to a PSM-covered process, but that instead an interconnected vessel must pose a risk of catastrophic release of HHCs. In a linguistic/grammatical analysis that is now becoming common in regulatory construction cases under Kisor v. Wilke, Wynnewood asserted that the language in the definition should be read such that “any group of vessels which are interconnected” and “separate vessels which are located” are two items in a series that are both modified by “such that a highly hazardous chemical could be involved in the potential release.”

The majority disagreed with Wynnewood’s interpretation. Instead, the court found that the phrase “such that a highly hazardous chemical could be involved in the potential release” applies only to “vessels which are located.” On this basis, the court determined that “the Commission did not err in concluding that the Secretary need not demonstrate that the Wickes boiler posed a risk of catastrophic release of highly hazardous chemicals in order to be part of a process. Rather, he need only prove that the boiler was interconnected with a PSM-covered process.”

While Wynnewood also took the position that the Wickes boiler was not “interconnected” with the FCCU, the court’s majority found that Wynnewood had waived its right to this argument because it did not include the argument in its opening brief and only addressed that position in its reply brief. As a result, the court simply adopted the OSHRC position on this point without analysis. Under the facts of the case below, a significant amount of piping and equipment ran between the last unit containing HHCs and the Wickes boiler. This means that the question of whether a series of connections in a complex facility constitutes “interconnection” under the definition of process is not fully addressed by this decision.

Implications

Two significant OSHA interpretations on the scope of the PSM standard have been upheld by this Tenth Circuit decision. First, the boundary of a process can extend beyond vessels and piping that contain HHCs, and can include vessels and piping that do not contain HHCs. Second, the boundary of a process is not limited to vessels and piping that pose a risk of catastrophic release if there is interconnection. 

The decision does not settle the question of whether interconnection exists where a significant amount of piping and equipment runs between the last vessel or pipe containing HHCs and a connected unit that does not contain HHCs (i.e., an indirect connection), where that connected unit does not appear to pose a risk of a release of HHCs. The OSHRC decision that Wynnewood challenged did take that position, however, and the failure of the court to address this leaves the regulated community with continued uncertainty on this important question. As a result of this decision, employers should expect that OSHA will continue to take an expansive view of the definition of “process” and should revisit their process boundaries and consider whether they need to be modified to be consistent with standard’s definition of “process” and OSHA’s interpretations.