September 4, 2020
Andrew Wheeler
Office of the Administrator
U.S. Environmental Protection Agency
Room 3000
WJC South Building
1200 Pennsylvania Ave., NW
Washington, DC 20460
Wheeler.andrew@Epa.gov
Re:

Petition for Reconsideration of Generic Maximum Achievable Control Technology
Standards Residual Risk and Technology Review for Ethylene Production; Final Rule, 85
Fed. Reg. 40,386 (July 6, 2020), Docket No. EPA-HQ-OAR-2017-0357.

BY E-MAIL AND FIRST-CLASS MAIL
Dear Administrator Wheeler:
This is a petition under Clean Air Act (“CAA” or “the Act”) § 307(d)(7)(B), 42 U.S.C.
§ 7607(d)(7)(B). The parties submitting this petition are RISE St. James (8581 Hwy 18, St.
James, Louisiana 70086), Air Alliance Houston (2520 Caroline St., Houston, TX 77004), Center
for Biological Diversity (P.O. Box 710, Tucson, AZ 85702), Clean Air Council (135 S. 19th St.,
Suite 300, Philadelphia, PA 19103), Community In-Power & Development Association (600
Austin Avenue, Port Arthur, TX 77640), Environmental Integrity Project (1000 Vermont Ave.
NW, Suite 1100, Washington, D.C. 20005), Louisiana Bucket Brigade (4226 Canal St., New
Orleans, LA 70119), Louisiana Environmental Action Network (P.O. Box 66323 Baton Rouge,
LA 70896), Sierra Club (2101 Webster Street, Suite 1300, Oakland, CA 94612), and Texas
Environmental Justice Advocacy Services (“t.e.j.a.s.”) (900 North Wayside Drive, Houston, TX
77023).
Petitioners request that EPA reconsider certain aspects of the final action taken at 85 Fed.
Reg. 40,386 (July 6, 2020), and titled “National Emission Standards for Hazardous Air
Pollutants: Generic Maximum Achievable Control Technology Standards Residual Risk and
Technology Review for Ethylene Production; Final Rule” (“Final Rule” or “2020 Rule”), and
strengthen the Final Rule to protect public health and satisfy the Clean Air Act. As discussed
further below, EPA must grant reconsideration because it has failed to provide the requisite
opportunity to comment on key rationales that the agency newly submits as bases for its
finalized, harmful, and unlawful exemptions, for an inadequate health risk assessment and risk
determinations, and for its refusal to require fenceline monitoring as necessary. Several of those
rationales make clear that the aforementioned EPA actions in this rulemaking are arbitrary and
capricious under the Act. EPA also has failed to provide the requisite opportunity to comment
on, or a lawful or rational basis for, its arbitrary and capricious refusal to use new health risk
1001 G STREET, NW,
T: 202.667.4500

F: 202.667.2356

SUITE 1000

WASHINGTON, DC

E: dcoffice@earthjustice.org

20001

W: www.earthjustice.org

 information on toxic pollutants, such as benzene and toluene. Because of the massive proposed
growth of this dangerous industry sector, which emits local toxic air pollution and causes
tremendous greenhouse gas impacts, granting reconsideration and strengthening these standards
is of the greatest importance and urgency.
FACTUAL AND LEGAL BACKGROUND
Ethylene production facilities spew over 4,000 tons per year of toxic air pollution into the
atmosphere along with approximately 64 million tons per year of greenhouse gases. Proposed
Rule, 84 Fed. Reg. 54,278, 54,327 (Oct. 9, 2019). 1 The vast quantities of pollutants emitted by
ethylene facilities can cause cancer and an array of adverse health impacts for exposed
populations, including serious neurological and respiratory problems. Final Residual Risk
Assessment (“RRA”) (EPA-HQ-OAR-2017-0357-0070) at 5-6.
This rulemaking is critically important to protect public health of communities now
exposed to ethylene production sources’ pollution – and for communities facing a potential
onslaught of new polluting sources. The petrochemical industry has indicated that it seeks to
build hundreds of facilities in high cancer-risk communities across the United States. 2 EPA
anticipates “massive growth” for this sector will increase toxic pollution in communities already
exposed to dangerous levels of carcinogens and other harmful pollutants. 3 EPA cites estimates
from the U.S. Department of Energy that capacity for ethylene production (and intermediate
petrochemical products) in the U.S. is expected to increase by over 85 percent between 2018 and
2040. 4 The planned FG LA, LLC (“Formosa”) Chemical Complex in St. James, Louisiana, for
example, would bring 28 ethylene cracking furnaces, and the facility as a whole would emit over
800 tons per year of hazardous air pollutants including known carcinogens like ethylene oxide,
benzene, 1, 3-butadiene, and formaldehyde in the epicenter of Cancer Alley, one mile from an
elementary school, and in a majority African American district where some residents live just a

1

See CBD Petition (2019) (attached to CBD Comment), EPA-HQ-OAR-2017-0357-0050 (citing CIEL
Plastics & Climate Report at p. 46, citing U.S. Environmental Protection Agency, GHG Reporting
Program Data Sets, Greenhouse Gas Reporting Program,
https://www.epa.gov/ghgreporting/ghgreporting-program-data-sets).
2
Id. at 4-5 (citing data from American Chemistry Council seeking to build or expand 333 facilities and
increase North American plastic production by 35% by 2025).
3
See Response to Comments (EPA-HQ-OAR-2017-0357-0074) at 6 (“RTC”); see also Final RRA (EPAHQ-OAR-2017-0357-0070) App. 1 at 26 tbl.2 (Nov. 2017 Memo on Emissions Data Used: predicting
“large growth” and discussing proposed expansions at five facilities and five new sources in Texas,
Louisiana, and Pennsylvania).
4
U.S. Department of Energy, Report to Congress: Ethane Storage and Distribution Hub in the United
States, at 7 fig.ES-4 (Nov. 2018),
https://www.energy.gov/sites/prod/files/2018/12/f58/Nov%202018%20DOE%20Ethane%20Hub%20Rep
ort.pdf.

2

 half mile of the planned complex. 5 Four of Formosa’s 14 planned units, if allowed to move
forward, would be subject to the ethylene production rule. 6
EPA found a high level of cancer and other health risks – including a lifetime cancer risk
of at least 100-in-1 million from inhalation of ethylene production sources’ emissions alone, and
this is likely higher based on the record, as discussed later in this petition. EPA also found that
exposure to ethylene production sources’ pollution, alone, causes an extra case of cancer every
decade – which EPA should find unacceptable, especially in view of the cumulative exposure
communities experience from ethylene production and collocated sources. 7 EPA’s risk
assessment numbers do not fully capture the real-world impacts that communities face. EPA’s
risk assessment likely significantly undercounts the health risks for fenceline communities as
Petitioners’’ 2019 Comments explained. Petitioners’ 2019 Comments (Dec. 6, 2019), (EPA-HQOAR-2017-0357-0044) at 24-63. Indeed, for years the ethylene production sector has had a
pattern of repeated upsets and malfunctions that have gone unremedied and undercounted under
the previous, unlawful general exemption for such emissions. 8
The Clean Air Act requires EPA to regulate hazardous air pollution from ethylene
production to protect public health with “an ample margin of safety.” 42 U.S.C. § 7412(f)(2) ;
see also § 7401(a) (purpose); see also §§7412(b), (d)(1)-(3), (6) (requiring control of hazardous
air pollutants emitted by ethylene production sources). Pollution controls and monitoring
methods are widely available, yet so far EPA has refused to set emission standards that ensure
the health protection the Act requires under § 7412(f)(2) and (d)(6).
The Act directs EPA to assess health risk remaining within eight years of promulgating
§ 7412(d) emission standards, i.e. the “residual risk.” Based on this assessment, EPA must set
emission standards to protect “the individual most exposed” to ethylene production facilities’
pollution by eliminating all unacceptable health risk, and providing an “ample margin of safety
to protect public health.” 42 U.S.C. § 7412(f)(2) (emphasis added). The term “ample margin of
5

Final RRA (EPA-HQ-OAR-2017-0357-0070) App. 1 at 27; RISE St. James comments on proposed
Formosa permits at 6 (Aug. 12, 2019) (attached to Petitioners’ Comments, EPA-HQ-OAR-2017-03570044).
6
The proposed facility’s air permits are currently on appeal by RISE St. James, Louisiana Bucket
Brigade, Center for Biological Diversity, Sierra Club, and other groups, represented by Earthjustice.
https://earthjustice.org/news/press/2020/appeal-challenges-louisianas-air-permits-for-formosas-massivepetrochemical-complex-in-cancer-alley.
7
See 85 Fed. Reg. at 40,392; Final RRA at 6, 40.
8
The record contains data on startup, shutdown, and malfunction releases submitted from ethylene
production sources in response to EPA’s Information Collection Request, though these data sets are likely
underestimates. ICR Data Memo (EPA-HQ-OAR-2017-0357-0004). See also EIP, Plastics Pollution on
the Rise at 3-5 (Sep. 5, 2019), https://www.environmentalintegrity.org/wpcontent/uploads/2019/09/Plastics-Pollution-on-the-Rise-report-final.pdf (citing pollution during upset or
malfunction incidents from ethylene production and other plastics plants including Exxon Baytown and
other sources, for which there was no state enforcement). EIP also found that: “From 2015 to 2017, Texas
or EPA only imposed penalties on 7 percent (57 out of 872) of unpermitted pollution releases from the
plastics industry in the Houston region during malfunctions or other industrial “upset” incidents. These
872 incidents released 11 million pounds of air pollution, but the fines were minimal, totaling only about
6 cents per pound of illegal pollution.” Id. at 7.

3

 safety” in § 7412, as interpreted by the court with respect to § 7412(d)(4), means an actual
protective buffer. Sierra Club v. EPA, 895 F.3d 1, 13 (D.C. Cir. 2018).
Pursuant to § 7412(d)(6), EPA also must “review, and revise as necessary” emission
standards promulgated for a source category pursuant to § 7412 every eight years. 42 U.S.C.
§ 7412(d)(6). This includes assessing and ensuring the standards “tak[e] into account,” i.e.,
reflect and follow, “developments” in pollution controls, practices, and technologies. Id. Where
a development exists demonstrating the achievability of further emission reductions or control
improvements, EPA must revise the standards to reflect them and ensure the “maximum” degree
of emission reductions that is “achievable.” 42 U.S.C. § 7412(d)(2). In addition, EPA is required
to make all other changes “necessary” to assure compliance with the Act—such as removing
illegal exemptions, and setting limits on uncontrolled or inadequately controlled emissions.
Louisiana Envtl. Action Network v. EPA, 955 F.3d 1088, 1096 (D.C. Cir. 2020) (“LEAN”).
Ethylene Production Rule Proposed Revisions
In 2019, EPA issued a proposed rule that included a health risk assessment and residual
risk proposal for ethylene production facilities pursuant to § 7412(f)(2). EPA also proposed
action under § 7412(d)(6) to “review and revise, as necessary,” the ethylene production standards
previously promulgated, including to set certain emission limits for the first time pursuant to
§ 7412(d)(2)-(3). 84 Fed. Reg. at 54,278.
In that proposal, EPA recognized that emission standards must be “continuous” and apply
at all times under the Clean Air Act and D.C. Circuit precedent. Id. at 54,320. EPA proposed to
strengthen emission standards applicable to specific emission points, including by adding
operational and monitoring requirements for flares, requiring repair of additional leaks from heat
exchange systems, and setting work practice standards for the first time for air toxics emissions
from decoking operations of ethylene cracking furnaces. EPA also removed the unlawful general
exemption for startup, shutdown, malfunction emissions and specific exemptions for particular
equipment, and required electronic reporting. EPA also proposed to strengthen the storage vessel
standards to require certain storage vessels with lower capacity and maximum true vapor
pressure to reduce and control organic HAP emissions.
However, EPA unlawfully proposed new malfunction exemptions from the emission
standards for flares and pressure relief devices (“PRD”), allowing exceedances once or twice
every three years with no consequence. EPA also proposed blanket exemptions for such
exceedances any time they are associated with a so-called “force majeure event,” such as loss of
external power or severe weather, that the Administrator deems to be beyond the facility’s
control.
EPA additionally performed an assessment of health risks from refineries and found
health risks to be “acceptable” even though the total maximum individual risk from routine
inhalation alone was 100-in-1 million —in addition to other serious health risks— and even
though ethylene production pollution disproportionately affects children, communities of color,
and low-income people. Final Residual Risk Assessment at 5-6; 84 Fed. Reg. at 54,330 (citing

4

 disproportionate health risks for individuals who are African American, Hispanic or Latino, Over
25 Without a High School Diploma, and Below the Poverty Level.)
In 2019, Petitioners filed detailed comments explaining why EPA should remove the
PRD and flare malfunction exemptions from the rule. 9 Petitioners also commented that the Act
and EPA’s own scientific guidance require that the agency add all cancer risks together in order
to meet EPA’s responsibility to assess risk and protect the “individual most exposed” to ethylene
production facility pollution. Doing so would tip the cancer risk above 100-in-1 million, EPA’s
presumptive acceptability “benchmark” under § 7412(f)(2). Petitioners commented that, due to
the cumulative risks ethylene production pollution causes to neighboring communities, including
from high acute risk, EPA must find that the health threats are “unacceptable” and set stronger
emission standards for such facilities that reduce emissions and provide an “ample margin of
safety to protect public health.” 42 U.S.C. § 7412(f)(2). Petitioners additionally urged EPA to
require fenceline monitoring for fugitive emissions from ethylene production facilities to capture
uncounted emissions harming communities living proximately to these facilities, noting that the
agency had proposed and finalized fenceline monitoring for the substantively similar refineries
source category.
2020 Rule
In the Final Rule, EPA largely rejected Petitioners’ 2019 comments, refused to remove
the unlawful exemptions, and declined to recognize that health risks from ethylene production
sources are unacceptable or strengthen the emission standards. 85 Fed. Reg. at 40,386. In the
Final Rule, EPA relied on new rationales and new data for these conclusions that it has not
subjected to public notice-and-comment.
In particular, EPA attempts to justify the flare and PRD exemptions with new rationales
not presented for notice-and-comment:
(1) EPA provides a new rationale for its contention that the exemptions for PRDs and
flares do not violate the Act’s requirement for emission standards to be “continuous,”
contending that some components of work practice standards apply during these
periods, and contending that a flare is not an emission point that need satisfy § 7412
of the Act;
(2) EPA newly claims that the exemptions meet §§ 7412(h) and 7412(d)(2)-(3) even
though they do not match the emission limitations achieved by sources operating
under the California standards that EPA described as the “floor.”
For years, these releases have increased the pollution to which Petitioners’ members are
frequently exposed. 10 The data EPA collected for this rule show approximately 578 reported
malfunction incidents over a 5-year period, with hundreds of thousands of pounds of HAPs
released. See ICR Data Memo and App. D. While preparing for this filing, some of Petitioners’
members experienced new spikes in toxic air pollution due to ethylene production sources’
9

Petitioners’ 2019 Comments (Dec. 6, 2019), EPA-HQ-OAR-2017-0357-0044.
See, e.g., sources cited in supra note 5.

10

5

 inadequate or delayed preparation for a hurricane – this time, Hurricane Laura. 11 For example,
the ethylene production plant Motiva (formerly Flint Hills) in Port Arthur, Texas experienced a
leak and severe flaring – and CIDA members also witnessed substantial flaring at the nearby
Chevron Phillips and BASF Fina facilities (also ethylene production sources). 12 EPA’s illegal
malfunction exemptions perpetuate health impacts and injustice by granting regular three-year
allowances for this pollution from PRDs and flares, and allowing these releases any time there is
a so-called “force majeure event.”
EPA also provides new rationales to try to justify its finding that the health risks to
communities are acceptable:
(1) EPA newly claims it is not appropriate, from a scientific standpoint, to sum the
multipathway (ingestion) cancer risk screening value that the agency developed with
inhalation risks, in order to assess total cancer risk.
(2) EPA newly argues that it can ignore its own guidelines and the best available science
finding that cancer risk is additive by assuming that the person most exposed to the
highest inhalation risks is not also exposed to the highest multipathway risk.
Thus, EPA now unlawfully attempts to justify its determination that risk is acceptable and
evade its own longstanding benchmark of 100-in-1 million, by contending it can ignore excess
cancer risk above that benchmark, so long as the additional cancer risk does not stem from
inhalation alone.
In the Final Rule, EPA refused to mandate fenceline monitoring for ethylene production
to bring their toxic fugitive emissions under better control by efficiently assessing fugitive
emissions and mandating corrective action for fenceline benzene concentrations above a set
threshold. EPA newly attempts to justify its failure to require fenceline monitoring:
(1) EPA newly argues that the agency has broad discretion to decline to require fenceline
monitoring for this source category under law.
(2) EPA suggests, for the first time, that despite the fact that the agency had required
fenceline monitoring for the Refineries Sector, ethylene production facilities’ fugitive
emissions are less likely to be “understated” as compared to those from refineries.
11

Rebecca Hersher, Millions Of Pounds Of Extra Pollution Were Released Before Hurricane Laura's
Landfall, NPR.org (Aug. 28, 2020), https://www.npr.org/sections/healthshots/2020/08/28/906822940/millions-of-pounds-of-extra-pollution-were-released-before-laura-madelandfall (attached); Zoya Teirstein, Even shut down, Texas oil refineries in Hurricane Laura’s path will
emit nearly 4 million pounds of pollution, Grist.org (Aug. 26, 2020), https://grist.org/climate/even-shutdown-texas-oil-refineries-in-hurricane-lauras-path-will-emit-nearly-4-million-pounds-of-pollution/
(attached); Ron Brackett, Oil, Chemical Plants Released Tons of Pollutants While Shutting Down for
Hurricane Laura, weather.com (Aug. 31, 2020), https://weather.com/news/news/2020-08-31-hurricanelaura-pollution-refineries-chemical-plants.
12
Video of flaring at Motiva, Port Arthur, TX, by Hilton Kelley, CIDA (posted Aug. 29, 2020 at 1:01
pm), https://www.facebook.com/hilton.kelley/videos/10157712254168233/; see also Erin Douglas,
Facilities in the Beaumont and Port Arthur area estimated that more than 4 million pounds of emissions
were put in air from closures, Houston Chron. (Aug. 29, 2020) (attached).

6

 (3) EPA claims that comments received at proposal did not highlight particular,
unregulated fugitive emissions demonstrating the need for fenceline monitoring and
that the ethylene production facility with the highest maximum individual cancer risk
(“MIR”) from inhalation, Shell Norco, is presently conducting fenceline monitoring
given that it is co-located with a petroleum refinery.
(4) EPA newly argues that the agency would violate notice-and-comment provisions if it
were to require fenceline monitoring in the Final Rule.
Further, in the 2020 Rule, EPA uses a different, undisclosed factor to assess acute risk than it
did in the 2019 proposed rule. EPA also provides two new rationales for not using the California
OEHHA reference exposure level (“REL”) to assess acute risk from benzene, flouting the
agency’s own scientific guidelines:
(1) EPA asserts that it “disagrees with some aspects of the methodology used to derive
the RELs for 1,3-butadiene, benzene, and nickel, and the derivation methodology
may partly underlie the apparent discrepancy between the 1- to 14-day MRL and the
1-hour REL for benzene;”
(2) EPA states that the ATSDR acute inhalation MRL represents the daily human
exposure to benzene that is likely to be without appreciable risk of adverse noncancer health effects for up to 14 days of exposure. EPA claims that it has therefore
not used the acute REL for benzene (0.027 mg/m3) because, EPA asserts, the alleged
similarity between these values “indicates that exposure to this concentration for up to
2 weeks is not expected to be associated with adverse effects.” RRA at 22.

EPA’s new arguments are misleading, unfounded and without evidence. Further, because
EPA’s belated presentation of these new rationales prevented Petitioners from raising critical
objections at proposal, EPA’s new rationales violate notice and comment. 42 U.S.C. §
7607(d)(3)-(6), (9), (h).
EPA also did not use the 2020 OEHHA health REL for toluene in the risk assessment. 13
As those were just finalized in August 2020, and are values EPA considers authoritative as the
best available science, and because these factors show toluene causes greater acute and chronic
non-cancer risks than EPA considered, EPA must update the risk assessment to apply these
factors here.
Because EPA did not present its new rationales and new data was not available for public
notice-and-comment until the 2020 Final Rule, it was “impracticable” for Petitioners to object in
their 2019 comments and show why EPA’s regulatory provisions based on these rationales are
unlawful and arbitrary for the reasons provided in this petition. 42 U.S.C. § 7607(d)(7)(B).
Further, the grounds for the objections discussed in this petition arose after the closure of the
period for public comment. Therefore, and because, as discussed below, Petitioners’ objections
are “of central relevance to the outcome of the rule,” EPA must grant reconsideration pursuant to
§ 7607(d)(7)(B). See, e.g., Chesapeake Climate Action Network v. EPA, 952 F.3d 310, 321 (D.C.
13

Cal. EPA OEHHA, Toluene RELs (Aug. 20, 2020),
https://oehha.ca.gov/media/downloads/crnr/toluenerel082020.pdf.

7

 Cir. 2020) (“CCAN”) (finding that EPA had unlawfully denied reconsideration where EPA failed
to disclose in the proposed rule or grant reconsideration regarding the agency’s “process” for
applying § 7412(d)(2)-(3), and the “critical reasoning behind its . . . analysis”).
I.

EPA MUST GRANT RECONSIDERATION ON ITS NEW RATIONALE FOR
THE PRESSURE RELIEF DEVICE AND SMOKING FLARE EXEMPTIONS.
A. Inability to Raise the Objection in Comments.

Cloaked as a novel work practice standard for pressure relief devices (“PRDs”) and
smoking flares, EPA in this rulemaking proposed and finalized exemptions removing liability for
one or two uncontrolled atmospheric releases of toxic air pollution during every 3-year period,
and during any so-called “force majeure event.” 40 C.F.R. § 63.1103, § 63.1107; 85 Fed. Reg. at
40,426, 40,433.
After Petitioners commented at proposal that the pressure release device (“PRD”) and
flare provisions contained unlawful exemptions for malfunction releases, 14 EPA in the Final
Rule advances a new rationale contending that “continuous,” though piecemeal, standards apply
to PRDs and flares, in lieu of removing these exemptions as the law demands.
EPA’s finalized malfunction exemptions are unlawful and arbitrary under the Clean Air
Act and D.C. Circuit precedent. In their 2019 Comments, Petitioners explained that allowing
facilities episodically to exceed the visible emissions requirement in the “emergency flaring”
work practice standard and to periodically release uncontrolled emissions from PRDs, means that
the standards do not apply at all times, in violation of the CAA. Petitioners’ 2019 Comments at
104-161. As Petitioners explained, it is unlawful for EPA to allow the flare and PRD standard
exceedances – which permit an unlimited amount of HAP pollution to be released to the
atmosphere – to be deemed “not a violation” in view of the plain language in § 7602(k) requiring
standards to be “continuous” and apply “at all times.” See Sierra Club v. EPA, 551 F.3d 1019,
1028 (D.C. Cir. 2008). Further, EPA has no statutory authority to allow uncontrolled emissions
from PRDs and smoking flares. Doing so flouts the plain text of the CAA and encroaches on
judicial power granted to courts by § 7604 to evaluate potential non-compliance and appropriate
civil penalties, not to EPA, under § 7413 or § 7412. NRDC v. EPA, 749 F.3d 1055 (D.C. Cir.
2014).
Lacking any rational justification for the finalized malfunction exemptions for flares and
PRDs, EPA in the Final Rule presents a new rationale to attempt to buttress the exemptions. EPA
newly argues that some, but not all component parts of the final work practice standards
constitute “measures [that] must be complied with at all times,” and thus that the standards
“apply at all times.” 85 Fed. Reg. at 40,406 (discussing the agency’s work practice standard for
PRDs); see also 85 Fed. Reg. at 40,404-405 (arguing that certain components of the agency’s
work practice standard for flares apply continuously).

14

Comments of Louisiana Bucket Brigade et al. (Dec. 6, 2019), EPA-HQ-OAR-2017-0357-0044.

8

 EPA’s new rationale is wholly unavailing and cannot satisfy the requirements of the
CAA or the D.C. Circuit’s mandate in Sierra Club. The law makes clear that standards compliant
with § 7412 must apply continuously. Sierra Club v. EPA, 551 F.3d at 1027-28 (D.C. Cir. 2008)
(“Congress has required that there must be continuous section 112-compliant standards.”;
“Congress … did not authorize the Administrator to relax emission standards on a temporal
basis.”). The allowance of episodic exemptions from emission standards cannot meet § 7412(h) –
which requires EPA to satisfy § 7412(d)(2)-(3) – because such allowances regularly provide for
no control of HAP emissions. Thus, work practice standards, which apply in only a piecemeal
fashion, could not reflect the maximum achievable degree of emission reduction by any source,
or the average emission limitation achieved by the best-performing sources, as § 7412(d)(2)-(3)
directs. Indeed, EPA in Sierra Club considered and rejected the argument that the continuous
application of meager, legally insufficient requirements could suffice to satisfy § 7412(d)(2)-(3).
551 F.3d at 1027-28.
Moreover, the Act requires control of all emitted HAPs. 42 U.S.C. § 7412(d)(1)-(3); Nat’l
Lime Ass’n v. EPA, 233 F.3d 625, 641 (D.C. Cir. 2000). The D.C. Circuit recently reaffirmed
this principle, holding that it is “necessary” under § 112(d)(6) for EPA to eliminate all
uncontrolled HAP emissions for this source category and to set any “missing limits.” LEAN, 955
F.3d at 1096 (“We read the statutory text to require EPA during its section 112(d)(6) review to
establish any missing limits.”). Therefore, leaving any HAP emissions uncontrolled – including
fugitive emissions from PRDs – violates the Act’s requirement that standards for HAPs must
apply at all times.
The uncontrolled emission loopholes within the final work practice standards also violate
§ 7412(f)(2) because the porous standards that EPA finalizes do not provide any margin of safety
to protect public health, much less an ample one. Unlike other limited statutory provisions that
allow for narrow “exceptional events” allowances, § 7412 allows for no such malfunction
exemption. 15
Because EPA had not previously attempted to contend that there are standards that apply
at all times during the malfunction periods, much less rely on such an argument as a rationale,
objecting to EPA’s novel contention during the public comment period was “impracticable”
within the meaning of Clean Air Act § 307(d)(7)(B). There was no reasonable way Petitioners
could have attempted to comment without seeing EPA’s rationale “in the [Notice of Proposed
Rulemaking].” CCAN, 952 F.3d at 320. The grounds for our below-discussed objections arose
after the close of the comment period under § 307(d)(7)(B), when EPA unveiled its new rationale
in the Final Rule.

15

Compare 42 U.S.C. § 7619(b) (addressing air quality monitoring data affected by an “exceptional
event”) with § 7410 (considering suspending state implementation plan requirements during an “energy
emergency”) and § 7545(c)(4)(C)(ii) (allowing temporary waiver of fuel or fuel additive controls or
prohibitions due to a “natural disaster, Act of God,” etc.).

9

 B. Objection: EPA’s Contention That It May Require Only Partial Standards
During Malfunctions Is Unlawful and Has Failed to Demonstrate That There
Are Continuous § 7412-Compliant Emission Standards Applicable to Pressure
Relief Devices and Flares, As the Act Requires.
To argue that standards apply at all times, the agency relies on a new claim that there are
some provisions in each work practice standard that do apply during the allowances even though
EPA undisputedly lifts other provisions that it put in place specifically to assure § 7412(h) and
§ 7412(d)(2)-(3) compliance. EPA’s new rationale is neither explained nor supported and is both
unlawful and arbitrary.
For PRDs, EPA states that facilities must comply with the preventative measures and
monitoring “at all times,” and must perform a root cause analysis where there is a force majeure
event. 85 Fed. Reg. at 40,406. For flares, EPA states that facilities must comply with
combustion efficiency standards, in particular, limits on the net heating value in the combustion
zone, at all times, including during a force majeure event. Id. at 40,404-405.
EPA’s argument fails, however. First, it is undeniable that EPA has lifted certain core
requirements for flares that are necessary for § 7412(h) and 7412(d)(2)-(3) compliance, and has
done so simply to allow facilities to avoid meeting the Act’s requirements during certain
malfunctions. Specifically, EPA has exempted flares from the visible emissions prohibition once
or twice every three years, and any time there is a “force majeure event” that is determined to
cause the release. 40 C.F.R. § 63.1103; 85 Fed. Reg. at 40,426. Similarly, although satisfaction
of § 7412 demands continuous prohibition of PRD releases, EPA has instead allowed periodic
uncontrolled, unlimited HAP releases from PRDs at least once or twice every three years, and for
an untrammeled number of instances during so-called “force majeure events.” 40 C.F.R. §
63.1107; 85 Fed. Reg. at 40,433.
Even assuming that the only requirements that apply during the exemption-triggering
events could constitute a “limitation” on emissions from ethylene production facilities (which
they cannot), Sierra Club makes clear that the requirements that EPA promulgates under § 112—
whether they are numeric emission limits or work practice requirements—must be both
“continuous” and “§ 112-compliant.” 42 U.S.C. § 7412(d)(2)-(3); Sierra Club, 551 F.3d at 102728. The D.C. Circuit has made it abundantly plain that even if EPA could craft a work practice
standard that could possibly “apply to the wide range of possible malfunctions,” “the EPA would
need to determine that the standard would ‘reduce emissions of hazardous air pollutants,’ an
evidence-based standard that is difficult (perhaps impossible) to apply to the unpredictable
circumstances of malfunctions.” U.S. Sugar Co. v. EPA, 830 F.3d 579, 608 (D.C. Cir. 2016)
(reh’g granted on remedy 844 F.3d 268 (D.C. Cir. 2016) (changing remedy to remand instead of
vacatur)). There is no applicable standard that “reduce[s] emissions” during the malfunction
periods that EPA exempts. Id.
Further, the text of the Clean Air Act makes plain that any work practice requirements
that EPA promulgates under § 112(h) must be consistent with § 112(d)—i.e. reduce emissions by
the “maximum” degree that is “achievable” and, at a minimum, to the level already “achieved”
by the relevant best performing sources. See also U.S. Sugar, 830 F.3d at 608 (“EPA would

10

 need to determine that the standard would ‘reduce emissions of hazardous air pollutants’”).
EPA’s work practices for pressure relief devices and flares are unlawful and arbitrary,
notwithstanding the agency’s new rationale for them, because the only requirements that are
“continuous” are not “§ 112-compliant.”
The work practice standards finalized by EPA for PRDs and smoking flares plainly
contain explicit exemptions that break the standards’ continuity. EPA has finalized gaping holes
in the emission standards during malfunctions that unlawfully excuse emissions that would
otherwise qualify as a violation. During the malfunctions the exemptions allow, there is simply
no § 7412-compliant standard in place. That deficiency is fatal to the standards’ legality. The
Act does not allow a piecemeal approach where, as here, EPA sets a standard that purports to
meet § 7412, but crafts regulatory provisions that periodically and unjustifiably lift key
requirements from that standard. See 42 U.S.C. § 7602(k).
None of the limited regulatory provisions that apply during the malfunction exemptions
negate the fact that EPA has excused facilities from the components of the standards that are
needed to satisfy § 7412(h) and 7412(d)(2)-(3). Petitioners do not discount the value of the
various measures that apply during these times such as the combustion efficiency requirement
for flares and the PRD release preventative measures. However, the purpose of those
requirements is to monitor the performance of equipment and personnel and/or to help prevent
additional future emissions releases, rather than to impose a limit on emissions or require
emissions reductions. Further, EPA has not, and could not, claim that implementation of the
continuous requirements represents the maximum achievable emissions reduction achieved by
the best performing sources, as the Act requires. Id. § 7412(d)(2)-(3). In particular:
•

The combustion efficiency requirements EPA points to are important but they still do
nothing to restrict the potentially extremely high net total amount of HAPs released
during exempted events. Indeed, EPA’s “above smokeless capacity” provision unlawfully
permits flares to operate inefficiently during malfunctions, by allowing operators to avoid
the visible emissions requirements without consequence, and to allow much higher than
usual HAP releases. 16 During the malfunction releases that EPA finalized, visible
smoking flares demonstrate to the naked eye that the destruction of HAPs is substantially
sub-optimal, making clear that the exemptions that EPA affords for visible flare
emissions do not assure the necessary maximum achievable degree of emission limitation
(“MACT”) restriction (98-percent HAP destruction efficiency) on HAPs routed to those
flares. Although EPA contends otherwise, EPA has not shown this and could not do so.
For example, in the 2015 Refineries Final Rule, EPA promulgated the entire “suite” of
flare requirements (i.e., visible emissions prohibition, maximum flare tip velocity,
combustion zone operating limits, and pilot flame requirements) to ensure that flares
reduce HAPs by 98 percent. 2015 Refinery Rule RTC at 83, EPA-HQ-OAR-2010-06820802; see also 80 Fed. Reg. at 75,182; 2015 Refinery Rule RTC at 8 (“we proposed and
are finalizing significant new operating and monitoring requirements for flares to ensure

16

As EPA recognized in 2015: “smoking flares indicate reduced combustion efficiency. Furthermore, . . .
[the no visible emissions] requirement is part of a suite of flare requirements that ensure that the flare is
well-operated as necessary to achieve the MACT requirements for affected emissions sources using flares
as a control device.”). 2015 Refinery Rule RTC at 83 (EPA-HQ-OAR-2010-0682-0802).

11

 flares are achieving the required control efficiency.”); Id. at 75 (“98 percent destruction
efficiency . . . forms the basis for the MACT flaring provisions”). All of these
requirements are needed to maintain 98 percent destruction efficiency. Yet EPA
illogically now argues that flares can still maintain a 98 percent destruction efficiency
even when core requirements no longer apply during periods of operation above
“smokeless capacity”—periods when HAP emissions will actually be greater than
periods of normal operation due to the increase in gases routed to flares. Further, when
smoking, a flare increases its emission of the products of incomplete combustion. 17 The
rule’s periodic lifting of the provisions prohibiting visible emissions allows flares to emit
an uncontrolled amount of HAPs, violating § 7412(d).
•

The PRD requirements that continuously apply impose no actual limit on PRD emissions,
require no actual reduction of PRD emissions, are extremely vague, and provide a
substantial number of alternate compliance options for facilities – including options that
solely require facilities to monitor equipment performance without mandating any
corrective action at all. EPA states that “[t]he work practice standard for PRDs requires a
number of prevention measures that operators must undertake to prevent PRD release
events, and the installation and operation of continuous monitoring device(s) to identify
when a PRD release has occurred.” RTC at 129. The provisions allow facilities to select
and implement any three “redundant prevention” measures, which include choices
amongst the following: “flow, temperature, liquid level and pressure indicators with
deadman switches, monitors, or automatic actuators . . . [d]ocumented routine inspection
and maintenance programs and/or operator training.” 40 C.F.R. § 63.1107. Prevention
and monitoring are not alone pollution controls, and do not alone constitute § 112compliant emission limitations. The “redundant prevention” measures are therefore
insufficient to assure § 7412 compliance and EPA has arbitrarily failed to make any
demonstration that implementation of such measures represent the average emission
limitation achieved by the best-performing sources or the maximum achievable degree of
emissions reduction. 42 U.S.C. §§ 7607(d)(9), 7412(d)(2)-(3).

Indeed, EPA could not do so. For example, EPA has not required any preventative
measures specifically designed to try to prevent and avoid smoking flares or PRD releases
connected to force majeure events. The Chemical Safety Board (“CSB”) has recognized many
17

EPA Flare Study (EPA-HQ-OAR-2010-0682-0191, incorporated into this docket, see 84 Fed. Reg. at
54,293); Cal. EPA OEHHA, Analysis of Refinery Chemical Emissions and Health Effects (Mar. 2019)
(“OEHHA 2019”); https://oehha.ca.gov/media/downloads/faqs/refinerychemicalsreport032019.pdf see
also Cheryl L. Weyant et al., Black Carbon Emissions from Associated Natural Gas Flaring, 54 Environ.
Sci. Technol. 2075–2081 (2016), https://doi.org/10.1021/acs.est.5b04712; Grzegorz Wielgosiński,
Pollutant Formation in Combustion Processes in ADVANCES IN CHEMICAL ENGINEERING 296-324 (Dr.
Zeeshan Nawas ed., 2012); Fahad M. Al-Fadhli et al., Impact of Flare Destruction Efficiency and
Products of Incomplete Combustion on Ozone Formation in Houston, Texas, 51 Ind. Eng. Chem. Res.
12663–12673 (2012); W.B. Kindzierski, Importance of human environmental exposure to hazardous air
pollutions from gas flares, 8 Environ. Rev. 41–62 (2000) (“emission of a variety of compounds such as
volatile organic compounds (VOC's), polycyclic aromatic hydrocarbon (PAH's) as well as soot occur
when the flaring process ensues with incomplete combustion”).

12

 such releases are avoidable, including in its 2017 Safety Alert for facilities on safe shutdown and
startup procedures and in its report on one of the major releases that occurred in the wake of
Hurricane Harvey in 2017. 18 As the CSB then-Chairperson Vanessa Sutherland explained in
releasing the report on the Arkema fire after Hurricane Harvey that included many such
measures designed to prevent so-called “natech” disasters and releases 19: “Considering that
extreme weather events are likely to increase in number and severity, the chemical industry must
be prepared for worst case scenarios at their facilities. We cannot stop the storms, but working
together, we can mitigate the damage and avoid a future catastrophic incident.” 20 One of the
preventative measures—the root cause analysis after a PRD release has occurred—is also
valuable and may help prevent the next release. But like the other preventative measures, this
does nothing to prevent or remedy the release that already occurred, and that it is the subject of
the analysis.
It is indisputable that none of the provisions EPA points to as applying continuously
during malfunctions actually limit the amount of HAPs released during a malfunction, much less
do so to the extent necessary to assure at least “the average emission limitation achieved” by the
best-performing sources, and to assure the “maximum achievable degree of emission reduction”
during such releases. 42 U.S.C. § 7412(d)(2)-(3). Allowing PRDs and flares to release an
uncontrolled amount of pollution during the exemption periods that EPA deems “not a violation”
thus means that there is no meaningful or legally sufficient limit on such pollution during these
malfunctions. By contrast, the D.C. Circuit found that, based on the unique circumstances
present in EPA’s MACT rule for industrial boilers (including EPA’s imperfect information and
identified safety hazards), the startup and shutdown work practice standards there passed muster
under 42 U.S.C. § 7412(d)(2)-(3) in part because they created no malfunction exemption and
actually restricted HAPs. Sierra Club v. EPA, 884 F.3d 1185 (D.C. Cir. 2018). 21
Contrary to EPA’s new rationale, the fact that some requirements apply during the
exempted malfunction events does not change the fact that they do not control emissions
continuously as required to meet § 7412(h) and § 7412(d)(2)-(3). The prohibition on visible
18

CSB, Safety Alert, After Harvey: Precautions Needed During Oil and Chemical Facility Startup (Aug.
2017), https://www.csb.gov/assets/1/20/csb_harvey2017_05.pdf; CSB, News Release: CSB Urges Oil and
Chemical Facilities to Take Special Safety Precautions during Startups Following Hurricane Harvey
(Aug. 27, 2017), https://www.csb.gov/csb-urges-oil-and-chemical-facilities-to-take-special-safetyprecautions-during-startups-following-hurricane-harvey/; CSB, Investigation Report No. 2017-08-I-TX:
Organic Peroxide Decomposition, Release, and Fire at Arkema Crosby Following Hurricane Harvey
Flooding (May 2018), http://www.csb.gov/file.aspx?DocumentId=6068.
19
The CSB uses the term “natech” event to mean a pollution release or industrial malfunction (“tech”) on
top of and related to a natural disaster such as a hurricane or earthquake.
20
CSB, News Release: CSB Releases Arkema Final Report (May 24, 2018), https://www.csb.gov/csbreleases-arkema-final-report/ (emphasis added).
21
The circumstances present in that case are not the same as the circumstances here. Further, the flare and
PRD provisions here cover malfunctions—not startup or shutdown periods. See U.S. Sugar, 830 F.3d at
608 (crafting § 112-compliant work practice standard is perhaps impossible for malfunction periods). And
unlike the startup-shutdown standards in Sierra Club, these do not contain any of the built-in limitations
the court found relevant that actually restricted the amount of HAPs sources emitted during those periods
as § 7412(h) and § 7412(d)(2)-(3) require.

13

 emissions requirements is essential to limit emission. A flare smokes once it is over its
“smokeless capacity” and is therefore emitting a higher quantity of HAPs, destroying a lower
quantity of HAPs, and creating more products of incomplete combustion than when it does not
smoke. Thus, for EPA to meet § 7412(d)(2)-(3), EPA must, at least, deem smoking to be a
violation. 22 Similarly, the prohibition on uncontrolled emissions from PRDs that applies only if
an exceedance is deemed to result from an operator or maintenance error, after a facility has once
or twice exceeded the limit, and if there is no precipitating “force majeure event,” is the only
restriction that limits the amount of emissions from these valves and similar devices. The PRD
is either open or closed. If it is open, it must be routed to a control device or it will emit freely
into the air. In order to comply with § 7412(d)(2)-(3)’s requirements for limiting PRD releases,
EPA must remove the exemptions.
EPA cannot justify the free-pollution allowances by suggesting that they are continuous
emission standards. Instead, they are non-standards. During certain periods of time, the full suite
of provisions controlling HAP emissions apply to flares and PRDs. However, during
malfunctions once or twice every three years, and during any “force majeure event,” the Final
Rule allows exemptions from critical requirements within the work practice standards;
requirements that the agency itself found important and necessary to meet the § 7412(d)
requirements for CAA-compliant emission standards. Under the Final Rule, if facilities release
HAPs that would otherwise exceed requirements, such an emissions event, in a number of
prescribed instances, is not deemed a violation, and the facility is not subject to civil penalties or
any other consequences. Indeed, EPA’s novel claim that the PRD and smoking flare exemptions
are justified because some form of preventative measure or monitoring requirement, no matter
how meager, applies consistently to those emissions sources is an argument that has been
rejected by the D.C. Circuit in the very case that EPA purports to address in this rulemaking. In
Sierra Club v. EPA, EPA attempted to cite analogous preventative measures and requirements
such as SSM plans and the “general duty” provision that applied continuously but did not limit
pollution during SSM periods to attempt to defend the legality of the agency’s unlawful
exemptions. In striking down the exemptions, the court held that the measures and requirements
that were continuously in place did not obviate the fact that no continuous § 112-compliant
emission standards applied during the SSM exemption. 551 F.3d at 1027-28.
EPA attempts to suggest, contrary to Petitioners’ comments at proposal, that stripping
citizens’ power in this way does not violate the Act’s citizen-suit and penalty provisions, stating
that “the regulations do not specify that the EPA Administrator would make a binding
determination regarding whether a PRD release is in compliance or a violation, and the issue
could be argued and resolved by a court in the context of a citizen suit.” 85 Fed. Reg. at 40,408.
EPA’s argument is nonsensical. EPA has expressly deemed the exempted emissions events not a
“violation” of the regulations. See, e.g., 85 Fed. Reg. at 40,433 (citing 40 C.F.R.§ 63.1107(h) and
prescribing PRD release events that constitute regulatory violations, specifically excluding all
first-time releases not deemed to result from “operator error or poor maintenance,” all secondtime releases that do not share a root cause with the first-time release, and all releases not
resulting from “force majeure events.” ). EPA has also made clear that certain uncontrolled
releases are “not a deviation” from the regulatory standards. See, e.g., RTC at 107-08. In such
22

EPA must also set additional restrictions on flares’ routine emissions, to satisfy § 7412(d)(2)-(3), as
Petitioners discussed in their 2019 Comments, as well as prohibiting smoking flares.

14

 instances, EPA has prevented an affected community member from bringing an enforcement suit
and has attempted to prevent a court from being able to adjudicate a violation of the emission
standards during malfunction periods. Under the regulations as written for the free passes to
pollute, EPA has wholly exempted uncontrolled releases from the standards. These exemptions
unlawfully encroach on judicial authority under 42 U.S.C. § 7412 and 7413, just as the court held
in NRDC. v. EPA, 749 F.3d 1055. A search of EPA’s enforcement database and experience of
Petitioners has found no significant enforcement of any kind due to the types of malfunction
releases this rule would allow at ethylene production sources or any similar sources, since
2016. 23
Further, it is arbitrary for EPA to contend that it can apply some, but not all, work
practice components to these pieces of equipment during the exempted periods—especially
when, as discussed below, the work practice standards EPA has set do not meet § 7412(d)(2)-(3).
See 42 U.S.C. § 7607(d)(9). The free passes EPA unlawfully authorizes constitute escape hatches
from the PRD and smokeless flare standards, giving the facility incentive to fail to invest
sufficient employee time and to fail to perform the checks needed to ensure that no such releases
occur. These escape hatches allow facilities to face no serious consequences beyond, at
maximum, the performance of a root cause analysis – and at minimum no consequence at all –
even if every PRD and every flare at that facility has multiple close consecutive releases,
propelling hazardous pollutants into the air. The record shows substantial emissions from
malfunctions, including from the types of malfunction releases exempted from meaningful
emissions controls here. See ICR Data Memo and App. D (providing database with information
received from sources on these releases, including at least 278,000 pounds labeled as
malfunction releases).
Other conclusions in the Final Rule further demonstrate that EPA’s rationale cannot save
its unlawful action. EPA admits that the PRD releases it is allowing to be uncontrolled “are
triggered by equipment or process malfunction.” 85 Fed. Reg. at 40,408 (emphasis added). EPA
further admits that the same is true for the flaring allowances: “these flare emissions are
emissions due to a sudden increase in waste gas entering the flare, typically resulting from a
malfunction or an emergency shutdown at one or more pieces of equipment that vents emissions
to the flare . . . during malfunction events.” Id. at 40,404 (emphasis added). 24 EPA’s concessions
reiterate that the permitted releases are transparently just another version of the same
23

EPA, Enforcement and Compliance History Online, https://echo.epa.gov/;
https://cfpub.epa.gov/enforcement/cases/. Further, EPA’s own enforcement has declined substantially in
recent years. See, e.g., EPA OIG, Report: EPA's Compliance Monitoring Activities, Enforcement Actions,
and Enforcement Results Generally Declined from Fiscal Years 2006 Through 2018, Report No. 20-P0131 (Mar. 31, 2020), https://www.epa.gov/office-inspector-general/report-epas-compliance-monitoringactivities-enforcement-actions-and.
24
EPA admits that: “Other than for those specific emission points discussed in section III.C of this
preamble, the EPA determined that no additional standards are needed to address emissions during
periods of SSM.” Id. at 40,390. And again EPA admits “as discussed in sections III and IV of this
preamble, we are adding separate work practice standards to the final rule for the following SSM
activities/events: (1) Periods of SSM for when flares are used as an APCD, (2) periods of SSM for certain
vent streams (i.e. PRD releases and maintenance vents), (3) vent control bypasses for certain vent streams
(i.e., closed vent systems containing bypass lines, in situ sampling systems, and flares connected to fuel
gas systems), and (4) decoking operations for ethylene cracking furnaces.” Id. at 40,391.

15

 malfunction exemption the D.C. Circuit struck down in 2008. See Sierra Club v. EPA, 551 F.3d
at 1028.
There can be no doubt that the PRD and smoking flare exemptions allow sources to avoid
the established emission standards during malfunction periods, as did the explicit SSM
malfunction exemption and affirmative defense to penalties provisions which the D.C. Circuit
previously found unlawful. Id.; see also NRDC v. EPA, 749 F.3d 1055 (holding that EPA could
not create an affirmative defense to civil penalties for malfunction releases “to account for the
tension between requirements that emissions limitations be ‘continuous’ and the practical reality
that control technology can fail unavoidably”). Indeed, the D.C. Circuit has specifically and
continually held that EPA’s authorization of facilities’ avoidance of emissions standards and
evasion of consequences for violating those standards contravenes the Act, notwithstanding
EPA’s repeated attempts to craft new versions of the same exemptions as it does here.
Malfunction exemptions, like those finalized, are simply not emission standards and cannot
survive judicial scrutiny.
EPA is wrong that the Clean Air Act does not require it to fix these legal defects,
including to control these uncontrolled emissions, when it promulgates RTRs. Section 112(d)(6)
requires EPA to review its § 112 rules every eight years and revise them “as necessary.” 42
U.S.C. § 7412(d)(6). It is “necessary” that EPA’s Clean Air Act rules comply with the Clean Air
Act. As The D.C. Circuit recently confirmed, “[t]he section 112(d)(6) requirement that EPA,
when it undertakes its eight-year review, revise emission standards ‘as necessary’ means that
EPA must conform them to the basic requisites of ‘emission standards’ under section 112,
including by setting controls on previously unaddressed hazardous air pollutants.” LEAN, 955
F.3d at 1098 (emphasis added). The PRD and smoking flare exemptions constitute uncontrolled
emissions that flout the court’s holding in LEAN—because EPA has left holes in the standards
that allow repeated PRD and flare releases without limits.
In its response to comments, EPA also incorrectly suggested that Petitioners have not
highlighted any unregulated fugitive emissions. RTC at 129. However, Petitioners described in
detail in their comments fugitive emissions from uncontrolled PRD releases (and products of
incomplete combustion). Petitioners’ 2019 Comments at 104-161. Cleanly removing the regular
exemptions (once or twice every three years) and the “force majeure event” exemptions would
satisfy EPA’s obligation under LEAN to revise the Ethylene Production standards “as necessary”
to assure controls on all emitted HAPs. LEAN, 955 F.3d at 1091, 1096-97; see also 42 U.S.C. §
7412(d)(6).
The Final Rule’s allowance of uncontrolled PRD and smoking flare emissions also
violates § 7412(d)(2) by failing, definitionally, to require the maximum achievable degree of
emissions reduction for these sources. That is, regulatory provisions providing for no control at
all during these periods cannot possibly achieve the maximum degree of reduction. EPA has no
authority to grant a broad exemption from § 7412 regulations by regulation.
That Congress specifically chose to provide EPA express authority to permit departures
from ordinarily applicable requirements during exceptional events or natural disasters in other
parts of the Act that do not address emissions standards, but did not do so here, indicates

16

 Congress’s clear intent not to allow any force majeure event or other malfunction exemption
applying to emissions standards similar to those created by the Final Rule. 25
C. This Objection Is of “Central Relevance” to the Rule’s Outcome Because Section
112-Compliant Emission Standards Must Be Continuous and the Exemptions
Have Dangerous Consequences for Public Health.
This objection is “of central relevance” § 7607(d)(7)(B) because it goes to the “legality”
of the PRD and flare exemptions, providing “substantial support for the argument that the
regulation should be revised.” CCAN, 952 F.3d at 320. It is indisputable, and EPA admits, that
ethylene production facilities must be subject to § 7412-compliant emission standards that are
“continuous” and apply “at all times.” 42 U.S.C. § 7602(k); see also 85 Fed. Reg. at 40,390. The
D.C. Circuit has repeatedly vacated other similar exemptions where EPA has attempted to justify
excusing non-compliance or lifting penalties in advance due to malfunctions and should do so
again if EPA does not remove the exemptions here. See, e.g., Sierra Club v. EPA, 551 F.3d
1019; NRDC v. EPA, 749 F.3d 1055.
The issue is also “of central relevance” because of the serious harm to communities
exposed to pollution from these exemptions, and the strong need for Petitioners to have EPA
reconsider this new rationale and recognize that it cannot allow the malfunction exemptions.
Extremely high amounts of toxic emissions are released directly into local communities’ air
during the types of routine malfunction events for which EPA’s standards allow uncontrolled
PRD and smoking flare pollution releases. For example, on August 9, 2015, a single PRD release
at Shell Deer Park released more than 150 tons of 1,3-Butadiene in less than one hour. 26 These
emissions increase cancer and other chronic and acute health threats to nearby communities
already facing extremely high health threats, as EPA’s risk assessment here and in the 2015
Refinery Rule show. 27 Thus, the pollution impact of these exemptions makes this objection an
important health issue that is of central relevance because of the Act’s purpose of preventing
pollution and protecting public health that § 7412 implements. 42 U.S.C. § 7401(a).
The objections carry particular relevance because they show the illegality of exemptions
that would exacerbate harm to public health and safety at times when the public most needs
pollution protection. For example, during natural disasters, which qualify under EPA’s
regulatory revisions as triggering potential “force majeure event” pollution exemptions,
emissions and health threats from air pollution increase substantially. During such events,
25

See e.g., 42 U.S.C. § 7619(b) (addressing air quality monitoring data affected by an “exceptional
event”); § 7410 (considering suspending state implementation plan requirements during an “energy
emergency”); § 7545(c)(4)(C)(ii) (allowing temporary waiver of fuel or fuel additive controls or
prohibitions due to a “natural disaster, Act of God,” etc.).
26
2016 Refinery Rule Comments at 18-20 (citing TCEQ Emissions Event Inventory, Incident 218482)
(attached).
27
Refinery Rule RRA at 42 (finding additional cancer and acute health risks from “non-routine”
emissions or malfunction events the standards allow); but see Ethylene Production RTC at 20 (stating that
EPA’s health risk assessment includes elevated health risks from these emissions, though not identifying
them or showing how EPA has accounted for these additional risks).

17

 communities need facilities to have a strong incentive to avoid releases, not an advance free pass
to pollute. As just one series of powerful instances showing the need to avoid such exemptions,
a 2018 EIP report found that huge HAP releases from refineries and other chemical facilities
were reported to the Texas Commission on Environmental Quality during and after Hurricane
Harvey in 2017, finding a total of 8.3 million pounds of excess HAP released. 28 Many of these
occurred days after the Governor declared a state of emergency, as facilities continued operating
instead of preparing adequately for an orderly and safe shutdown. 29 EPA would allow such
releases to qualify for the Final Rule’s “force majeure event” regulatory exemption, even though
hurricanes are common occurrences and fall, foreseeably and like clockwork, across the nation
every year during hurricane season.
Hurricane Harvey and the high releases associated with this storm happened in 2017.
EPA is well aware of these events. 30 Yet, it does not appear that EPA even considered these
during this rulemaking. In a 2019 report resulting from an investigation, EPA’s Office of
Inspector General highlighted major flaws in EPA air quality data collection and emergency
response during and after Hurricane Harvey when chemical facilities were known to release
millions of pounds of HAPs and other harmful pollutants. 31 Thus far, there has been little or no
enforcement by EPA at any of these facilities known to have had such releases. It is not publicly
known how many facilities may have taken advantage of the “force majeure event” exemption,
as EPA has not released this information.
The circumstances surrounding COVID-19 provide another chilling example of why EPA
must reconsider and remove the “force majeure event” exemptions from the revised ethylene
production standards. According to the Federal Emergency Management Agency “[a]ll 50 states,
the District of Columbia, and 4 territories have been approved for major disaster declarations”
related to COVID-19.” 32 The COVID-19 health crisis amply demonstrates how such an
unlawful exemption could threaten to swallow the rule wholesale, by potentially allowing
polluting facilities to seek, and EPA to grant, unlimited free passes for PRD releases and visible
flare emissions during the pendency of a years-long pandemic.
Recent developments related to the COVID-19 pandemic reinforce the egregious and
unlawful nature of the “force majeure event” exemptions that EPA has finalized here. After the
President issued a national emergency declaration on March 13, 2020, and the COVID-19 virus
cases began to increase exponentially, the American Petroleum Institute asked the EPA for
28

EIP, Preparing for the Next Storm at 1 (Aug. 16, 2018), https://environmentalintegrity.org/wpcontent/uploads/2018/08/Hurricane-Harvey-Report-8.16.18-final.pdf; EIP, News Release: Preparing for
the Next Storm(Aug. 16, 2018), https://environmentalintegrity.org/reports/preparing-for-the-next-storm/.
29
Id. at 2.
30
See, e.g., Petitioners’ May 2018 Comments, 15-17 (May 25, 2018),
https://www.regulations.gov/document?D=EPA-HQ-OAR-2010-0682-0953.
31
EPA OIG, EPA Needs to Improve Its Emergency Planning to Better Address Air Quality Concerns
During Future Disasters, Report No. 20-P-0062 (Dec. 16, 2019),
https://www.epa.gov/sites/production/files/2019-12/documents/_epaoig_20191216-20-p-0062.pdf.
32
U.S. Federal Emergency Management Agency, Covid-19 Disaster Declarations (last updated July 27,
2020), https://www.fema.gov/disasters/coronavirus/disaster-declarations.

18

 extraordinarily broad enforcement and regulatory exemptions (such as from fenceline monitoring
for pollution releases, leak detection and repair requirements to identify and end pollution
releases, and other undefined “regulatory noncompliance”) due to the virus, described as “nonessential compliance discretion.” 33
Shortly thereafter, on March 26, 2020, EPA’s Assistant Administrator for Enforcement
(“OECA”) published a new “temporary” enforcement discretion policy regarding “implications”
of COVID-19 that, when originally issued, applied for an indefinite period of time. 34 In this
policy, EPA provided advance notice that it would not seek penalties for violations of
monitoring, reporting, or compliance certification requirements “where the EPA agrees that
COVID-19 was the cause of the noncompliance.” 35 The policy also states that EPA is open to
considering similar direct violations of administrative settlement agreements and consent decrees
as “force majeure,” and excusable as caused by COVID-19. 36 EPA states that the policy should
not “be read as a willingness to exercise enforcement discretion in the wake of [an accidental
hazardous chemical] release,” implicitly referring to § 7412(r). 37 This seems to signal, however,
that in nearly any other circumstances, including circumstances pertaining to compliance with
routine air pollution regulations, EPA was inviting requests from industry to not enforce or seek
penalties for violations using COVID-19 as an excuse. On April 2, 2020, EPA issued a letter to
members of Congress describing this memo as a temporary policy illustrating how EPA would
exercise its “case-by-case” enforcement discretion “after the pandemic is over.” 38 EPA later
updated its website to state that it was no longer in effect as of August 31, 2020.
EPA has occasionally issued narrowly tailored advance enforcement statements during
past emergencies, and this policy appears to be unusually broad, illustrating how sweepingly
industry could attempt to seek, and EPA could attempt to apply, the “force majeure event”
regulatory exemption in the final rule, because the rule imposes no decisive constraints on how
EPA may choose to interpret a “force majeure event,” which is defined expansively within the
regulations as a PRD or flare emissions release that the Administrator deems to “result from an
event beyond the owner or operator’s control, such as natural disasters; acts of war or terrorism;
loss of a utility external to the ethylene production unit . . . and fire or explosion originating at a
near or adjoining facility outside of the ethylene production unit that impacts the ethylene
33

Letter from Sr. Vice Pres. Frank J. Macchiarola, Am. Petrol. Inst., to Adm’r Wheeler on “Compliance
Discretion” (Mar. 23, 2020) (attached) (providing a list seven pages long of “detailed examples of issues
for which industry is seeking temporary relief through enforcement discretion, waivers or revised
compliance timeframes in response to the COVID-19 pandemic”).
34
EPA Enforcement Policy Memo from Susan Bodine, Ass’t Adm’r, OECA, to All Governmental and
Private Sector Partners, Re: COVID-19 Implications for EPA’s Enforcement and Compliance Assurance
Program (Mar. 26, 2020), https://www.epa.gov/sites/production/files/202003/documents/oecamemooncovid19implications.pdf (attached) (“OECA Policy on COVID-19”). This
memo expired on August 31, 2020 but EPA has not released information showing the full impact it had in
authorizing exemptions or allowing dangerous emissions.
35
Id. at 2-3.
36
Id. at 4.
37
Id. at 7.
38
See, e.g., Letter from Ass’t Adm’r, OECA, to Sen. Feinstein (Apr. 2, 2020) (attached) (“OECA Letter
to Feinstein”).

19

 production unit’s ability to operate.” 40 C.F.R. § 63.1103; 85 Fed. Reg. at 40,426. Certain states
have classified COVID-19 as a natural disaster in order to trigger executive powers associated
with a natural disaster declaration, and certain state courts have upheld such declarations. 39
EPA’s enforcement policy memorandum further makes clear that the types of challenges the
agency envisages as being associated with COVID-19, a presidentially declared nationwide
emergency, include facility staffing shortages that EPA misguidedly, and unlawfully, believes
warrant absolution from statutorily-mandated enforcement activity. 40
EPA’s enforcement policy memorandum, which provided guidance on how EPA may
choose to exercise its enforcement discretion, where industry makes a case-by-case showing that
“COVID-19 was the cause of non-compliance,” shows how harmful and wide-ranging an
advance regulatory exemption that authorized broad non-compliance could become. 41 The fact
that API and EPA are already pointing to the COVID-19 crisis as a “force majeure” event
illustrates how dangerous and unlawfully broad the advance regulatory exemption in this rule is.
A federal government plan responding to COVID-19 has suggested that the pandemic could
extend for as long as 18 months. 42 The harm from allowing industry to evade requirements for
such an extended period under the cloak of a “force majeure event” would be extraordinarily
severe. For communities where families are sheltering at home and facing increased exposure to
routine and non-routine air pollution, EPA’s suggestion that it would entertain requests from
industry to use COVID-19 as an excuse for violating health-protective regulations and exceeding
standards that protect people vulnerable to both air pollution and COVID-19 is unconscionable.
Moreover, the exemption in the regulation is not limited to releases related to natural
disasters or the examples provided within the regulatory provisions. Under the regulatory
language, industry may attempt to rely on the force majeure event exemption for any release that
it contends “result[s] from an event beyond the refinery owner or operator’s control,” with no
stopping point. 40 C.F.R. § 63.1103; 85 Fed. Reg. at 40,426. The Act does not allow this kind of
back-and-forth between regulated industries and EPA in defining what regulations apply, when.
In contrast to a negotiated enforcement agreement between parties that has a “force majeure”
clause curtailing the scope of the agency’s enforcement discretion, the ethylene production
standards under the Clean Air Act must be “continuous” and apply at all times to assure the level
of health protection required. Sierra Club, 551 F.3d at 1027-28. The statute is not a contract that
industry may negotiate.

39

Jerry Kimmitt & Jacob Esparza, Briefings: COVID-19 Classified as a Natural Disaster by
Pennsylvania Supreme Court; Could This Affect Coverage Litigation Arising From COVID-19?,
HOLMAN FENWICK WILLAN LLP (last visited Sep. 3, 2020), https://www.hfw.com/COVID-19-Classifiedas-a-Natural-Disaster-by-Pennsylvania-Supreme-Court-Could-this-Affect-Coverage-Litigation-Arisingfrom-COVID-19.
40
OECA Policy on COVID-19 at 2.
41
OECA Policy on COVID-19 at 3.
42
U.S. Dep’t of Health & Human Services, PanCAP Adapted U.S. Government COVID-19 Response
Plan at 4 (Mar. 13, 2020), https://int.nyt.com/data/documenthelper/6819-covid-19-responseplan/d367f758bec47cad361f/optimized/full.pdf#page=1.

20

 At the same time, the EPA enforcement policy memo also illustrates that creating an
advance regulatory exemption is neither rational nor necessary. In the event an unusual
circumstance comes up that makes it impossible to comply, EPA has well demonstrated that it
believes it has other options where it deems that facility compliance was truly impossible. It can
simply exercise enforcement discretion case-by-case, it can issue a temporary policy to this
effect in narrow circumstances, or it can use other legal authority granted by Congress to address
emergencies. EPA’s letter to Congress highlighted examples of past exercises of enforcement
discretion during and after hurricanes that provide ways EPA has chosen to address compliance
concerns it deemed valid—without the need for a broad, advance exemption to be promulgated
into a permanent regulation. 43 Where a facility operator can demonstrate that all available
advance preparation to prevent releases occurred but compliance was still impossible due to a
natural disaster, EPA enforcement discretion may well be warranted in limited circumstances,
and a court may well apply similar discretion in a private enforcement suit pursuant to § 7604.
However, EPA may not lawfully or rationally authorize non-compliance by regulation, before
any such planning, before there is any such result, and before the facility makes all available
attempts to prevent the violations occur, and in such broad terms that a facility loses any
incentive to even attempt to comply.
Allowing regular malfunction exemptions and even broader exemptions for some of the
most dangerous releases during times EPA attempts to deem so-called “force majeure events,”
such as natural disasters like Hurricane Harvey and Hurricane Laura as discussed elsewhere in
these comments, means there is no pollution control during these particularly dangerous times
even though facilities can and should take additional precautions and have a strong incentive to
avoid pollution during these periods that are particularly dangerous to human health. Thus, this
objection is of “central relevance.” EPA must grant reconsideration to correct its inaccurate
characterization of the ethylene production regulations, to remove the unlawful malfunction
exemptions, and to set limits that satisfy the Act’s requirement for emission standards to apply at
all times.
II.

EPA MUST GRANT RECONSIDERATION ON ITS NEW RATIONALE FOR
THE PRESSURE RELIEF DEVICE PROVISIONS UNDER 7412(d)(2)-(3).
A. Inability to Raise the Objection in Comments.

In the 2019 proposed rule, EPA contended that the PRD work practice standards satisfy
§ 7412(h) and meet the floor and beyond-the-floor requirements of § 7412(d)(2)-(3) as § 7412(h)
also requires. Petitioners demonstrated that the exemptions are not consistent with the best
performers’ emission limitations as shown by the Bay Area Air Quality Management District
(“BAAQMD”) and South Coast Air Quality Management District (“SCAQMD”) Rules.
Petitioners’ 2019 Comments at 127-129.
Perhaps aware that the rationale provided for the exemptions could not withstand
scrutiny, EPA now advances new rationales and new information to justify the provisions
allowing them, as discussed below. In particular, EPA attempts to ignore measured data
43

OECA Letter to Feinstein at 2 (describing 41 examples of enforcement discretion, one fuel waiver, and
21 force majeure letters during Hurricanes Katrina and Rita, as well as other similar examples).

21

 available from PRD releases, demonstrating EPA cannot satisfy § 7412(h)’s test to set a work
practice standard instead of a numerical limit of zero. EPA also attempts, without evidence or
merit, to justify the PRD provisions as satisfying § 7412(d)(2)-(3) while newly admitting that the
final standard strips away core restrictions contained in the SCAQMD, BAAQMD and Chemical
Accident Prevention Provisions rules (84 FR 54303, October 9, 2019)– rules that the agency
contends represent the standards applying to facilities achieving the PRD emissions “floor” of
zero. EPA suggests – for the first time – that the final requirements are “comparable” to these
more restrictive rules, directly contradicting the rationale that the agency offered at proposal. 85
Fed. Reg. at 40,410. 44
Because EPA has provided these new rationales solely in the Final Rule, it was
“impracticable” within the meaning of Clean Air Act § 307(d)(7)(B) to object to these rationales
and EPA’s reliance on them during the public comment period. 42 U.S.C. § 7607(d)(7)(B). In
addition, the grounds for the below objection to the new rationales arose after the close of the
comment period under § 307(d)(7)(B), when EPA unveiled its new rationales with the Final
Rule.
B. Objection: EPA Has Failed to Demonstrate That The Work Practice Standards
For Pressure Relief Devices Meet § 7412(h) and § 7412(d)(2)-(3) In Reliance on
the California Local Air District Rules.
As a first new rationale, EPA contends in response to Petitioners’ 2019 Comments that,
because it is setting a work practice standard, it can ignore the numerical emissions data that is
available from the SCAQMD and elsewhere demonstrating that the best-performing sources in
this source category are emitting zero HAP emissions from PRDs. See 85 Fed. Reg. at 40,410.
Section 112(h), however, directs that work practice requirements must be consistent with
§ 112(d)—i.e. require the “maximum” degree of reduction in emissions that is “achievable” and,
at a minimum, require reductions to the level “achieved” by the best performing sources. The
data that EPA ignores demonstrate that sources have achieved, and can achieve, far greater
reductions than EPA’s work practice requirements require. Accordingly, EPA’s work practice
requirements are unlawful and the agency’s decision to ignore these data is unlawful and
arbitrary.
EPA cannot satisfy § 7412(h) where such data have been measured and it is “feasible” to
set a numerical limit. That the data show that it is “feasible” for the best performing ethylene
production facilities to reduce emissions to zero demonstrates that EPA is required by law to set
a numerical limit for PRDs rather than a work practice standard. The PRD work practice
standard therefore violates § 7412(h). EPA gives no lawful or rational justification to explain
how or why, when those PRD release data exist, it has not set a numerical emissions limit for
PRDs. See 42 U.S.C. § 7607(d)(9).
EPA now concedes that the emissions reductions achieved by the best performing sources
is zero, 85 Fed. Reg. at 40,409, but arbitrarily fails to show that “measurement technology
limitations” described at proposal prevent facilities from documenting and reporting that a PRD
44

EPA has since rolled back the 2017 Chemical Disaster Rule amendments that it had pointed to as
relevant evidence of what sources should be doing here.

22

 release has occurred. To the contrary – at proposal, EPA specifically stated that the agency had
“identified . . . monitoring systems capable of alerting an owner or operator of when a PRD
release occurs.” 84 Fed. Reg. at 54,302-303. Given the agency’s own prior concession, EPA’s
new claim that it can ignore available data cannot possibly excuse it from making a credible
showing that setting a numeric emissions limit of zero for PRDs is infeasible.
Further, even if EPA were justified in setting a work practice standard for PRDs in lieu of
a numeric emissions limit, which it is not, EPA has failed to meet the Act’s requirements for
setting appropriately protective work practice standards. In the Final Rule, EPA recognizes that it
has not required, as part of the finalized work practice standard, the most stringent measures
applicable in the BAAQMD, SCAQMD and Chemical Accident Prevention Provisions rules that
EPA describes as the emissions “floor.” EPA newly admits that the final standards for PRDs do
not “exactly mirror” the BAAQMD, SCAQMD and Chemical Accident Prevention Provisions
rules. 85 Fed. Reg. at 40,410. However, EPA newly argues that the requirements are
“comparable,” id., even though the BAAQMD rule more stringently requires that the operator
must control all PRDs that discharge for a second time in a 5-year period and that the SCAQMD
rules “include a similar provision, but offer as an alternative payment of a fee of $350,000 for
each PRD that is not controlled.” Id. at 40,409. Unlike the finalized work practice standards for
PRDs, the “SCAQMD rules also require control of any PRD that has a single large release of
greater than 2,000 pounds per day (lbs/day).” Id. Moreover, the local rules more stringently limit
emissions than EPA’s finalized work practice standards for PRDs in other respects. See id.
EPA’s shifting explanations for the PRD work practice standards from proposal to final
rule highlight the unlawful, pretextual and arbitrary nature of those standards. 42 U.S.C. § 7412
(d)(2)-(3), (h): § 7607(d)(9). In the final rule, EPA directly contradicts the rationale that EPA
offered at proposal that the agency “expect[s] that the best performing ethylene production
facilities have implemented a program for PRDs that vent to atmosphere that consists of using at
least three prevention measures and performing root cause analysis and corrective action.” 84
Fed. Reg. at 54,303. Instead, the agency now claims that its work practice standard, while
significantly less stringent than the BAAQMD and SCAQMD rules that it describes as the floor
assuring maximum emissions reductions, is offset by certain differing requirements in the final
rule. 85 Fed. Reg. at 40,410. EPA argues, for example, that it is requiring a root cause analysis
for PRD releases of any size, whereas the SCAQMD does not require such an analysis for
releases less than 500 lbs/day, rendering the work practice standard “comparable” to the
otherwise stronger SCAQMD rule. Id. But the agency points to no actual improvements made to
offset the weaker requirements than the BAAQMD rule that EPA finalizes here, stipulating only
that “[w]e also adopted the three prevention measures requirements in the BAAQMD rule with
limited modifications.” Id.
EPA arbitrarily provides no evidence-based analysis in support of its assertion that the
finalized requirements for PRDs are “comparable” to the BAAQMD, SCAQMD and Chemical
Accident Prevention Provisions rules, merely reciting conclusory assumptions and suggesting
that the supposedly offsetting modifications are “reasonable.” See id.; 42 U.S.C. § 7607(d)(9).
Whether the modifications are “reasonable” is beside the point. The statute directs EPA to set
emission standards that assure the average emission limitation “achieved” by the relevant best
performers. 42 U.S.C. § 7412 (d)(2)-(3), (h). EPA has failed to do this, by failing to set rules that

23

 are at least as stringent as the rules to which it points that the best performers are meeting.
Instead, EPA unlawfully and irrationally cherry-picks the parts of each set of rules that it wishes
to include during malfunctions. Petitioners support the enhancements EPA has included from
the BAAQMD and SCAQMD rules. But EPA gives no lawful or rational justification for
allowing a higher number of releases prior to considering the owner or operator to be in violation
of the work practice standard than even the SCAQMD Rule 1173 allows, or for failing to
continuously prohibit all large releases of greater than 2,000 pounds per day as the SCAQMD
rules also require. See id. EPA does not demonstrate how allowing a greater number of
uncontrolled releases than permitted by the state rules upon which the MACT floor is supposedly
based satisfies, or could satisfy, § 7412(d)(2)-(3) (even if EPA could allow any uncontrolled
releases). EPA has also failed to match the rules that it asserts represent the MACT floor by
refusing to require more preventative measures (BAAQMD requires 5 instead of 3) and failing to
adopt the other additional requirements in the more stringent BAAQMD rules. EPA gives no
lawful or rational justification for refusing to set more stringent limitations on PRD releases in
order to satisfy § 7412(d)(2)-(3).
EPA’s new rationale that it can ignore the numerical emission data available because it is
arbitrarily labeling a selection among some parts of state work practice standards the “floor” fails
to satisfy the Act or fundamental requirements of reasoned decisionmaking. 42 U.S.C.
§ 7607(d)(9). EPA appears to have simply chosen what it wishes to require from the SCAQMD,
BAAQMD and Chemical Accident Prevention Provision rules and labeled that weaker
combination of requirements the “floor.” EPA has unlawfully and arbitrarily labeled its preferred
choice of parts of the SCAQMD and BAAQMD rules as the “floor” and refused to recognize that
it must further restrict emissions from PRDs to zero as the Act directs.
EPA further suggests that “significant additional emissions” would be generated by
installing additional flares or a cogeneration plant to reduce releases creating a “net
environmental disbenefit.” 85 Fed. Reg. at 40,405. Yet EPA fails to demonstrate that any
additional emissions would be produced by augmenting flare capacity, particularly in light of the
SCAQMD data showing that the best-performing sources have very few or no PRD emissions.
Further, EPA has failed to consider ways to reduce or prevent such disbenefits, such as requiring
additional preventative measures applicable to PRDs (the BAAQMD requires five preventative
measures instead of the three finalized here), flare gas recovery systems, or through additional
methods to control and reduce emissions from flares themselves (which the agency has
unlawfully and arbitrarily relaxed here instead of ensuring apply at all times). EPA has therefore
failed to justify refusing to satisfy § 7412(d)(2)-(3) by strengthening standards applying to PRDs
and flares.
EPA must reconsider the ethylene production rule to further strengthen the PRD
requirements and ensure that it sets standards that satisfy § 7412(d)(2)-(3)) as discussed in these
comments.

24

 C. This Objection Is Of “Central Relevance” to the Rule Because It Demonstrates
the Inadequate Stringency of the Pressure Relief Device Standards.
Because this objection goes directly to EPA’s new rationale for failing to set a numerical
PRD emissions limit of zero and failing to set a far more stringent work practice standard
limiting PRD releases as required by the Clean Air Act, it is of “central relevance” to the
outcome of the rule. 42 U.S.C. § 7607(d)(7)(B). Petitioners object to EPA’s unlawful failure to
set a numeric emissions limit of zero in light of EPA’s new concession that the best performing
PRD sources emit zero emissions. Petitioners also object to the agency’s new, unlawful attempt
to justify its work practice standards’ substantial divergences from the local rules that EPA
represents as the MACT floor by arguing, with evidence or merit, that minor enhancements or
modifications offset the stripping away of core requirements contained within the local rules.
Petitioners’ objections call for the substantial strengthening of the weak standards regulating
PRD emissions under the final rule, rendering the objections of “central relevance” and meriting
a new round of public comment proceedings. Id.
In addition, EPA’s failure to restrict PRD emissions routinely, and to prohibit them
(including during malfunctions) is particularly harmful and unlawful given that the available data
further show the need to restrict emissions connected to force majeure events. The OIG and EIP
reports (cited above) summarizing the high amounts of HAPs and co-emitted pollutants from
refineries and chemical plants during and after Hurricane Harvey, Laura, and similar incidents
illustrate the strong need for EPA to remove the force majeure event exemption. These data
show that, instead, EPA must require additional preventative and corrective action measures for
HAP releases connected to hurricanes in hurricane-prone areas like the Gulf or other natural
disasters such as earthquakes in earthquake-prone areas.
EPA must therefore convene proceedings to reconsider the PRD work practice standards,
including the exemptions from such standards.
III.

EPA HAS FAILED TO DEMONSTRATE THAT THE WORK PRACTICE
STANDARDS FOR FLARES MEET § 7412(d)(2)-(3).

A. Inability To Raise the Objection in Comments
In the Final Rule, EPA finalized a new provision that removed applicability of the novisible emissions requirements once or twice every three years, and any time there is a “force
majeure event.” At proposal, Petitioners demonstrated that even if the unlawful exemptions were
permitted to stand, the new emergency flare provisions are inconsistent with the best performers’
emission limitations on flares. Petitioners’ 2019 Comments at 105-107, 132-134. For example,
Petitioners commented that there is no rational connection between EPA’s finding that the “best
performers would have no more than one event every 7 years, or a probability of 14.3 percent of
having an event in any given year,” 84 Fed. Reg. at 54,296, and EPA’s now finalized proposal to
set a standard allowing up to two uncontrolled emission releases every three years for certain
causes and an unlimited number of uncontrolled emissions caused by “force majeure” events.
Petitioners’ 2019 Comments at 132-133.

25

 In the Final Rule, EPA advances new rationales to attempt to justify the unlawful
exemptions it has promulgated for emergency flaring. Shifting course dramatically, EPA now
suggests that, although EPA has relied on § 7412(d)(2)-(3) to set the flaring operational and
monitoring requirements, the agency need not meet the statute’s requirement to set standards
achieving the maximum achievable degree of emissions reduction for flares. EPA newly
contends that maximally protective standards are not required for flares because “[it] is therefore
not a specific emission source within the EMACT standards.” 85 Fed. Reg. at 40,404. EPA
further states that the agency “do[es] not set [emission] limits for thermal oxidizers, baghouses,
or other control devices that we desire to remain operational during malfunction events to limit
pollutant emissions to the extent practicable.” Id. EPA additionally newly argues that it did not
set the flare standards to ensure maximally achievable control of flares, but to assure 98 percent
destruction or reduction efficiency at the emitting equipment routed to the flares. Id.
Because EPA only provided this new rationale in the Final Rule, it was “impracticable”
to object to EPA’s rationale during the public comment period. 42 U.S.C. § 7607(d)(7)(B). In
addition, the grounds for our below-discussed objections arose after the close of the comment
period under § 307(d)(7)(B), when EPA unveiled its new rationale with the Final Rule.
B. Objection: EPA’s New Rationale That It May Set Standards for Flares That Do Not
Satisfy § 7412(d)(2)-(3) Is Unlawful and Arbitrary.
EPA is incorrect that it need not satisfy § 7412(d)(2)-(3) for flares because there are
underlying emission standards that apply to equipment routed to flares, and flares are being used
as a control device.
Petitioners support EPA’s action in the Final Rule to set important new flare operational
and monitoring requirements to assure compliance with the underlying standards. Generally,
those standards require the equipment to meet 98% destruction efficiency for HAPs. EPA found
significant problems achieving that level of HAP destruction and the improvements should
reduce HAP emissions from the underlying equipment routed to the flare. 84 Fed. Reg. at
54,294. The Final Rule aims to address those problems. Petitioners also support EPA’s
recognition in the Final Rule that sources can and must meet the new maximum flare tip velocity
requirement at all times. 85 Fed. Reg. at 40,404; RTC at 104.
However, it is indisputable that flares are not merely control devices—they are also
emission points. In the process of destruction of certain HAPs, flares create new HAPs—the

26

 products of incomplete combustion. 45 Thus, flares emit new HAPs into the air. The Act requires
EPA to set a limit on all emitted HAPs. 42 U.S.C. § 7412(d)(1)-(3); Nat’l Lime Ass’n v. EPA,
233 F.3d at 641. The D.C. Circuit reaffirmed this principle in holding that it is “necessary”
under § 112(d)(6) for EPA to correct the problem of uncontrolled HAP emissions in this
rulemaking. LEAN, 955 F.3d at 1096 (“We read the statutory text to require EPA during its
section 112(d)(6) review to establish any missing limits.”). Therefore, EPA must set a floor and
consider setting beyond the floor standards for the uncontrolled HAP emissions flares release,
just as Petitioners explained in their 2019 Comments. By refusing to set limits, as the Act
requires, on the total HAPs released from flares and ignoring the increased emissions that occur
during smoking flares, EPA is violating the Act. Indeed, although EPA failed to propose (or
finalize) a sufficiently protective limit for flares, the agency at proposal recognized that flares
must abide by requirements establishing a MACT floor, stating that “for purposes of the MACTallowable risk analysis, we are required to evaluate whether it is necessary to tighten the existing
MACT standard and subsequent level of performance a flare is expected to achieve.” 84 Fed.
Reg. at 54,285-86.
EPA’s contention in the Final Rule that the 98-percent combustion efficiency is the only
limit on flares necessary to satisfy the CAA is contrary to law. 42 U.S.C. § 7412(d)(2)-(3). The
Act requires EPA to set limits on each hazardous air pollutant that a source category emits, and it
requires these limits to reduce the pollutant or pollutants to which they apply to the maximum
achievable degree. 42 U.S.C. § 7412(d)(2). At a minimum, the Act requires EPA’s limits to
reflect the actual emission levels achieved for the pollutant or pollutants to which they apply by
the sources that are best performing with respect to such pollutant or pollutants. Id.;
§ 7412(d)(3). A standard that merely reduces the total volume of hazardous air pollutants by a
percentage does not—and cannot possibly—satisfy these requirements. For example a source
achieving 98-percent reduction of its total emissions of hazardous air pollutants might be
achieving no reduction at all in emissions of pollutants such as dioxins or polycyclic organic
matter that are emitted at low volumes, even though significant reductions in such pollutants are
“achievable” and in fact being “achieved” by some sources. 42 U.S.C. § 7412(d)(3).
Further, even if EPA is correct that 98 percent of the HAP emissions sent to the flare are
typically destroyed, in a hydraulic flaring event much greater amounts of gases are sent to the
45

80 Fed. Reg. at 75,215 (“smoke in the flare exhaust is an indication of incomplete combustion.”); EPA
Flare Study (in the docket); OEHHA 2019 App. E (listing examples of pollutants released during “nonroutine” refinery events, including flaring); see also Weyant, C.L., Shepson, P.B., Subramanian, R.,
Cambaliza, M.O.L., Heimburger, A., McCabe, D., Baum, E., Stirm, B.H., Bond, T.C.: Black carbon
emissions from associated natural gas flaring. Environ. Sci. Technol. (2016).
https://doi.org/10.1021/acs.est.5b04712; Wielgosiński, G. Pollutant Formation in Combustion Processes.
Technical University of Lodz, Faculty of Process and Environmental Engineering Poland; Fahad M. AlFadhli, Yosuke Kimura, Elena C. McDonald-Buller, David T. Allen, Impact of Flare Destruction
Efficiency and Products of Incomplete Combustion on Ozone Formation in Houston, Texas, Ind. Eng.
Chem. Res. 2012, 51, 39, 12663-12673; Kindzierski W.B. (2000) Importance of human environmental
exposure to hazardous air pollutions from gas flares, Environ. Rev. 8, 1, 41-62 (“emission of a variety of
compounds such as volatile organic compounds (VOC's), polycyclic aromatic hydrocarbon (PAH's) as
well as soot occur when the flaring process ensues with incomplete combustion”).

27

 flare. The final work practice standard imposes no underlying limit based on the “average
emission limitation achieved” by the relevant best-performing flares or sources. Importantly, the
more a flare combusts, the more new HAP pollution it creates from the process of incomplete
combustion. 46 Thus, by permitting exemptions for visible flaring, EPA is unlawfully allowing
much greater total amounts of HAPs to be released into the air.
Accordingly, EPA must perform the floor and beyond-the-floor analysis required to ensure
that refinery flares cannot emit more than the Act allows, no matter how much HAP stream a
facility sends to the flare. By failing to impose limits on flares that are fully compliant with
§ 7412(d)(2)-(3), EPA is allowing virtually unlimited levels of HAPs to be released through these
emissions points, in violation of the Act.
EPA’s new rationale is arbitrary because it is, by its own admission, acting pursuant to
§ 7412(d)(2)-(3) in strengthening the flare requirements but asserts, contrarily, that it need not
fully meet the stringency test that applies to standards created under this provision. See 42 U.S.C.
§ 7607(d)(9). EPA’s new rationale shows it has not even attempted to satisfy § 7412(d)(2)-(3) in
creating its smoking flare work practice standards, even though it has appropriately relied on that
provision in promulgating the flare improvement provisions, 47 and has recognized it must
strengthen the flare standards pursuant to § 7412(d) elsewhere in the rulemaking record. For
example, EPA states that it is proposing “revisions to the operating and monitoring requirements
for flares . . . pursuant to CAA section 112(d)(2) and (3).” 85 Fed. Reg. at 40,309. In the final
rule, EPA also appropriately requires compliance with the maximum flare tip velocity operating
limit at all times, including when flares are smoking. 85 Fed. Reg. at 40,404; RTC at 104 (“in the
final rule, you must comply with the maximum flare tip velocity operating limit at all times and
there is no work practice standard for when the flare vent gas flow rate exceeds the smokeless
capacity of the flare and the tip velocity exceeds the maximum flare tip velocity operating limit.”)
(emphasis added). In explaining why EPA strengthened that requirement from proposed rule to
Final Rule, the agency stated that it agreed with comments that its “proposed determination of the
frequency of these events at the best performing sources is not supported.” 85 Fed. Reg. at 40,404.
In doing so, EPA recognized that § 7412(d)(2)-(3) requires that the agency determine a MACT
floor for flares based, in part, on emissions reductions achieved by the best performing sources,
and impose requirements that ensure emissions are reduced to that floor. However, EPA fails to
apply that logic to strengthen flare requirements across the board, as required by law.
EPA’s new argument is also irrational because the fact that a flare is being used as a
control device for other equipment, and is assisting that equipment in reducing its emissions,
does not excuse the agency from setting limits preventing flares from releasing unlimited
amounts of HAPs. 42 U.S.C. § 7607(d)(9). Even if EPA’s argument held water, EPA has failed
to justify how it can avoid setting any limits on the new HAPs a flare creates. The other control
devices that EPA cites as examples—thermal oxidizers and baghouses—are purely HAP control
devices. They do not create new HAPs in the process of destroying others. A flare’s use as a
control device does not justify EPA’s failure to satisfy § 7412(d)(2)-(3) by setting emission
limits to assure the “maximum achievable degree of emission reduction” from flares that it is at
least as stringent as the “average emission limitation achieved” by the best-performing flares.
46
47

See, e.g., sources cited in prior footnote.
See, e.g., Proposed Rule, 84 Fed. Reg. at 54,294; Final Rule, 85 Fed. Reg. at 40,389-90.

28

 (Notably, EPA must strengthen limits on flares for the independent reason that doing so would
help to assure the “ample margin of safety to protect public health” that is needed to satisfy
§ 7412(f)(2))).
Moreover, EPA is also violating the Act by creating a separate subcategory of flares –
those operating above the smokeless capacity – and allowing unlimited pollution from those
flares. Section 7412(d)(1) and § 7412(h) do not allow EPA to categorize sources or emission
points by whether or no=t they are having a malfunction. EPA has some authority to
“distinguish among classes, types, and sizes of sources,” in setting a standard reflecting the
maximum degree of emission reduction. 42 U.S.C. § 7412(d)(1). However, EPA may not
categorize sources by whether or not they are controlled. 48 The flares are all the same type of
source. The only distinction between the flare that EPA treats disparately is that some are
smoking as part of a malfunction event because they do not have a flare gas recovery system or
other methods in place to prevent such smoking even when a hydraulic flaring event occurs.
Thus, EPA has unlawfully set a weaker standard for smoking flares than for all other flares.
EPA must set standards that satisfy § 7412(d)(2)-(3) for all flares, rather than subcategorizing
them in a way that puts all of the best-performing sources in one group and the worstperforming, smoking flares in the other group, thereby unlawfully weakening the requirements
that apply to smoking flares.
C. The Objection Is of Central Relevance to the Outcome of the Rule Because It is the
Core Factor in the Stringency of the Flare Requirements.
Because this objection goes directly to the illegality and arbitrariness of EPA’s rationale
for not setting stronger flare standards as required by the Clean Air Act, it is of “central
relevance” to the outcome of the rule. 42 U.S.C. § 7607(d)(7)(B); CCAN, 952 F.3d at 322
(finding objection that raised concerns that “go to the very legality” of a standard centrally
relevant). Petitioners’ objection would require EPA to evaluate and set a MACT floor for flares
that restricts flare emissions to the extent achievable by the best performing sources, an act that
EPA expressly disavows performing in this final rule.
Data from Hurricane Harvey and other serious releases in recent years, including
examples cited in the first section of this petition from Hurricane Laura, additionally show that it
is critical for EPA to restrict emissions from flares connected to force majeure events rather than
providing an advance free pass to pollute. The OIG and EIP reports summarizing the high
amounts of HAPs and co-emitted pollutants from refineries and chemical plants during and after
48

See 42 U.S.C. § 7412(d)(1) (allowing EPA only to “distinguish among classes, types, and sizes” of
sources); see also Davis Cty. Solid Waste Mgmt. v. EPA, 101 F.3d 1395, 1399 (D.C. Cir. 1996), amended
on reh’g, 108 F.3d 1454 (D.C. Cir. 1997) (“[T]he MACT floor will obviously be lower if the category
includes more units with advanced pollution control devices than if the category contains fewer units with
such devices”); Sierra Club v. EPA, 895 F.3d at 15 (“But once the EPA identifies a source in a category it
must set the MACT floor based on the ‘best’ performing sources.”); id. (“The EPA has the authority to
‘distinguish among classes, types, and sizes’ of emissions sources and set separate MACT floors for
each.”).

29

 Hurricane Harvey illustrate the strong need for EPA to remove the force majeure event
exemption, and instead require additional preventative and corrective action measures for HAP
releases connected to hurricanes or other natural disasters such as earthquakes in earthquakeprone areas (like California and Utah). 49 Thus far, EPA appears not to have followed through
with any meaningful enforcement at any ethylene production or similar facilities since 2016.
Therefore, EPA must convene proceedings to reconsider the new rationale for the
smoking flare exemptions and flare work practice standards, and the standards that rely on that
new rationale.
IV.

EPA MUST GRANT RECONSIDERATION ON ITS NEW RATIONALES
FOR FINDING CANCER RISKS ACCEPTABLE, AND ON ITS RISK
DETERMINATION BASED ON THESE RATIONALES.

A. Inability To Raise the Objections in Comments.
In the Final Rule, EPA incorrectly determined that the residual health risks its standards
allow ethylene production facilities to cause nearby communities, including the highest risks to
the “individual most exposed” are “acceptable.” 85 Fed. Reg. at 40,309. EPA’s failure to sum, or
attempt to sum, cancer risk from inhalation and multipathway exposures violates the Act’s
directive to adequately assess the health risk to determine whether such risk is acceptable
pursuant to 42 U.S.C. § 7412(f)(2) and is arbitrary under 42 U.S.C. § 7607(d)(9). In the Final
Rule, EPA found the MIR, or maximum individual cancer risk, from inhalation is 100-in-1
million, from regular inhalation of ethylene production facility HAP emissions. RRA at App. 10
tbl. 2b. The record actually shows that the cancer risk is 104-in-1 million and EPA has not
explained why it has said the risk is lower. Id. (for NEI ID # 22089110013831201: MIR based
on allowable emissions of 1.04 x 10-4 or 104-in-1 million).
EPA’s reported MIR of 100-in-1 million is the benchmark that EPA deems presumptively
unacceptable for cancer risk from toxic industrial air pollution. 50 But, that was not the only
cancer risk EPA found from ethylene production facility pollution. EPA also found that these
sources cause significant additional cancer risk through the ingestion pathway of exposure, for
people exposed to locally grown vegetables or fish—as high as 30-in-1 million, according to the
Final Rule preamble. 85 Fed. Reg. at 40,392. The accompanying risk assessment documentation,
however, shows multipathway risk as high as 51-in-1 million at the Exxon Baytown Olefins
plant. Final RRA at App. 10 tbl. 5.a (Multipathway Cancer Screen Values - for NEI ID
48201110000463169). To reach its “acceptable” risk conclusion under § 7412(f)(2), EPA
ignored the fact that the cancer risk is actually above the level it deems presumptively
acceptable: 100-in-1 million. Adding together the MIR of 104-in-1 million (as EPA’s record
shows) with 30-in-1 million, would be a lifetime cancer risk of 134-in-1 million or 151-in-1
million total, based on the highest multipathway risk found. At a minimum, it would be 130-in-1
million -- far higher than EPA’s benchmark. As shown in the attached chart, if EPA added the

49

See reports cited, supra notes 19-21.
That benchmark is far too high and should be reduced, as Petitioners’ 2019 Comments discussed at 8893.
50

30

 highest MIR and multipathway risks for individual sources, the total cancer risk would also be
above 100-in-1 million for at least two sources.
In their reconsideration petition and 2019 Comments, Petitioners called for EPA to
follow its own guidelines and sum inhalation and multipathway cancer risks, to recognize the
cancer risk is higher than the agency’s own benchmark for presumptive acceptability,
necessitating the strengthening of the standards. Petitioners’ 2019 Comments at 69-72.
In the Final Rule, EPA expressly refused to sum the cancer risk, but implicitly
recognizing that it lacked justification for doing so in view of its own guidelines, EPA tried to
come up with a new rationale. To try to avoid finding risk to be above the presumptively
acceptable level, EPA newly claims that it may choose not to add together the cancer risks that
the ethylene production pollution causes, in reliance on new, conclusory assumptions that run
counter to the Act, EPA’s own guidelines, and the best available science. See Response to
Comments at 49.
Because EPA has suddenly provided these rationales in the Final Rule, it was
“impracticable” within the meaning of Clean Air Act § 307(d)(7)(B) to object to EPA’s
rationales during the public comment period. 42 U.S.C. § 7607(d)(7)(B). In addition, the
grounds for our below-discussed objections arose after the close of the comment period under
§ 307(d)(7)(B), when EPA unveiled its new rationale with the Final Rule, and so reconsideration
is required.
B. Objection: EPA’s New Rationales for Refusing to Add Cancer Risk Are Unlawful
and Unsupported by Science or the Record and Fail to Justify EPA’s § 7412(f)(2)
Determination.
EPA’s own risk assessment policy and guidelines recognize that carcinogenic risk adds
up with each increment of exposure. Petitioners’ 2019 Comments at 61-72 (citing scientific
sources; EPA guidelines). There is no safe level of exposure to carcinogens—they have no
threshold. Therefore, EPA appropriately added cancer risk from inhalation of multiple pollutants
together here, and also recognized the need to add cancer risk from inhalation of pollutants from
multiple sources in the ethylene production category together to determine the maximum
inhalation risk. See Final RRA at 13, 33 & App. 2. 51
Yet, EPA refused to add each type of cancer risk (from different exposure pathways) to
assess total cancer risk in this risk assessment. Doing so would increase the cancer risk above
EPA’s presumptive level of acceptability, to as high as 110-in-1 million, based on allowable
emissions under the final rule. 42 U.S.C. § 7412(f)(2); Petitioners’ 2019 Comments at 83-84;
Final RRA at 32 tbl. 3.2-3; id. App. 10 tbl. 5.a. EPA’s multipathway risk assessment “resulted in
a maximum Tier 2 cancer screening value of 30 for arsenic.” 85 Fed. Reg. at 40,392. EPA’s
Appendix 10 to the Final Residual Risk Assessment shows that the highest multipathway cancer
risk is actually 51 for the Baytown olefins plant. Final RRA App. 10 tbl. 5.a. However, EPA did
51

“We typically combine the impacts of all facilities within the same source category and assess chronic
exposure and risk for all census blocks with at least one resident (i.e., locations where people may
reasonably be assumed to reside rather than receptor points at the fenceline of a facility).”

31

 not add the cancer risk from routine inhalation (100) with the cancer risk from ingestion or
“multipathway” exposure to arsenic (30) in order to determine whether risk is unacceptable, and
what reductions are needed to achieve an acceptable level, and to provide the requisite “ample
margin of safety” to protect public health. 42 U.S.C. § 7412(f)(2); see also RTC at 49.
EPA for the first time attempts to justify refusing to come up with a cancer risk total by
contending that its screening value for multipathway risk overestimates risk and that it is
inappropriate to sum multipathway risk and inhalation risk from a scientific standpoint. EPA
states that the risk from allowable emissions under the standard (“MACT-allowable”) is based on
“conservative,” “health-protective assumptions” that assume that the “exposed individual
exhibits ingestion behavior that would lead to a high total multipathway exposure.” RTC at 49.
Nevertheless, the fact that EPA’s assessment includes certain health-protective assumptions and
components does not allow EPA to avoid adding the risks it has found—to assess cancer risk as a
whole from ethylene production sources’ pollution.
EPA’s express refusal to sum inhalation and ingestion cancer risk violates the Act and is
arbitrary. 42 U.S.C. § 7412(f)(2); id. § 7607(d)(9). The objective of § 7412(f)(2) and the health
risk rulemaking is to require EPA to take an advance, precautionary approach to assess risk and
protect public health from toxic air pollution before harm occurs. Health-protective assumptions
(also known as scientific “defaults”) and uncertainties are always present in risk assessment
where the goal is to attempt to predict and prevent future harm. Congress understood that there
would likely be scientific uncertainties that would make it difficult or impossible to assess the
exact amount of additional cancer risk a person faced from a given industrial source category.
So that scientific gaps or risk assessment uncertainties do not inadvertently lead to dangerous
gaps in health protection, Congress directed EPA to push forward: not just to take the most basic,
lowest level of action possible, but to assure standards provide “an ample margin of safety to
protect public health,” 42 U.S.C. § 7412(f)(2). Further, the Act does not just require this as a
general matter, but it directs EPA to protect the “individual most exposed”—based on the
maximum potential risk to human health. Id. In the context of public health protection and to
meet the statutory test that focuses on the “most exposed” and requires an “ample margin of
safety,” the statute favors a conservative approach in the face of any uncertainties.
Here, EPA cannot rationally assess risk based on record facts and the science, and then
turn around and ignore its own health risk assessment results as too “health protective” to be
useful. EPA recognized the need to ask key questions and account for the risks that its
assessment found. The agency cannot justify now ignoring the answers, just because they point
to the need for more stringent regulations. And, overall, EPA guidelines direct that “the primary
goal of EPA actions is protection of human health; accordingly, as an Agency policy, risk
assessment procedures, including default options that are used in the absence of scientific data to
the contrary, should be health protective.” 52
Therefore, EPA appropriately used certain health-protective assumptions or defaults in its
risk assessment, and they do not justify sidelining that risk assessment at the regulatory step in
52

EPA, Guidelines for Carcinogen Risk Assessment (Mar. 2005),
https://www.epa.gov/sites/production/files/2013-09/documents/cancer_guidelines_final_3-25-05.pdf.

32

 this rulemaking. Its new rationale that the numbers it found are conservative or “health
protective” and may be lower does not justify refusing to add together carcinogenic risk or to
recognize that this risk is above EPA’s benchmark of unacceptability. Due to gaps that remain in
the agency’s risk assessment approach, EPA likely undercounts, not overestimates, risk, showing
how important the assumptions it does use are. 53
The 100-in-1-million benchmark for maximum individual risk was established by EPA
itself, was accepted as reasonable (at EPA’s urging) by the D.C. Circuit at least based on the
record at the time of that decision, and is part of the baseline risk approach that the court found
Congress intended to affirm as a method from which EPA could not backslide when it enacted
the Clean Air Amendments of 1990. Even if EPA could somehow find health risk acceptable
where cancer risk was greater than 100-in-1 million, it would have to provide a reasoned
explanation based on the record and full picture of health risks communities face, and
demonstrate how this could be consistent with the Act. EPA has not done that here. Instead,
EPA now takes the position—without any explanation or justification whatsoever—that because
its calculations include certain assumptions that EPA considers conservative, the final risk
estimates it has generated are overestimates.
As a second new rationale in support of its cancer risk determination, EPA attempts to
justify its refusal to add cancer risks together by cursorily contending that “[i]t is highly unlikely
that the person with the highest inhalation exposure is the same hypothetical fisher or
farmer/gardener with the highest multipathway exposure.” RTC at 49. EPA cites no evidence
whatsoever for this conclusion and fails to articulate any rational connection between its
unsupported assertion and the facts at hand. To the contrary, fishing and gardening sites are
located near ethylene production facilities, 54 creating a substantial likelihood that EPA’s
assumption is false. As such, EPA’s contention does not meet the fundamental principles of
53

As discussed in Petitioners’ 2019 comment, EPA’s risk assessment approach is not the most health
protective or conservative—it actually contains serious gaps in assessing emissions and risk, which means
it likely underpredicts the cancer and other health risks that communities face from refinery pollution.
Petitioners have called for EPA to follow recommendations of the National Academy of Sciences
(“NAS”) and Science Advisory Board (“SAB”) that would further strengthen its risk assessment to
account for additional risks it currently ignores or predicts to be lower than the best available scientific
information would indicate.
54
Houston-area ethylene facilities emit pollution in close proximity to fishing and gardening sites. For
example, the Baytown Olefins Plant—with among the highest multipathway cancer risks of any ethylene
facility, RRA App. 10 tbl. 5.a—is less than two miles from a nature center with four fishing piers, and
less than three miles from a community garden that produces vegetables for a food pantry. Baytown
Nature Center, About the Nature Center, CITY OF BAYTOWN (last visited Sep. 3, 2020),
https://www.baytown.org/city-hall/departments/parks-recreation/baytown-nature-center/about-the-naturecenter; Texas Annual Conference of the United Methodist Church, ‘Lord’s Acre’ Provides Healthy
Harvest for Baytown (July 14, 2016), https://www.txcumc.org/newsdetail/lord-s-acre-provides-healthyharvest-for-baytown-5345063. Ethylene facilities are also located near gardening sites in Port Arthur,
Texas, where a recently installed community garden is roughly three miles from both the Chevron Phillips
ethylene facility and the Flint Hills Resources ethylene facility. Raegan Gibson, New Park to Open in
Port Arthur for Families to Plant Fruit, Vegetables, Enjoy Playground, KBMT (Apr. 15, 2019),
https://www.12newsnow.com/article/news/local/power-city/new-park-to-open-in-port-arthur-for-familiesto-plant-fruit-vegetables-enjoy-playground/502-38e7cf79-c1b4-4f9f-bdb5-0f7d53c6c129.

33

 reasoned decisionmaking or the best available science as required by the Act. 42 U.S.C. §
7607(d)(9); see also Allied Local & Reg’l Mfrs. Caucus, 215 F.3d at 61, 68.
EPA cannot rely on its unsupported assertion as a basis to refuse to even attempt to
determine the health risks faced by the person most exposed to cancer risk from both inhalation
and ingestion, as § 7412(f)(2) requires. The Act directs EPA to protect “the individual most
exposed” in setting health-protective emissions standards. 42 U.S.C. § 7412(f)(2). That term
necessarily requires EPA to account for the highest possible cancer risk to the nearby community
from any route of exposure—and to assume that an “individual” is experiencing that risk.
Refusing to consider or account for the maximum potential exposure for an individual from
aggregate cancer risk, when it is well established that community members both breathe air and
take in carcinogenic pollution through ingestion, violates the Act and is arbitrary. 42 U.S.C.
§ 7412(f)(2); id. § 7607(d)(9). Indeed, as EPA’s own source-specific risk data show, all ethylene
production sources cause both inhalation and multipathway risks. Final RRA App. 10 tbl. 5a.
Thus, even if EPA believes it is not “appropriate” to add the maximum inhalation and
multipathway cancer risks together, at least it should add these cancer risks that it found for each
source together. Doing so for at least some of those sources would lead to cancer risk above 100in-1 million. For example, EPA states that the Norco facility has an MIR of 100-in-1 million
based on allowable emissions (which is actually 104-in-1 million according to EPA’s chart) –
adding an additional 4-in-1 million to that would tip this source over EPA’s unacceptability
benchmark for cancer risk. Id. The same is true at the Baytown Olefins Plant, where adding
fishing and farming multipathway risks raises the cancer risk above 107-in-1 million.55 In
wholly failing to attempt to add these risks, Final RRA App. 10 tbl. 5.a, EPA has failed to
account for the full cancer risk from multipathway exposure in setting standards for this
category; EPA’s decision not to set stronger emission limits is unlawful as well as arbitrary and
capricious.
Further, § 7412(f) standards must be “based on the most current scientific knowledge and
on sound scientific judgment.” 56 Failure to aggregate cancer risk from multiple routes of
exposure flouts scientific consensus. As the National Academies of Sciences recognized,
“[b]ecause exposure to a specific chemical is rarely confined to a single route (although one
route might dominate), the total exposure must be calculated by summing air (inhalation),
dermal, and dietary (food and water) intakes. For example, pollutants that begin as ‘air
pollutants’ can generate substantial exposures through other media if they can move from air to
water, soil, or vegetation.” 57 To protect communities and the most-exposed individual from the
highest risk allowed under the ethylene production standards, EPA must add multipathway and
inhalation risk, just as OEHHA does. As OEHHA explained in its 2015 Risk Assessment
Guidelines: “If multiple substances are emitted, the substance-specific cancer risks for all
55

See Chart - Ethylene Production Cancer Risk Addition_09-03-2020, based on EPA’s RRA (attached).
Benzene Rule, 54 Fed. Reg. 38,044, 38,062-63 (Sept. 14, 1989).
57
National Research Council, Science and Judgment in Risk Assessment at 595 (1994),
https://www.nap.edu/catalog/2125/science-and-judgment-in-risk-assessment (emphasis added). EPA
cited and relied upon this report in creating the Residual Risk Report to Congress in 1999. EPA, Residual
Risk Report to Congress at 130-31 (Mar. 1999), https://www.epa.gov/sites/production/files/201308/documents/risk_rep.pdf (“Residual Risk Report”).
56

34

 exposure pathways are summed to give the (total) multipathway cancer risk at the receptor
location.” 58
EPA cannot ignore that there are people who both breathe in and ingest pollution from
ethylene production facilities at the highest assessed levels. 59 That is precisely why EPA
recognized the need to assess both inhalation and multipathway risk in this rulemaking. Evidence
shows that people who live and breathe air near ethylene production facilities have backyards,
attend schools, and frequent community gardens where they grow food that they consume. 60
Community urban gardens near ethylene production facilities also feed local residents who visit
local farmers’ markets and receive food. The ethylene production facility with the highest
inhalation cancer risk is an example of this; a garden at an elementary school merely 1.1 miles
from Shell Chemical Norco has supplied produce to a local farmer’s market. 61
Ethylene production facilities are located near small and large waterbodies, including the
Mississippi River and Sabine Lake, Capitol Lake and others near Baton Rouge, Lake Charles (in
Louisiana), the Houston Ship Channel, and more. Many of these have fishing advisories in place
because community members rely on these waterbodies for sustenance and fishing still occurs
notwithstanding the advisories due to the need for local communities to rely on this fish, 62
where community residents both breathe air and fish for sustenance because they need to, even

58

Cal. EPA OEHHA, Risk Assessment Guidance Manual at 5, 8-13 (Feb. 2015) (attached to 2014
Comments) (“Cancer risks from all carcinogens addressed in the HRA [health risk assessment] are
added.”); (“Cancer risks from different substances are treated additively in risk assessment generally, and
in the Hot Spots Program in part because many carcinogens act through the common mechanism of DNA
damage.”).
59
EPA’s 1999 Residual Risk Report shows visually how inhalation and other pathways of exposure can
cause cancer and other risk in the same community. Residual Risk Report at 36-37 & ex. 7 (1999),
https://www.epa.gov/sites/production/files/2013-08/documents/risk_rep.pdf.
60
See, e.g., Norco 4-H Works in Garden, L’Observateur (Oct. 17, 2014),
https://www.lobservateur.com/2014/10/17/norco-4-h-works-in-garden/#puzzle,1641,1598828547598
(describing sale at local farmer’s market of produce grown at school garden); ‘Lord’s Acre’ Provides
Healthy Harvest for Baytown, Texas Annual Conference of the United Methodist Church (Aug. 14,
2016), https://www.txcumc.org/newsdetail/lord-s-acre-provides-healthy-harvest-for-baytown-5345063
(describing garden growing produce for community food pantry).
61
Final RRA App. 10 tbl. 2.b; Staff Reports, Norco 4-H Works in Garden, L’OBSERVATEUR (Oct. 17,
2014), https://www.lobservateur.com/2014/10/17/norco-4-h-works-ingarden/#puzzle,1641,1598828547598.
62
For example, Devil’s Swamp & Devil’s Lake in North Baton Rouge (a few miles north of Exxon Mobil
Chemical, which has an ethylene production facility), has fish consumption advisories. See, e.g.,
Louisiana Dep’t of Envtl. Quality, Fishing Consumption and Swimming Advisories (last updated Feb. 7,
2019), https://deq.louisiana.gov/index.cfm?md=pagebuilder&tmp=home&pid=fishing-consumption-andswimming-advisories. This article discusses the fact that people still fish here due to the need for fish. See
Kevin Litten, Devil’s Swamp Lake fish too polluted to eat, DHH warns, NOLA.com (last updated Jul. 18,
2019), https://www.nola.com/news/politics/article_fdb9c84c-3c6c-5a43-b06c-6c60507b542e.html.

35

 where it may be unsafe to do so. 63 Fishing activities may be located where garden sites also
occur, as in Norco, where the Army Corps of Engineers advertises “fishing, crawfishing,
crabbing, and shrimping” at the Bonnet Carré spillway—within three miles of Shell’s ethylene
facilities and Norco Elementary School. 64 Research shows low-income, and communities of
color are more likely to fish, and grow food for sustenance. 65 In view of the disproportionate
exposure for people of color and the known impacts to these communities from ethylene
production facilities, it is particularly problematic and irrational for EPA to assume that the
person most exposed to inhalation-based cancer risk is not also highly, and even the “most,”
exposed to ingestion-based cancer risk.
Even assuming arguendo that EPA could justify not adding the highest level of ingestion
risk to the highest level of inhalation risk, the agency cannot lawfully or rationally avoid adding
at least some amount of ingestion to the inhalation to attempt to better account for the real-world
risk of the person most exposed to ethylene production pollution through multiple pathways.
EPA has not, and cannot, credibly claim that the person with the highest inhalation exposure
from this source category is unlikely to experience any level of multipathway exposure. Thus,
EPA is not excused from its responsibility to account for and sum inhalation risk and
multipathway risk in calculating the maximum individual cancer risk. Even if EPA disagrees that
the agency’s multipathway assessment cancer risk value of 30-in-1 million for arsenic should be
added to the agency’s calculated inhalation cancer risk of 100-in-1 million to determine the
maximum individual cancer risk, the agency must still account for multipathway ingestion risk
by estimating the total cancer risk faced by the most exposed individual in order to determine an
accurate MIR.
The accompanying chart illustrates that adding the inhalation and multipathway cancer
values that EPA found for at least two sources—the Shell Chemical Norco and Baytown Olefins
plants—shows cancer risk significantly above 100-in-1 million when adding exposure from
farming and fishing risks; applying the same method, additional sources present cancer risks
from 80-90-in-1 million. 66 Thus, EPA cannot contend that no person is exposed to a combined
cancer risk above 100-in-1 million – because the information the agency itself has provided
illustrates that is not true. Further, because the cancer risk values themselves are underestimates,
as discussed in Petitioners’ Comments (Residual Risk section), it is likely that more than only the
top two in the chart are causing unacceptable cancer risk.

63

The Texas Department of State Health Services has endorsed an assumption, based on EPA guidance,
that subsistence fishing “likely occurs” in Sabine Lake--next to Port Arthur--regardless of whether the
frequency of such fishing has been documented precisely. E.g., Texas Dep’t of State Health Services,
Characterization of Potential Adverse Health Effects Associated with Consuming Fish from Sabine Lake
(2011), https://dshs.texas.gov/WorkArea/linkit.aspx?LinkIdentifier=id&ItemID=8589961765.
64
Bonnet Carré Spillway Fishing, U.S. Army Corps of Engineers (last visited Sep. 3, 2020),
https://www.mvn.usace.army.mil/Portals/56/docs/Recreation/BCS/Brochures/BCS%20Fishing%20Broch
ure.pdf.
65
Nat’l Envtl Justice Advisory Council, Fish Consumption and Environmental Justice at 14-15, 26-29
(Nov. 2002), https://www.epa.gov/sites/production/files/2015-02/documents/fish-consumpreport_1102.pdf (“NEJAC 2002”).
66
See supra note 51, Cancer Risk Chart based on the RRA (attached).

36

 Further, EPA has also stated that it evaluated combined cancer risk from multiple
sources. See Final RRA at 13, 33 & App. 2. However, EPA has not provided the specific
combined cancer risk values at any community receptor locations, other than the source-specific
ones included in the chart attached. EPA should provide those to support its determination that
cancer risk is no higher than 100-in-1 million, if indeed that is the case. As some sources,
individually, are causing such high cancer risk, and those are located near other sources, it is
surprising that the combined multiple-source cancer risk values are also no higher than the
highest MIR EPA has provided for any single source (i.e., 100-in-1 million). If in fact combined
cancer risk from inhalation alone is higher than 100-in-1 million, EPA must recognize that
cancer risk is unacceptable. Because EPA has not provided this information within the final
Residual Risk Assessment, or anywhere else in the record to the best of Petitioners’ discernment,
EPA has also violated the public participation requirements of 42 U.S.C. Sec. 7607(d) (requiring
publication in the docket of all materials on which EPA relied for the final rule).
In order to dismiss excess ingestion cancer risk above its presumptive benchmark of 100in-1 million, as EPA has done here, EPA would have to provide a lawful and reasoned
justification for its policy change with regards to the level of cancer risk that the agency has
historically deemed presumptively unacceptable. See, e.g., FCC v. Fox Television Stations, Inc.,
556 U.S. 502, 515 (2009); Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S.
29, 42 (1983). EPA has not done so and thus has failed to justify ignoring the significant
additional cancer risk that its own assessment shows communities are experiencing. 42 U.S.C. §
7607(d)(9). Because the cancer risk is above 100-in-1 million as the data show, EPA has
violated § 7412(f)(2) by refusing to reduce this risk by strengthening the emission standards.
EPA may not lawfully or rationally assume that “the most exposed individual” whom the Act
directs the agency to protect is someone who either breathes air or ingests pollution, but never
both. Id.
EPA’s failure to account for any amount of multipathway risk in determining whether
risk is unacceptable and determining what is required to provide the mandated “ample margin of
safety to protect public health,” in violation of § 7412(f)(2) and contrary to § 7607(d)(9) appears
all the more egregious in light of the fact that EPA’s cancer risk assessment, including its
multipathway risk assessment, already represents a significant undercount of the true cancer risks
communities are facing. Petitioners’ 2019 Comments at 30-72. For example, EPA failed to
quantify the risk from dermal exposure, from exposure in utero, of infants in breastmilk, or of
children playing at home or on a playground near an ethylene production facility and ingesting
soil or dust containing the pollution—though it cannot deny that these exposures happen
regularly near ethylene production facilities around the United States and the best available
science demonstrates that they contribute to cancer risk. Petitioners’ 2019 Comments at 61-63.
Indeed, EPA’s multipathway assessment itself undercounts cancer risk for the ethylene
production source category. EPA acknowledged but did not quantify the additional cancer risk
from ingestion of fish from the Houston Ship Channel and other large waterbodies because it has
not yet created parameters to put numbers on this additional risk. RRA at 19. EPA cannot deny
that risk from dispersion into those waterbodies and ingestion of fish from those waters exists.
Yet, the agency has purposefully ignored such risk in assessing multipathway cancer risk from
ingestion. It is therefore indisputable that EPA’s own multipathway assessment – which EPA

37

 declines to use or account for in the calculation of the maximum individual cancer risk –
undercounts the risk from fish ingestion. EPA also failed to acknowledge increased exposure for
low-income, tribal and Native American communities who are more exposed to PBTs and the
resulting cancer risk due to increased consumption through fishing, hunting, and farming. 67 EPA
did not account for this increased exposure for any community members in the Risk Assessment.
Rather, EPA used the same default ingestion value for all community members. The National
Environmental Justice Advisory Council (“NEJAC”) has called EPA’s use of default values that
“are less protective of higher-consuming and subsistence subpopulations … [g]iven that these
subpopulations are in the main comprised of particular communities of color, low-income
communities, tribes, or other indigenous peoples, … deeply troubling.” NEJAC 2002 at iv. EPA
also appears to have left known toxic metals that persist in the environment out of the
multipathway risk assessment completely, such as chromium compounds. See RAA at 5, 35.
Thus, although metals such as chromium (including hexavalent chromium) are known to
contribute to cancer risk from ingestion, EPA’s assessment assumes the risk produced is zero.
Further, EPA chose to discount its finding that the facility-wide risks at facilities that
contain ethylene production sources and other kinds of sources, like refineries and other
chemical plants, are extremely high – well above EPA’s 100-in-1 million unacceptability
benchmark. 68 While EPA did not factor the facility-wide risk into its §7412(f)(2) determination
for ethylene production, the evidence showing that people are cumulatively exposed to
unacceptable levels of pollution from both ethylene production units and collocated units in other
source categories demonstrates that EPA must, at least, account for the full, combined cancer
risk from ethylene production plants and regulate them effectively to assure an “ample margin of
safety to protect public health.” For example, EPA notes that the highest cancer risk source from
inhalation, Shell Norco, is collocated with a refinery. EPA’s risk assessment for refineries found
that the MIR is 100-in-1 million and for the Shell Norco facility – also regulated under the
Ethylene Production Rule – the MIR is 24-in-1 million (based on allowable emissions). 69 Thus,
combining that with the EP risk this would create a facility risk from these two types of sources
as high as 134-in-1 million (110 from EP and another 24 from the refinery)—there is no question
that people exposed to the Shell Norco’s ethylene production emissions are experiencing an
unacceptable level of cumulative cancer risk from facilities regulated both by the Final Rule and
the Refineries Rule.
Similarly, EPA admits that people are simultaneously exposed to Ethylene Production
units and other types of co-located chemical plants that emit ethylene oxide. 70 As the facilitywide risk at these units is as high as 2,000-in-1 million, 20 times EPA’s benchmark of
unacceptability, EPA cannot ignore the evidence that the “individual most exposed” to ethylene
production emissions is exposed, cumulatively, to unacceptable total amounts of HAPs from the
combined impact of ethylene production and collocated sources’ emissions. The record shows
that people exposed to the Norco facility, and to the collocated EP-MON or EP-HON facilities
67

NEJAC 2002 at iv (“many American Indian and Alaska Native (AI/AN) communities are particularly
prone to environmental harm due to their dependence on subsistence fishing, hunting, and gathering”).
68
Refinery RRA (-0800) App. 10 tbl. 2.b.
69
Refinery RRA (-0800) App. 10 tbl. 2.b (NEI33031 – Shell Norco/formerly Motiva).
70
Final Rule, 85 Fed. Reg at 40,392; Ethylene Production RRA at 42.

38

 are experiencing unacceptable risk from the facility-wide emissions that include emissions from
ethylene production and collocated regulated major sources. 71 To satisfy the Act’s requirement
to provide an “ample margin of safety to protect public health,” EPA cannot justify refusing to
add inhalation and multipathway risks together, or setting stronger standards that account for this
facility-wide risk, to satisfy § 7412(f)(2). 72
EPA may not lawfully or rationally ignore the excess cancer risk from inhalation of
ethylene production pollution, including emissions released during the emission spikes EPA
allows as part of its malfunction exemptions, and the excess cancer risk EPA has found from
ingestion. EPA must admit that its own record shows the cancer risk MIR is above 100-in-1
million. EPA must therefore grant reconsideration to set standards to reduce cancer risk and the
other health threats ethylene production facilities cause, including acute risks from benzene and
chronic and acute risks from toluene, as discussed later in this petition. EPA cannot satisfy §
7412(f)(2) by ignoring the full extent of cancer risk, or any other kind of health risk, posed by
ethylene production facilities.
C. These Objections to EPA’s New Rationales on Cancer Risk Are Of Central
Relevance Because They Are A Core Basis for EPA’s Decision Not To Set Stronger
Standards Pursuant to § 7412(f)(2).
Because Petitioners’ objections pertaining to excess cancer risk fundamentally challenge
EPA’s rationale for finding health risk acceptable, and consequently, not setting stronger
emission standards pursuant to § 7412(f)(2), the objections are of “central relevance” to the
outcome of the rule. 42 U.S.C. § 7607(d)(7)(B). The risk assessment is the core determination
dictating whether standards are sufficiently protective of human health under this provision. If
EPA finds risk unacceptable, it must first reduce risk to acceptable levels, without consideration
of cost. EPA must thereafter provide an “ample margin of safety to protect public health.” 42
U.S.C. § 7412(f)(2). Petitioners’ objection, if deemed well-founded, would compel EPA to set
stronger standards pursuant to § 7412(f)(2).
The objections are also “of central relevance” because of the high cancer risk
communities proximate to ethylene production facilities are facing and because EPA has
undercounted this risk, as summarized above. If EPA were to sum the cancer risks it has found—
which already undercount the real-world health risks—it would find the cancer risk alone is
above the benchmark of what EPA has recognized to be presumptively unacceptable. Again,
EPA would be required to act pursuant to § 7412(f)(2) to set stronger standards to eliminate this
unacceptable cancer risk and provide an “ample margin of safety to protect public health” from
ethylene production facilities’ pollution. EPA must therefore grant reconsideration.

71

See, e.g., EPA, Miscellaneous Organic Chemical Manufacturing Residual Risk Assessment & App. 10
(Apr. 2020), https://www.regulations.gov/document?D=EPA-HQ-OAR-2018-0746-0189 (listing some
sources collocated with ethylene production sources, such as the Channelview, Baytown, and Sasol
plants) (attached).
72
See also Final RRA at 6, 40 (same cancer incidence).

39

 V.

EPA MUST GRANT RECONSIDERATION ON THE NEW ANALYSIS AND
NEW RATIONALE IT HAS USED TO ASSESS ACUTE RISK, AND TO
GROUND ITS § 7412(F)(2) DETERMINATION.

A. Inability to Raise the Objection in Comments.
Benzene is a dangerous pollutant that is a neurotoxicant and can cause acute respiratory
harm as well as cancer. Ethylene production sources (all 31 assessed) emit approximately 498
tons per year of benzene, according to EPA. Final RRA at 36 tbl. 3.1-1.
In the proposed and final 2020 Rule for Ethylene Production, EPA assessed acute risk
from benzene using an acute reference exposure level of 1.3 mg/m3. Final RRA at 36 (appearing
to refer to the pre-2014 Cal. EPA OEHHA “REL” of 1.3 mg/m3) (-0070). In comments,
Petitioners called for EPA to use the most current, then-recently finalized 2014 Cal. EPA
OEHHA reference exposure level for benzene of 27 µg/m3 (or 0.027 mg/m3). Comments at 3940. Applying that REL would have shown acute risk to be an order of magnitude greater, from
just this one pollutant.
EPA did not do so, however, and EPA’s final action is internally contradictory. In the
Final Residual Risk Assessment, EPA stated that it was considering the new OEHHA benzene
value. 73 By contrast, in the Response to Comments, EPA stated that it would not use the
OEHHA REL. RTC at 22-23. EPA’s Response to Comments makes it sound like EPA is using
an emergency-only value such as an AEGL or ERPG—yet the risk assessment does not show
this. Instead, the final RRA still refers to the old REL.
EPA’s own scientific guidelines, as affirmed by the Science Advisory Board, direct EPA
generally to apply the agency’s own IRIS toxicology values, then ATSDR, and then California
EPA OEHHA values as the priority health reference or dose-response values. 74 As EPA
explained in the proposed and final Residual Risk Assessment, “The prioritization process was
aimed at incorporating into our assessments the best available science with respect to doseresponse information.” Final RRA at 24. 75 EPA has prioritized use of the Cal. EPA OEHHA
values, as the only dose-response values EPA applies from any state. This is because, as EPA has
explained: “The process for developing these assessments is similar to that used by EPA to
develop IRIS values and incorporates significant external scientific peer review.” Id. at 25-26. It
appears that in the proposed rule, EPA had simply not updated the REL yet to recognize the
change in 2014, as it should have to follow these guidelines, but at least it was applying the prior
OEHHA REL.
73

RRA at 40 note f.
SAB, Review of EPA’s draft entitled, “Risk and Technology Review (RTR) Risk Assessment
Methodologies: For Review by the EPA’s Science Advisory Board with Case Studies – MACT I Petroleum
Refining Sources and Portland Cement Manufacturing,” EPA-SAB-10-007 at ii, 5-6 (May 7, 2010)
(attached) (“SAB May 2010”) (“The Panel found EPA’s approach to selecting dose-response chronic
toxicity values to be generally sound.”).
75
EPA further explains that it has prioritized use of the Cal. EPA reference dose values because of “(1)
conceptual consistency with EPA risk assessment guidelines and (2) level of peer review received.” RRA
at 24; see also id. at 25-26.
74

40

 In the Final 2020 Rule, however, without public notice-and-comment, EPA did
something completely incomprehensible. EPA refused to use the current OEHHA REL. EPA
responded to comments not to justify that action but to justify use of some other unidentified
value. EPA has given no information as to what acute health risk value EPA actually used. The
RRA table shows only the former REL (1.3). RRA at 36 tbl. 3.1-1 (1.3 shown in the benzene row
for the column “Health Benchmark Values for Acute Noncancer (mg/m3)”). If in fact EPA used
an AEGL or an ERPG, developed by an industry association rather than a value created by EPA
or any of the authoritative, prioritized governmental scientific agencies, that would follow an
approach that is not supported by the best available science or the SAB and was the opposite of
what Petitioners called for. EPA has provided inconsistent and inadequate information as to
what value it actually applied to assess acute risk.
EPA provides two new rationales for not using the OEHHA REL:
(1) “The ATSDR acute inhalation MRL of 0.029 mg/m3 represents the daily human
exposure to benzene that is likely to be without appreciable risk of adverse noncancer health effects for up to 14 days of exposure. The ATSDR acute (1- to 14-day)
MRL for benzene is very similar to the CalEPA’s acute (1-hour) REL for benzene.
This indicates that exposure to this concentration for up to 2 weeks is not expected to
be associated with adverse effects.”
(2) “the EPA disagrees with some aspects of the methodology used to derive the RELs
for 1,3-butadiene, benzene, and nickel, and the derivation methodology may partly
underlie the apparent discrepancy between the 1- to 14-day MRL and the 1-hour REL
for benzene.” EPA also states that this is “[b]ased on examination of these acute
RELs, including the methodology used in the derivation of the values (e.g., including
the identification of effect on which to base the value and the application of UFs).”
RTC at 22-23.
EPA does not provide the value it actually used to assess acute risk from benzene in the
RRA or the RTC, however. If indeed EPA applied the weaker emergency-only factor, then EPA
would dramatically underestimate the acute risk from ethylene production sources, by two orders
of magnitude. EPA also does not provide any document showing the “examination” of the acute
benzene REL explaining what issue it had with the methodology used by OEHHA.
EPA did not provide this rationale or calculation in the record for the Final Rule. EPA
also did not provide notice that it would apply an emergency factor, instead of the acute REL, to
assess risk, and reach a different acute risk HQ than using the REL. Therefore, it was
“impracticable” to object to use of the new rationale and an AEGL or ERPG value as EPA’s
basis to weaken the Residual Risk Assessment for ethylene production and to refuse to set
§ 7412(f)(2) standards. Further, the grounds for the below-discussed objections arose after the

41

 close of the comment period under § 307(d)(7)(B), when EPA unveiled its new rationale with the
2020 Final Rule.
B. Objection: EPA Unlawfully and Arbitrarily Changed the Dose-Response Value
Used to Assess Acute Health Risk from Benzene and This Led to an Unlawful and
Arbitrary Risk Assessment and § 7412(f)(2) Determination.
In the 2020 Rule, EPA’s new rationale for finding low acute risk and determining that
health risks are acceptable is that it has used a different factor to assess acute risk than it did in
the 2019 proposed rule: an unknown emergency-based AEGL or ERPG factor. EPA states that it
has done this in reliance on an “examination” of the REL. RRA at 22. Yet, the only information
Petitioners could find on EPA’s website is the Dose-Response Table EPA uses for these rules,
which includes the 2014 OEHHA REL value for benzene: 0.027 mg/m3 – not the old value
(1.3). 76 There is no indication in that Table that EPA does not believe that value reflects the best
available science or that EPA will not use that in risk assessments.
EPA gives basically no explanation or information of the “examination” on which EPA
relies. EPA states cryptically that it “disagrees with some aspects of the methodology used to
derive the RELs for 1,3-butadiene, benzene, and nickel, and the derivation methodology may
partly underlie the apparent discrepancy between the 1- to 14-day MRL and the 1-hour REL for
benzene.”
EPA states that the ATSDR acute inhalation MRL (0.029 mg/m3) represents the daily
human exposure to benzene that is likely to be without appreciable risk of adverse non-cancer
health effects for up to 14 days of exposure. Because of this, EPA states that it has not used the
acute REL for benzene (0.027 mg/m3) because EPA implies that the similarity between these
values “indicates that exposure to this concentration for up to 2 weeks is not expected to be
associated with adverse effects.” RRA at 22.
These appear to be consistent; the ATSDR MRL is quite close to the OEHHA value.
Thus, reading them consistently would suggest that the acute risk may occur immediately and
continue for the two weeks the ATSDR value covers. If anything, the MRL supports rather than
undermines the 2014 REL. If it undermines the OEHHA value, EPA has not shown how.
Alternatively, if anything the difference suggests that the ATSDR value should be more
stringent, recognizing greater harm over a 2-week period of exposure. EPA fails to explain why
the comparison of these values would suggest a problem with the 2014 OEHHA value as
opposed to the ATSDR value, which was finalized in 2007, and is 7 years older and more
outdated than the latest OEHHA value. 77 If EPA deems the ATSDR value to be more protective
or more accurate, EPA also fails to explain why that is—and if so, why EPA is not using that
instead of the unidentified AEGL or ERPG values which would be far weaker, and are not
appropriate factors for risk assessment from regulation of routine air pollution.

76

EPA, Table 2. Acute Dose-Response Values for Screening Risk Assessments (Jun. 18, 2018),
https://www.epa.gov/sites/production/files/2014-05/documents/table2.pdf (attached).
77
Agency for Toxic Substances & Disease Registry, Toxicological Profile for Benzene (Aug. 2007),
https://www.atsdr.cdc.gov/toxprofiles/tp3.pdf (attached).

42

 Further, EPA’s terse parenthetical attacking the OEHHA “methodology” is a conclusory
statement that does not provide support for EPA’s risk assessment and EPA’s refusal to use the
2014 OEHHA value. EPA says, “e.g., identification of effect on which to base the value and the
application of UFs),” but gives no information on which effect, which UFs, or the application
thereof by OEHHA are questioned. EPA has given no rational basis to question the new 2014
REL or refuse to use it here. It has provided no scientific information or any evidence at all for
ignoring the more protective value created by OEHHA scientists using a thorough review of all
scientific evidence, and peer-review process. 78
The 2014 OEHHA REL is well supported and reflects the best available science. 79 It is
the peer-reviewed factor that meets EPA’s own scientific integrity principles, and that is why
EPA’s own Risk Assessment prioritizes use of the OEHHA values. Final RRA at 24-26.
Yet, for benzene, EPA uses a value that is less protective than the 2014 REL. The factor
EPA used is unknown – but if it is the ERPG-1 (160 mg/m3), then further illustrating how
extremely problematic this is, the Occupational Safety and Health Administration’s permissive
exposure level for workers is lower than the value EPA used here. OSHA’s 8-hour PEL is 1
ppm, and its acute or short-term PEL is 5 ppm as averaged over any 15 minute period. 80 Thus,
the value EPA is using to consider how much protection it will consider for community
residents—who, unlike workers, have no protective equipment or other occupational safeguards,
who live near the facility (and do not just work for part of the day), and who may well be more
vulnerable to the health risks of pollution than the average worker—is 30 times less protective
than the value OSHA uses to address health and safety threats for workers.
The evidence shows EPA’s proposal significantly undercounts the acute health hazards of
the regulated facilities’ emissions, and thus the risk to “the individual most exposed” to ethylene
production sources’ emissions. It is both unlawful and arbitrary, as it is out of step with EPA’s
longstanding practice, an unsupported change with that practice, and a failure to apply the best
available science. In addition to underestimating acute non-cancer health hazards, EPA is
willingly and knowingly exposing communities of color and low-income communities to
increased and unnecessary harm from short-term benzene exposure. Because, as discussed in the
next section, EPA is also undercounting the acute risk from toluene, its simultaneous refusal to
apply the best available science to account for benzene risk means that EPA’s risk assessment
substantially underestimates the health risk to “the individual most exposed,” requiring
reconsideration.
By contrast with the REL, the emergency-focused value, ERPG-1, is not an appropriate
value to assess human health risk from regular and repeat exposure over time to ethylene
production sources’ emissions under § 7412(f)(2). This value was created by a nongovernmental entity, the American Industrial Hygiene Association (AIHA), for use in
78

Cal. EPA OEHHA, 2014 REL for Benzene (July 2014) (docketed with Petitioners’ 2014 Comments),
https://oehha.ca.gov/media/downloads/crnr/appendixd1final.pdf.
79
See id.
80
29 C.F.R. § 1910.1028(c)(1)-(2),
https://www.osha.gov/pls/oshaweb/owadisp.show_document?p_id=10042&p_table=STANDARDS.

43

 responding to chemical hazard incidents like massive leaks, fires, and malfunctions, not to assess
and decide how much air pollution to allow communities to breathe regularly. 81 EPA has
described the ERPG-1 not as a threshold where no acute harm occurs but rather: “ERPG 1 is the
maximum airborne concentration below which it is believed nearly all individuals could be
exposed up to one hour without experiencing other than mild transient adverse health effects or
perceiving a clearly defined objectionable odor.” 82
EPA has admitted that the Emergency Response Planning Guidelines (ERPG) values
were created for emergency exposure scenarios. Levels defined for “once-in-a-lifetime, shortterm exposures” and “emergency planning” for “single exposures” to chemical releases or
accidents are not appropriate tools to use to measure the acceptability of acute risks over a
lifetime from one or more potential exposures due to an industrial source’s emissions. 83 The
ERPG-1 is designed solely for emergency exposure response (such as when to send first
responders into a scene with an active chemical hazard and when that is too dangerous), and
cannot be considered health-protective for community members facing potentially repeat
exposures over a lifetime to acute risks from emission spikes. An emergency value like this
cannot provide assurance for or in any way justify EPA’s acceptability determination. EPA may
not use these values in lieu of health-protective reference values such as the REL, and has no
rational or scientific basis to assume that risks below these levels are safe. 84
EPA distorts the SAB’s recommendations in the Response to Comments. EPA states that
“in response to a key recommendation from the Science Advisory Board’s peer review of the
EPA’s RTR risk assessment methodologies, we now examine a wider range of available acute
health metrics than we do for our chronic risk assessments.” RTC at 31. That citation is
deceptive, however.
The SAB has approved use of the RELs but not the EPRGs without reduction “by a factor
of 3.” As the SAB has explained:
The incorporation of the available California Reference Exposure
Levels (RELs) for the assessment of acute effects is a conservative
and acceptable approach to characterize acute risks . . . . The Panel
has some concern with the use of the Acute Exposure Guidelines
Limits (AEGLs) and Emergency Response Planning Guidelines
(ERPGs) These limits were developed for accidental release
emergency planning and are not appropriate for residual risk
81

Nat’l Institute for Occupational Safety and Health, Benzene: Systemic Agent, CENTER FOR DISEASE
CONTROL AND PREVENTION (last updated May 12, 2011),
https://www.cdc.gov/niosh/ershdb/emergencyresponsecard_29750032.html; EPA.
82
EPA, Benzene Hazard Summary (last updated Jan. 2012),
https://www.epa.gov/sites/production/files/2016-09/documents/benzene.pdf.
83
Proposed Rule, 84 Fed. Reg. at 54,288.
84
See, e.g., SAB May 2010 at 6 (available at
https://yosemite.epa.gov/sab/sabproduct.nsf/4AB3966E263D943A8525771F00668381/$File/EPA-SAB10-007-unsigned.pdf) (attached to 2019 Comments).

44

 assessments without modification because, as described in the
AEGL and ERPG documentation, adverse effects may occur at these
levels. Some of the AEGLs and ERPGs listed in Table 2-5 are higher
than values used to protect healthy workers from acute effects in
occupational settings. The Panel recommends considering reducing
the AEGL1/EPRG-1 emergency guideline values by a factor of 3,
when the value is based on a LOAEL rather than a NOAEL. 85
Thus, the SAB has acknowledged, using the ERPG-1 would be expected to underestimate risk to
the maximum exposed individual. These values are therefore not appropriate to rely on to
attempt to protect health from unacceptable hazardous exposure in a § 7412(f)(2) residual risk
analysis.
Further, the D.C. Circuit recently held in reviewing an EPA risk assessment pursuant to
§ 7412(d)(4) that EPA had not sufficiently justified refusing to use the most up-to-date Cal. EPA
OEHHA reference values. 86 In this instance, EPA’s refusal to use the most up-to-date 2014 REL
for benzene is just as arbitrary and capricious as the court found in that case. EPA attempts to
respond to that case in the Response to Comments, but all EPA does is point to a prior decision
under § 112(f)(2) that did not explicate the “ample margin of safety” text specifically. RTC at
23. 87 That statement fails to explain how this responds to the holding, similarly applicable here,
that EPA has failed to justify not applying the OEHHA value. Further, the two cases should be
read consistently; the D.C. Circuit in the vinyl chloride decision recognized that EPA must
perform a second step to determine the “ample margin of safety to protect public health” just as
the court held in the brick kilns case. Compare NRDC v. EPA, 824 F. 2d at 1165, 1166, with
Sierra Club v. EPA, 895 F.3d at 1, 13. EPA’s citation to a separate program and statutory
section of the Clean Air Act – the NAAQs as opposed to the air toxics section – and to a
different statutory phrase “adequate margin of safety” instead of “ample margin of safety” does
not support its argument. RTC at 23. The fact that EPA and the court have interpreted the
entirely different NAAQS framework and different text differently is logical and irrelevant to the
issue here. The exact same text is used within the air toxics section 112; EPA has no statutory or
rational justification for refusing to interpret this consistently.
In view of EPA’s own scientific guidelines and practice to use the REL, EPA’s sudden
change without taking public notice-and-comment on the use of and calculation of acute risk
based on an emergency value, and failing to say what this is (and whether it is the ERPG-1 or
something else) instead of the REL, without providing a detailed explanation for changing the
risk factor it uses to reach the acceptability finding is arbitrary. See, e.g., FCC v. Fox, 556 U.S.

85

SAB May 2010 at 6 (emphasis added) (attached to 2019 Comments).
See, e.g., Sierra Club v. EPA, 895 F.3d at 11-12.
87
EPA states: “Specifically, we note that the ample margin of safety determination, under section
112(f)(2) is conducted separately, in accordance with the two-step framework set forth in the Benzene
NESHAP and NRDC v. EPA (the Vinyl Chloride Decision), 824 F. 2d at 1165, 1166 (D.C. Cir. 1987) in
contrast to the ample margin of safety determination that is called for in setting health based standards
under section 112(d)(4). See for example, NRDC v. EPA, 902 F. 2d 962, 973-74 (D.C. Cir. 1990)
(distinguishing the NAAQS process, whereby the margin of safety analysis is incorporated as part of the
standard without a two-step analysis, from residual risk determinations).” RTC at 23.
86

45

 at 515. It is also unlawful because EPA has recognized the need to apply the “best available
science” to protect “the individual most exposed” to emissions under § 7412(f)(2).
In addition and independently, EPA did not provide public notice of, and does not cite or
place the “examination” or any evidence on which it relies into the docket to find lower acute
risk from benzene. Yet EPA relies on such undisclosed information and an unidentified factor to
refuse to find risk unacceptable and strengthen the emission standards. Therefore, EPA has
violated both notice-and-comment requirements and the procedural rulemaking requirement to
include such material with the final rule to allow for effective judicial review. 42 U.S.C.
§ 7607(d)(3)-(d)(6), 7607(d)(9), 7607(h).
C. This Objection to EPA’s Determination of the Acute Risk Value for Benzene Is Of
Central Relevance Because This Is A Basis for EPA’s Decision Not To Set Stronger
Standards Pursuant to § 7412(f)(2).
This objection is “of central relevance” to the rule because EPA has relied in part on the
allegedly, low acute risk (stated to be below the HQ of 1) to attempt to justify its finding that
health risks from ethylene production sources are “acceptable” under § 7412(f)(2). 85 Fed. Reg.
at 40,392-93. That is the core determination under this provision at step one of the regulatory
stage. If EPA finds risk unacceptable, it must reduce risk to acceptable levels without
consideration of cost. Then, it must provide an “ample margin of safety to protect public
health.” 42 U.S.C. § 7412(f)(2). EPA’s use of an undisclosed emergency-based factor, and
assessment that there is lower acute health risk based on that factor is of “central relevance” to
the outcome of the final rule: EPA’s determination not to set stronger standards pursuant to
§ 7412(f)(2) to reduce benzene and other emissions from ethylene production sources.
Consequently, community residents, including Petitioners’ members, face higher benzene
emissions, and co-emitted hazardous air pollutants, than they would if EPA were to recognize the
need to reduce the benzene emissions. Therefore, EPA must convene proceedings to reconsider
the finding on acute risk from benzene, and the finding of acceptable risk and regulatory
determination based in part on that finding.
VI.

EPA MUST GRANT RECONSIDERATION ON ITS RISK ASSESSMENT
AND § 7412(F)(2) DETERMINATION DUE TO NEW INFORMATION EPA
MUST USE TO ASSESS HEALTH RISKS FROM TOLUENE.

A. Inability to Raise the Objection in Comments.
Toluene is a pollutant known to have neurotoxic effects to the central nervous system
including disabilities, and that can also cause sensory harm (i.e., eye and upper respiratory
irritation), or even vision impairment. 88 Because of their developing nervous systems, Cal. EPA
OEHHA has recognized that “[c]hildren may be more sensitive to toluene neurotoxicity than
88

Supra note 10, Cal. EPA OEHHA, Toluene RELs (Aug. 20, 2020),
https://oehha.ca.gov/media/downloads/crnr/toluenerel082020.pdf.

46

 adults because of the sensitivity of their developing nervous systems to neurotoxicants.” 89
Ethylene production sources (31) emit approximately 347 tons per year of benzene, according to
EPA—the second largest amount of any emitted hazardous air pollutant. Final RRA at 36 tbl.
3.1-1.
In the proposed and final 2020 Rule for Ethylene Production, EPA assessed risk from
toluene using the following health reference values:
(1) Acute: 37 mg/m3 (old REL) = 37,000 µg/m3 (based on the prior OEHHA REL).
(2) Chronic non-cancer Rfc: 5 mg/m3 = 5000 µg/m3 (based on the 2005 IRIS value).
Final RRA at 36 tbl. 3.1-1.
The use of these factors formed part of the basis for EPA’s chronic non-cancer and acute
risk assessments, and EPA’s determinations that chronic non-cancer and acute risk are
“acceptable,” and that no further emission reductions were needed to provide an “ample margin
of safety” under § 7412(f)(2). 85 Fed. Reg. at 40,391-92.
In 2020, OEHHA adopted more protective benchmarks for toluene exposure:
(1) Acute REL (for a 1-hour exposure): 5000 micrograms per cubic meter (µg/m3)
(2) 8-Hour REL (for repeated 8-hour exposures): 830 µg/m3
(3) Chronic non-cancer REL (for long-term exposures): 420 µg/m3
The 2020 Cal. EPA supportive documentation for these new values is attached. 90
These new Cal. EPA OEHHA values are an order of magnitude more protective than the
values EPA used for toluene.
EPA’s own scientific guidelines, as affirmed by the Science Advisory Board, direct EPA
generally to apply the agency’s own IRIS toxicology values, then ATSDR, and then California
EPA OEHHA values as the priority health reference or dose-response values. 91 As EPA
explained in the proposed and final Residual Risk Assessment, “The prioritization process was
aimed at incorporating into our assessments the best available science with respect to doseresponse information.” Final RRA at 24. 92 EPA has prioritized use of the Cal. EPA OEHHA
values, as the only dose-response values EPA applies from any state. This is because, as EPA has
89

Cal. EPA OEHHA, Notice of Adoption of Reference Exposure Levels for Toluene (Aug. 20, 2020),
https://oehha.ca.gov/air/crnr/notice-adoption-reference-exposure-levelstoluene?utm_source=Toluene+REL+Adoption&utm_campaign=Proposition+65++Proposed+Maximum+Allowable+Dose+Levels+for+Chlorpyrifos&utm_medium=email (attached).
90
Supra note 10.
91
SAB May 2010 at ii, 5-6 (“The Panel found EPA’s approach to selecting dose-response chronic toxicity
values to be generally sound.”).
92
EPA further explains that it has prioritized use of the Cal. EPA reference dose values because of “(1)
conceptual consistency with EPA risk assessment guidelines and (2) level of peer review received.” RRA
at 24; see also id. at 25-26.

47

 explained: “The process for developing these assessments is similar to that used by EPA to
develop IRIS values and incorporates significant external scientific peer review.” Id. at 25-26. It
appears that in the proposed rule, EPA had simply not updated the reference values yet to
recognize the changes in 2020, as it should now do to follow its own guidelines, but it was
applying the prior OEHHA REL for acute risk at least.
The new OEHHA reference values were adopted after the comment period closed. Thus,
it was “impracticable” to object to use of the factors EPA used for toluene in the proposed
Residual Risk Assessment for ethylene production. Further, the grounds for the below-discussed
objections arose after the close of the comment period under § 307(d)(7)(B), when OEHHA
adopted the new RELs.
D. Objection: EPA Unlawfully and Arbitrarily Used Outdated Health Reference
Values to Assess Health Risks from Toluene and This Led to an Unlawful and
Arbitrary Risk Assessment and § 7412(f)(2) Determination.
In the 2020 Rule, EPA’s new rationale for finding low acute and chronic non-cancer risks
and for determining that health risks are acceptable is that it has used outdated acute and chronic
non-cancer values for toluene. RRA at 36 (citing the old values).
EPA generally applies the OEHHA RELs. EPA previously applied the acute REL for
toluene, and now that OEHHA has updated this value EPA should apply the new, much more
protective value. Doing so would lead to an order of magnitude increase in the acute risk found
for toluene.
In addition, EPA should apply the more protective OEHHA value for chronic non-cancer
risk than EPA’s now 15-year old IRIS value. 93 That was a value IRIS created before knowledge
of the Cal. EPA OEHHA and the studies on which it relies—particularly the evidence showing
significant disproportionate neurological harm to children from exposure, and in some instances,
disability due to prolonged exposure. See OEHHA REL at 1. The OEHHA value is more
protective based on newer information. EPA’s 2005 IRIS value did not consider early-life
exposure, harm to children, disabilities among children as the OEHHA value discusses, or apply
any factors to account for the disproportionate risk children face. 94 In such circumstances,
EPA’s own prioritization policy directs it to use the best available science, which would include
the Cal. EPA OEHHA value. Consistent with EPA’s longstanding policy, OEHHA’s more
protective value meets EPA’s consistency, reliability, and peer-review qualifications and should
be used. At a minimum, before refusing to use OEHHA’s more current and more protective
value, EPA must evaluate it on reconsideration and when doing so EPA will likely find that this
like the acute REL meets EPA’s criteria for the best available science.

93

EPA IRIS, Chemical Assessment Summary: RfC for Toluene (Sep. 23, 2005),
https://cfpub.epa.gov/ncea/iris/iris_documents/documents/subst/0118_summary.pdf#nameddest=rfc
(attached).
94
Id.

48

 The 2020 OEHHA RELs are well-supported and reflect the best available science. 95
These are the peer-reviewed factors that best meet EPA’s own scientific integrity principles, and
that is why EPA’s own Risk Assessment prioritizes use of the OEHHA values. Final RRA at 2426.
The evidence shows EPA’s risk assessment significantly undercounts the chronic and
acute health hazards of the regulated facilities’ emissions, and thus the risk to “the individual
most exposed” to ethylene production sources’ emissions. It is both unlawful and arbitrary and a
failure to apply the best available science for EPA not to use the 2020 OEHHA values. In
addition to underestimating acute and chronic non-cancer health hazards, EPA is willingly and
knowingly exposing communities of color and low-income communities to increased and
unnecessary harm from toluene exposure. Because, as discussed in the prior section, EPA is also
undercounting the acute risk from benzene, its simultaneous refusal to apply the best available
science to account for toluene means that EPA’s acute and chronic risk both substantially
underestimate the health risk to “the individual most exposed,” requiring reconsideration.
Further, the D.C. Circuit recently held in reviewing an EPA risk assessment pursuant to
§ 7412(d)(4) that EPA had not sufficiently justified refusing to use the most up-to-date Cal. EPA
OEHHA reference values. 96 In this instance, EPA’s refusal to use the most up-to-date 2020
RELs for toluene is just as arbitrary and capricious as the court found in that case. EPA attempts
to respond to that case in the Response to Comments, but all EPA does is point to a prior
decision under § 112(f)(2) that did not explicate the “ample margin of safety” text specifically.
RTC at 23. 97 That statement fails to explain how this responds to the holding, similarly
applicable here, that EPA has failed to justify not applying the OEHHA value.
Further, it is possible to read the two cited cases consistently, as the D.C. Circuit in the
vinyl chloride decision, recognized that EPA must perform a second step to determine the
“ample margin of safety to protect public health” just as the court held in the brick kilns case.
Compare NRDC v. EPA, 824 F. 2d at 1165, 1166, with Sierra Club v. EPA, 895 F.3d at 1, 13].
EPA’s citation to a separate program and statutory section of the Clean Air Act – the NAAQs as
opposed to the air toxics section – and to a different statutory phrase “adequate margin of safety”
instead of “ample margin of safety” does not support its argument. RTC at 23. The fact that
EPA and the court have interpreted the entirely different NAAQS framework and different text
differently is logical and irrelevant to the issue here. The exact same text is used within the air
toxics section 112; EPA has no statutory or rational justification for refusing to interpret this
consistently. Thus, EPA also has failed to satisfy the Act’s directive to provide an “ample
margin of safety to protect public health,” because EPA has provided no such margin at all as
Commenters explained (Comments at 94-95), as well as failing to protect “the individual most
exposed” to ethylene production sources’ emissions from health risks. In view of EPA’s own
scientific guidelines and practice to use the OEHHA values, EPA’s refusal to apply the most
updated OEHHA RELs, without providing a detailed explanation is arbitrary. It is also unlawful
because EPA has recognized the need to apply the “best available science” to protect “the
individual most exposed” to emissions under § 7412(f)(2).
95

See OEHHA RELs for Toluene, supra note 10.
See, e.g., Sierra Club v. EPA, 895 F.3d at 11-12.
97
Supra note 83.
96

49

 B. This Objection to EPA’s Inadequate Determination of the Health Risks from
Toluene Is Of Central Relevance Because This Is A Basis for EPA’s Decision Not To
Set Stronger Standards Pursuant to § 7412(f)(2).
This objection is “of central relevance” to the rule because EPA has relied in part on the
allegedly, low acute and chronic non-cancer risk (stated to be below the TOSHI of 1 and an HQ
of 1) to attempt to justify its finding that health risks from ethylene production sources are
“acceptable” under § 7412(f)(2). 85 Fed. Reg. at 40,392-93. That is the core determination under
this provision at step one of the regulatory stage. If EPA finds risk unacceptable, it must reduce
risk to acceptable levels without consideration of cost. Then, it must provide an “ample margin
of safety to protect public health.” 42 U.S.C. § 7412(f)(2). EPA’s failure to use the best
available science on health risks from toluene, and resulting assessment that there is lower acute
health risk than would be shown using the OEHHA reference values are of “central relevance” to
the outcome of the final rule: EPA’s determination not to set stronger standards pursuant to
§ 7412(f)(2) to reduce toluene and other emissions from ethylene production sources.
Consequently, community residents, including Petitioners’ members, face higher toluene
emissions, and co-emitted hazardous air pollutants, than they would if EPA were to recognize the
need to reduce the toluene emissions. Therefore, EPA must convene proceedings to reconsider
the findings on acute and chronic non-cancer risk from toluene, and the finding of acceptable risk
and regulatory determination based in part on that finding.
VII.

EPA MUST GRANT RECONSIDERATION ON ITS NEW RATIONALE FOR
FAILING TO REQUIRE FENCELINE MONITORING.

A. Inability To Raise the Objections in Comments.
In the proposed rule and preamble to the proposed rule, EPA did not mention, much less
propose to require, fenceline monitoring or a fenceline corrective action level for any HAPs
emitted by ethylene production sources. 84 Fed. Reg. at 54,278-352. In the proposed rule
preamble, EPA did not acknowledge fenceline monitoring as a “development” under
§ 7412(d)(6), nor did EPA state that it was choosing to not require fenceline monitoring or an
associated corrective action level for total fugitive emissions. Id. Rather, the preamble to the
proposed rule stated that EPA identified two other equipment leak-related developments. 98
Petitioners commented that EPA must require fenceline monitoring and a corrective
action level at least as stringent as that required for benzene in the Petroleum Refineries Sector
rule, 40 C.F.R. Part 63 Subpart CC, 40 C.F.R. § 63.658; Petitioners’ 2019 Comments at 183-198.
Petitioners argued that refusing to do so would violate the Act’s prohibition against arbitrary and
capricious agency action and would fail to meet the Act’s mandates to “take into account
98

Proposed Rule, 84 Fed. Reg. at 54,316 (“Our technology review for equipment leaks identified two
developments in LDAR practices and processes: (1) Lowering the leak definition for valves in gas and
vapor service or in light liquid service from 500 ppm to 100 ppm and (2) lowering the leak definition for
pumps in light liquid service from 1,000 ppm to 500 ppm.”).

50

 developments in practices, processes, and control technologies,” § 7412(d)(6), to protect public
health with an “ample margin of safety” under § 7412(f)(2), and to require enhanced monitoring
for major stationary sources and specify conditions to assure continuous compliance with the
Act’s requirements. Petitioners’ 2019 Comments at 185-200; see also 42 U.S.C. §§ 7607(d)(9),
7414(a)(3), § 7661c(a); 40 C.F.R. § 70.6(a)(3), (c)(1).
In the Final Rule, EPA refused to require fenceline monitoring to protect public health.
In the Response to Comments accompanying the Final Rule, EPA expressly acknowledged that
EPA “evaluated fenceline monitoring as a development,” RTC at 192. Yet, EPA failed to revise
the standards “as necessary” by “taking into account [this development].” 42 U.S.C. §
7412(d)(6); RTC at 192. EPA also refused to require fenceline monitoring and a fenceline
corrective action level to assure the “ample margin of safety to protect public health” that §
7412(f)(2) directs the standards must provide. EPA declined to require fenceline monitoring to
assure compliance with existing, equipment-specific fugitive emission standards, such as the
equipment leak and PRD work practice standards, even though compliance assurance was part of
the justification for requiring such monitoring in the Refinery Rule, as Petitioners observed in
their comments, supra. EPA additionally arbitrarily failed to require such monitoring for this
source category, in violation of the Act and in conflict with these core statutory objectives. 42
U.S.C. § 7607(d)(9).
In contrast with its inaction here, EPA proposed and finalized fenceline monitoring
requirements for petroleum refineries because the agency determined that such monitoring is a
“development in processes, practices and control technologies for measuring and controlling
fugitive emissions from petroleum refineries,” and that it is “necessary” under section 112(d)(6)
of the Act. Refinery Rule RTC at 209. EPA also finalized the requirement applicable to
refineries because it found that fenceline monitoring was necessary to assure compliance with
the emission standards. 79 Fed. Reg. 36,880, 36,920 (June 30, 2014); 2015 Refinery Rule RTC at
259. In that rulemaking, EPA explained that the agency “expect[s] and encourage[s] refinery
owners or operators to proactively use the monitoring data to identify and correct high emitting
sources.” Refinery Rule RTC at 259. Pursuant to the Refineries Rule, operators must measure the
annual average ambient benzene concentration at their fencelines on a two-week rolling average,
determine if the annual average exceeds 9 µg/ m3 —the corrective action threshold established
by EPA – and take necessary action to reduce the ambient fenceline concentration below this
level. 79 Fed. Reg. at 36,922; EPA, Fenceline Monitoring Technical Support Document at 3,
EPA-HQ-OAR-2010-0682-0210; see also 40 C.F.R. §63.658(a), (e).
In its response to Petitioners’ comments on ethylene production, EPA recognized that
fenceline monitoring “is a recent development and viable monitoring technology for fugitive
emissions management.” RTC at 192. But EPA gave the following novel reasons to try to justify
refusing to require fenceline monitoring here:
(1) EPA newly claims that Petitioners have not highlighted “HAP fugitive emission
sources [that] EPA has failed to previously regulate and set MACT standards for
in the Ethylene Production source category,” RTC at 192.
(2) Attempting to respond to Petitioners’ comment that the agency should require
fenceline monitoring here as it did for petroleum refineries, EPA newly claims

51

 that “for petroleum refinery sources, that rulemaking record showed sufficient
evidence that emissions from their sources may be understated” but that the
ethylene production record shows no such evidence. Id. at 192. This new rationale
relies on but is distinct from the rationale offered in the Equipment Leaks Memo,
which focused solely on “magnitude” and “uncertainty” rather than the contention
that fugitive emissions from refineries are “understated” to a greater degree than
fugitive emissions from ethylene production facilities. Id. The relative magnitude
of fugitive emissions stemming from ethylene production facilities is unconnected
to whether reported fugitive emissions from those facilities are understated.
Similarly, “uncertainty” captures the possibility that reported fugitive emissions
are consistently overstated or fluctuate in accuracy. EPA relatedly contends that
fugitive emissions from ethylene production sources are “more tightly controlled”
than at refineries. Id. EPA cites its statement that “there is some uncertainty in the
cost-effectiveness of implementing such a program where it is anticipated that the
magnitude of emissions reduction would be less.” Equipment Leaks Memo at 9
(cited by RTC at 192) (emphasis added). EPA also cites and reaffirms its
recognition that the refinery rule’s sorbent tubes-based fenceline monitoring
method is “inexpensive,” even though that would support its use here. Id.
(3) Faced with evidence submitted by Petitioners that several ethylene production
facilities presently implement fenceline monitoring as a best practice, EPA newly
claims that the agency is “afforded significant discretion when it comes to
implementing compliance assurance practices such as monitoring.” RTC at 192.
EPA states that it will not require this despite acknowledging that “some ethylene
production facilities are currently performing various types of fenceline
monitoring [such as] ethylene production facilities collocated with petroleum
refineries subject to the fenceline monitoring provisions . . . facilities under
consent decree via enforcement actions, or facilities who have agreed to perform
some level of fenceline monitoring via settlement agreements with nongovernmental organizations. . .” RTC at 192-93.
(4) EPA newly claims that fenceline monitoring is not required to protect public
health under § 7412(f)(2) because “from a public health perspective, we note that
the facility contributing to the highest MIR for the source category of 100-in-1million, Shell Norco, is currently conducting fenceline monitoring for benzene
given that it is collocated at a petroleum refinery.” RTC at 193.
(5) Finally, EPA contends that it cannot require fenceline monitoring because
“finalizing any such fenceline monitoring provisions would violate our clear
notice and comment obligations under CAA section 307(d).” Id.
In the Notice of Proposed Rulemaking published in the Federal Register, EPA did not
mention fenceline monitoring or provide these rationales for not requiring this in the preamble to
the proposed rule for public notice-and-comment. 84 Fed. Reg. at 54,278. As the D.C. Circuit
explained in CCAN, EPA must grant reconsideration where EPA has failed to disclose “in the
[Notice of Proposed Rulemaking]” the agency’s “process” for applying the Act and the “critical
reasoning behind its . . . analysis.” 952 F.3d at 320. Petitioners could not have commented on the
52

 above-listed rationales that EPA did not provide until the Response to Comments was released
after publication of the Final Rule. As discussed below, EPA’s new rationales and consequent
refusal to require fenceline monitoring are unlawful, unavailing and misleading. Because EPA
unveiled the new rationales in order to defend its failure to require fenceline monitoring and
corrective action to reduce total fugitive emissions solely in the Final Rule, it was
“impracticable” within the meaning of Clean Air Act § 307(d)(7)(B) to object to these rationales
during the public comment period. 42 U.S.C. § 7607(d)(7)(B). Petitioners’ grounds for
objection to the new rationales arose after the close of the comment period, consistent with
§ 307(d)(7)(B).
A. Objection: EPA’s New Rationales for Refusing to Require Fenceline Monitoring
are Arbitrary and Unlawful under the Act
EPA’s novel rationales lend no legitimacy to the agency’s unlawful failure to require
fenceline monitoring and to set a corrective action level for total fugitive emissions at the
fenceline.
First, EPA’s contention that Petitioners’ comments did not highlight particular,
unregulated fugitive emissions showing the need for fenceline monitoring is incorrect. Notably,
EPA shied away from expressly claiming that there are no such uncontrolled fugitive emissions
from this source category. That is for good reason: Petitioners’ comments squarely targeted
EPA’s refusal to set adequate standards for fugitive emission sources by challenging the routine
and “force majeure event” exemptions for uncontrolled PRD releases and flare emissions that
EPA proposed and has now finalized. Petitioners’ comments at 104-161; see Parts I-III, supra.
EPA’s refusal to restrict these emissions at all during the malfunction periods EPA defines as not
violating the work practice standards means that the emissions constitute uncontrolled leaks and
create other types of fugitive emissions. This is the very type of emission for which fenceline
monitoring would make a serious difference in addressing, just as EPA explained in the 2015
Refinery Rule. See, e.g., 79 Fed. Reg. at 36,920.
Moreover, EPA’s obligation to require fenceline monitoring pursuant to §§ 112(d)(2)-(3),
(6), and (f)(2), does not depend on whether EPA has regulated all types of fugitive emissions
within the source category as EPA now claims. RTC at 192-93. As EPA recognized in the
refineries rulemaking, fenceline monitoring complements and strengthens regulation of specific
emission sources and types and assures the maximum achievable emissions reduction,
§ 112(d)(3), by compensating for the “impracticality of directly monitoring many of these
fugitive emission sources on a regular basis,” and allowing facilities to “identify a significant
increase in emissions” in a more “timely manner.” 79 Fed. Reg. at 36,920. EPA arbitrarily fails
to address or distinguish any of these specific rationales cited by the agency in the Petroleum
Refinery Sector rulemaking. 42 U.S.C. § 7607(d)(9). Thus, even if it were the case that EPA
has controlled all fugitive emissions from ethylene production sources, EPA has failed to justify
not requiring fenceline monitoring to assure compliance with such standards, to “tak[e] into
account” this development “as necessary” under § 7412(d)(6), and to assure fugitive emissions
do not increase to dangerous levels under § 7412(f)(2).
Second, EPA’s new reliance on the agency’s supposed “significant discretion” in
implementing monitoring practices is misleading and unavailing. RTC at 192. Fenceline

53

 monitoring is not solely a compliance assurance practice for existing emissions standards, though
that is one important value. Implementing fenceline monitoring would permit EPA to set a
standard limiting total benzene concentrations at the fenceline, as the agency did in the
Refineries Sector rulemaking, and to mandate corrective action to reduce excess ambient
fenceline benzene concentrations to a health-protective threshold. The corrective action level and
requirements associated with such fenceline monitoring – again, as the agency required for
petroleum refineries, 79 Fed. Reg. at 36,922; 40 C.F.R. §63.658(a), (e) – could be used to require
at least some facilities to further reduce fugitive emissions beyond the reductions necessitated by
existing emissions limits applying to individual fugitive sources, as well as to assure compliance
with those limits by finding leaks and problems. In the Organic Liquids Distribution (“OLD”)
rulemaking, EPA acknowledged that fenceline monitoring was at least and possibly more
effective at controlling fugitive emissions than some of the source-specific control methods EPA
had in place and was considering. OLD Proposed Rule, 84 Fed. Reg. 56,288 (Oct. 21, 2019). 99
Thus, fenceline monitoring is not merely a compliance monitoring method as EPA now attempts
to label it.
Indeed, requiring fenceline monitoring and associated corrective action for excess
fenceline benzene levels is non-discretionary under the Act. Under the Act, any work practice
requirements that EPA promulgates under § 112(h) must be consistent with § 112(d) and reduce
emissions to the “maximum” degree that is “achievable” and, at a minimum, to the average level
already “achieved” by the relevant best performing sources. 42 U.S.C. §§ 7412(d)(2)-(3),
7412(h). See also U.S. Sugar, 830 F.3d at 608 (“EPA would need to determine that the standard
would ‘reduce emissions of hazardous air pollutants.’”).
Fenceline monitoring must therefore be considered and required as part of the standard
assuring emissions reductions to the MACT floor for each newly established work practice
standard applying to types of fugitive emissions. In the Final Rule, EPA promulgates several new
work practice standards applicable to specific types of fugitive emissions at ethylene production
facilities. See, e.g. 40 C.F.R. § 63.1103(e)(5); 85 Fed. Reg. at 40,400 (describing work practice
standards applying to ethylene process vents for the ethylene production source category); 40
C.F.R. § 63.14; Fed. Reg. at 40,397 (describing work practice standard applying to heat
exchange system leaks for ethylene production facilities). EPA has also recognized that several
ethylene production facilities currently perform fenceline monitoring to reduce overall fugitive
emissions, as Petitioners noted in comments at proposal. RTC at 192-93 (acknowledging that
several ethylene production facilities are performing fenceline monitoring, including those
collated with petroleum refineries and those subject to consent decrees or settlement
agreements.)
Work practice standards consistent with the MACT floor may, and must, include the
institution of a continuous monitoring program, where such a practice is necessary for the
implementation of corrective action to reduce emission levels to the maximum degree feasible.
42 U.S.C. § 112(h), § 112(d)(2)-(3). EPA has recognized as much both in the Final Rule and in
prior rules promulgated pursuant to the Act. See, e.g. 40 C.F.R. § 63.8192 (requiring the
99

In the OLD Final Rule, EPA did not finalize the fenceline monitoring requirements it had proposed but
it also did not abandon or change its conclusion that it would have been particularly effective at finding
and reducing fugitive HAP emissions. 85 Fed. Reg. 40,740 (July 7, 2020).

54

 institution of a cell room monitoring program to continuously monitor mercury vapor
concentrations as part of a work practice standard applying to mercury cell chlor-alkali plants).
EPA expressly stated that fenceline monitoring was required consistent with the MACT floor for
fugitive emissions in the petroleum refineries rulemaking. 79 Fed. Reg. at 36,920 (“We sought to
develop a maximum fenceline concentration action level that is consistent with the emissions
projected from fugitive sources compliant with the provisions of the refinery MACT standards as
modified by the additional controls proposed in this action . . .”). Indeed, EPA proposed and
finalized the fenceline monitoring requirements for petroleum refineries because the agency
found that “[h]istorically, improved information through measurement data has often led to
emission reductions.” 79 Fed. Reg. at 36,920. Therefore, implementation of fenceline
monitoring, and requiring corrective action for total fugitive emissions exceedances, is necessary
to reach the MACT floor achieved by the best performing sources. EPA has arbitrarily failed to
demonstrate otherwise, in violation of the Act. 42 U.S.C. § 7607(d)(9).
EPA has not argued that instituting fenceline monitoring would not be cost effective. It
could not do so based on the record. Just as well: the agency may not consider costs in setting
initial maximum achievable control technology “floors;” it may do so only in determining
whether to require “beyond the floor” reductions in emissions. See Nat’l Lime Ass'n
v. EPA, 233 F.3d at 640; see also NRDC v. EPA, 489 F.3d 1364, 1375-76 (D.C. Cir.
2007). Instead, EPA has admitted that the costs for implementing fenceline monitoring are
modest. RTC at 192 (citing Equipment Leaks Memo (EPA-HQ-OAR-2017-0357-0010) at 9
(“implementation of a fenceline monitoring program such as the one finalized in the petroleum
refinery NESHAP is quite inexpensive”)).
EPA has also failed to revise the standards “as necessary” by “taking into account [this
development].” 42 U.S.C. § 7412(d)(6); RTC at 192. Section 112(d)(6) commands the
Administrator to “review, and revise as necessary” the technology-based standards in light of
technological developments at least every eight years. 42 U.S.C. § 7412(d)(6). Nat. Res. Def.
Council v. E.P.A., 529 F.3d 1077, 1080 (D.C. Cir. 2008). Where “developments” have occurred,
it is “necessary” for EPA to assure that the emissions standards “tak[e] into account” those
developments under § 7412(d)(6). If developments are present, EPA must set standards
accounting for those developments under the Act. Section 112(d)(6) requires EPA to impose
standards accounting for any developments that EPA identifies as part of the technology review.
Where a development exists demonstrating the achievability of further emission reductions or
control improvements, EPA must revise the standards to reflect that development and ensure the
“maximum” degree of emission reductions that is “achievable.” 42 U.S.C. § 7412(d)(2). If EPA
declines to revise the standards based on those developments, EPA must demonstrate that
implementation of the developments is not “achievable,” or would not result in any further
source category emissions reductions consistent with § 7412(d)(2), the threshold for what
emissions standards EPA deems “necessary” under that subsection—that is, the “maximum
degree of reduction in emissions that is deemed achievable.” 42 U.S.C. § 7412(d)(6).
EPA has failed to make such a demonstration here. Indeed, such a demonstration cannot
be made given that the agency has previously found fenceline monitoring to be achievable and
effective in reducing fugitive emissions in the refineries rulemaking. EPA even admits this
would be “inexpensive” here. Equipment Leaks Memo (-0010) at 9. Due to the history of
malfunctions under the prior SSM exemption, and the need to assure industry compliance with

55

 the existing and new standards EPA has set, the record shows fenceline monitoring would likely
assure greater fugitive emission reductions and control here. Thus, the agency identified
fenceline monitoring as a development pursuant to § 7412(d)(6) and subsequently failed,
unlawfully, to require such monitoring or to require the fenceline corrective action level that
such monitoring would permit. RTC at 192. Even if § 112(d)(6) permitted consideration of cost,
imposition of a fenceline monitoring program is still necessary because EPA correctly found that
fenceline monitoring is a development that is cost-effective and technically feasible. Equipment
Leaks Memo (-0010) at 9.
Third, EPA offers no evidence to support the agency’s novel claim that fugitive
emissions from the Ethylene Production source category are understated to a lesser degree (in
addition to being more “uncertain,” as the agency argued at proposal) than fugitive emissions in
the Refinery Sector. EPA’s rationale, wholly unsupported by direct evidence or logic, is
therefore arbitrary and capricious under the Act. 42 U.S.C. § 7607(d)(9); see also RTC at 19293. To attempt to bolster its argument, EPA claims that emission standards for individual fugitive
emissions sources, such as those that apply to certain leaks, are more stringent in the Ethylene
Production source category than the Refinery Sector category. RTC at 192-93. However, EPA
has failed to rationally articulate how the stringency of the emissions limitation is connected to
the accuracy or certainty of reported emissions for individual sources or compliance with such a
limitation. The data EPA collected for this rule suggest that ethylene production facilities’
reported fugitive emissions are similar to those of refineries in type and in control needs. See
ICR Data Memo and App. D (providing data reported in response to the information collection
request). And, the malfunctions reports in the record show that ethylene production facilities’
fugitive emissions are likely substantially understated, as malfunctions are rarely accurately or
completely reported. Id.
And, the agency implicitly refuted its own irrational claim that the stringency of the
emissions limitation is linked to the accuracy of reported emissions in the refineries rulemaking,
where EPA stressed that uncertainty in determining the amount of fugitive emissions resulted
from the infrequency of required inspection and monitoring for certain individual fugitive
emission sources, not of the degree of emissions limitations applying to those sources. 79 Fed.
Reg. at 36,920. Yet, EPA here relies on such an assertion to attempt to distinguish ethylene
production facilities as needing neither fenceline monitoring nor a fenceline corrective action
level, ignoring the counterevidence and without providing adequate evidence to support its
claim. The agency has further failed to make any comparison between the frequency of
inspecting or monitoring of individual sources’ fugitive emissions in the Refinery Sector and the
Ethylene Production Sector, and thus cannot justify treating these source categories differently
with respect to requiring fenceline monitoring and a fenceline corrective action level. If there is
any actual evidence behind EPA’s unsupported assertion, EPA’s refusal to provide the public
with such evidence or the data upon which EPA relies to apparently determine that ethylene
production facilities’ fugitive emissions are less “understated” than those from refineries violates
the Act’s directive to provide adequate comment and notice pursuant to § 307(d). Thus, EPA has
failed to justify refusing to require this common-sense monitoring method here.
Notably, EPA has a history of underestimating fugitive emissions in the rulemaking
process for petrochemical sources. A 2020 Environmental Integrity Project report assessing
refineries’ fugitive emissions shows higher levels of benzene at the fenceline than EPA predicted

56

 would occur in the refineries rulemaking. EIP, Monitoring for Benzene at Refinery Fencelines:
10 Oil Refineries Across U.S. Emitted Cancer-Causing Benzene Above EPA Action Level, Note
25, Addendum (Feb. 6, 2020) (as updated Feb. 18, 2020) (attached). Some refineries even
exceeded the fenceline action level for benzene during 2019. Id. The report demonstrates that it
is likely that EPA undercounted the benzene emissions in its 2015 Residual Risk Assessment for
refineries, and is similarly likely to have done so here. Id.
EPA has recognized strong similarity between the refinery and ethylene production
source categories and their fugitive emission problems throughout this rule, but arbitrarily fails to
do so with respect to fenceline monitoring. 42 U.S.C. § 7607(d)(9). In particular, EPA’s own
decisions in this rule that it is necessary to apply the 2015 Refinery flare improvements to
ethylene production, and its decision to apply the Refinery work practice standards for PRDs and
flares (albeit illegal and arbitrary in regard to the exemptions they contain), only further
demonstrate EPA’s admission of the similarity of these sources and their inadequate control of
fugitive emissions, directly contrary to the rationale presented in the Response to Comments.
RTC at 192-93.
Further, even if EPA’s new rationale claiming that ethylene production facility fugitive
emissions are less “understated” and “uncertain” than those stemming from regulated refineries
were justified – and it is not – that alone could not explain why the agency failed to set a
limitation of the total ambient pollutant concentration near the facility and ensure a holistic cap
on fugitive emissions across all individual fugitive emissions sources to protect public health to
the extent feasible. See 42 U.S.C. §§ 112(h), 112(d)(2)-(3), 7607(d)(9); RTC at 192-93. Thus,
EPA has failed to make this action consistent with the findings and the rationales previously
relied upon by the agency in regulating total fugitive emissions from refineries, and this is
arbitrary and capricious. 42 U.S.C. § 7607(d)(9). In fact, monitoring data collected as part of the
Refinery Sector rule provides some evidence that the emissions from ethylene production
facilities are not understated as suggested in the Response to Comments. For the sampling period
ending on December 12, 2019, Motiva Port Arthur Refinery excluded three monitored fenceline
concentrations from the annual average, citing that an “off-site event from a non-MACT CC
source (Ethylene production facility) impacted monitor.” 100 This off-site event from an ethylene
production facility resulted in a benzene concentration of 122 µg/m3 at the refinery fenceline –
more than 13 times above EPA’s action level, and the highest recorded value at Motiva Port
Arthur Refinery’s fenceline monitoring network to date.
On EPA’s fourth rationale, EPA failed to respond to the comment that EPA should
require fenceline monitoring to provide the “ample margin of safety to protect public health” that
§ 112(f)(2) requires. RTC at 192-193 (making no mention of an “ample margin of safety” to
protect public health). EPA included only one conclusory line in the RTC that does not make
clear whether EPA is responding to Commenters’ argument or trying to justify its refusal to
require fenceline monitoring under § 7412(f)(2). EPA states: “from a public health perspective,
we note that the facility contributing to the highest MIR for the source category of 100-in-1million, Shell Norco, is currently conducting fenceline monitoring for benzene given that it is
collocated at a petroleum refinery.” RTC at 193. EPA appears to be implying that it has
100

Motiva Port Arthur Refinery Fenceline Monitoring Report, Quarter 4, 2019. At Sample Results.
Submitted in compliance with 40 CFR §63.655(h)(8).

57

 protected community members from the source causing the highest inhalation-based cancer risk
(Shell Norco) through the fenceline monitoring requirements in the Refinery Sector Rule. But
EPA’s own regulations belie that.
As EPA is well aware the Refinery Sector Rule provides a method for sources to correct
for any other source emissions. That Rule does not require Shell Norco to monitor the ethylene
production plant, to report the fenceline concentration that the ethylene production unit is
causing or contributing to, or to take any corrective action based on the ethylene production unit.
Thus, the Refinery Rule requirements applicable to the Shell Norco refinery cannot substitute for
setting independent fenceline monitoring requirements for Ethylene Production, including both
monitoring and corrective action. Implementing such requirements would mean that Shell Norco
would be legally required to monitor, report, and reduce emissions if needed based on the
ethylene production unit’s emissions, in addition to the refinery’s emissions. And as discussed in
comments and in the risk section of this petition, EPA has not satisfied § 7412(f)(2) for ethylene
production because EPA has set stronger standards for refineries that do not apply to ethylene
production.
In fact, EPA’s sole citation to the Shell Norco facility sheds light on another reason why
EPA must require fenceline monitoring at ethylene production facilities. The portions of the
Shell Norco facility regulated by the Ethylene Production Rule are, according to EPA, causing an
inhalation-based cancer risk of 100-in-1 million, RRA at App. 10 tbl. 2b. The record actually
shows that the cancer risk is 104-in-1 million and EPA has not explained why it has said the risk
is lower. Id. (for NEI ID # 22089110013831201: MIR based on allowable emissions of 1.04 x
10-4 or 104-in-1 million). Meanwhile, the collocated refinery is causing additional, cumulative
cancer risk for nearby community members, which recent fenceline data indicates may be greater
than the risk calculated for the Refinery Sector Rule. In the first quarter of 2020, the most recent
data publicly available at the date of petition submittal, Shell Norco Manufacturing Complex’s
fenceline benzene concentrations exceeded the annual action level for each sampling period
beginning with the sampling period ending on January 6, 2020 until March 26, 2020. This
included a peak annual average of 13.72 µg/m3, over 50% above the action level, and with a
peak two-week average of 78.1 µg/m3, for the period ending on March 12, 2020. 101 The action
level in the Refinery Sector Rule was established using the maximum annual modeled benzene
concentration at refinery fencelines nationally, coincidentally at the Shell Norco facility. 102
Given that actual sampling data has greatly exceeded the modeled fenceline concentration data
used in the Refinery Sector rulemaking, the cumulative cancer risk from the refinery portion of
the facility may be a more significant contributor. The parts of the facility regulated under the
Ethylene Production Rule are therefore contributing a substantial part of the total cancer risk for
the facility as a whole. Requiring fenceline monitoring and corrective action for fugitives from
only part of the facility when the facility as a whole creates cancer risk above 100-in-1 million,

101

Shell Norco Manufacturing Complex Fenceline Monitoring Report, Quarter 1, 2020. Submitted in
compliance with 40 CFR §63.655(h)(8).
102
Memorandum to Brenda Shine, USEPA. “Fenceline Ambient Benzene Concentrations surrounding
Petroleum Refineries.” January 7, 2014. Docket ID Number EPA–HQ–OAR–2010–0682-0208 (attached).

58

 fails to protect “the individual most exposed” to the ethylene production source category’s
emissions from unacceptable risk as the Act directs.
In view of the serious underestimation of health risks as Comments explained,
Petitioners’ 2019 Comments at 23-97, and EPA’s allowance of uncontrolled emissions in the socalled work practice standards, EPA must set a meaningful buffer to account for uncertainty and
provide the “ample margin of safety to protect public health” that § 7412(f)(2) requires. See,
e.g., Sierra Club v. EPA, 895 F.3d at13. In a different part of the Response to Comments, EPA
attempts to distinguish the terms “ample margin of safety” in § 7412(f)(2) and (d)(4), which the
cited case addressed. RTC at 23. But EPA gives no statutory or rational ground for refusing to
recognize that the precise same language used in both parts of § 7412 can and must be read
consistently. While the provisions include different ways in which the “ample margin of safety”
is used, the D.C. Circuit’s holding on the meaning of that plain text applies equally in
§ 7412(f)(2).
Finally, EPA contends that it would violate notice-and-comment provisions to require
fenceline monitoring. That rationale also fails to support EPA’s inaction. Of course, notice-andcomment is required for any rule of this kind under the Clean Air Act – and EPA neither
proposed to require or not to require fenceline monitoring in the Notice of Proposed Rulemaking.
Nevertheless, EPA could have before now and, still could now take public comment on precisely
such a proposal to require fenceline monitoring and corrective action, just as it did in the
Refinery Rule. That is exactly what Petitioners have called for and continue to advocate in this
petition. EPA’s notice-and-comment rationale, if anything, only provides support for granting
reconsideration on this issue and engaging in the necessary public notice-and-comment to require
fenceline monitoring.
Thus, EPA has failed to assure sufficient monitoring and control of all hazardous air
pollutants as needed to serve the core statutory objectives of reducing air toxics and protecting
public health – and its action is also arbitrary and capricious. See, e.g., Comments at 14-19.
Thus, each of the rationales EPA has provided is unlawful, arbitrary and unsupported by
the record, and does not justify EPA’s refusal to require fenceline monitoring in this rulemaking.
First, Petitioners identified examples of unregulated fugitive emissions in their comments at
proposal, as described above. Fenceline monitoring is, moreover, not merely a compliance
assurance practice since its implementation itself lowers emissions, as EPA has previously
recognized, and allows the agency to set a fenceline corrective action level for total fugitive
emissions. Second, EPA has no discretion to decline to require fenceline monitoring here.
Fenceline monitoring constitutes part of the MACT floor for newly promulgated emissions
standards applying to subcategories of fugitive emission sources – particularly due to the
evidence of its use at some ethylene production and similar sources. 42 U.S.C. §§ 7412(d)(2)-(3),
7412(h). EPA has also failed to revise the standards “as necessary” by “taking into account”
fenceline monitoring as a new development. 42 U.S.C. § 7412(d)(6). Third, EPA’s claim that
fugitive emissions from the Ethylene Production source category are understated to a lesser
degree than fugitive emissions in the Refinery Sector is unsupported and arbitrary. 42 U.S.C. §
7607(d)(9). Fourth, EPA has arbitrarily failed to respond to Petitioners’ comment that EPA
should require fenceline monitoring to provide the “ample margin of safety to protect public
health” that § 112(f)(2) demands. Lastly, EPA cannot rely on the fact that it would need to

59

 engage in a new round of notice and comment rulemaking to avoid complying with the Act’s
dictate that the agency require fenceline monitoring to ensure a protective degree of emissions
reduction. That is precisely what should now happen on reconsideration.
Therefore, EPA must grant reconsideration to propose and finalize fenceline monitoring
and corrective action requirements at least as stringent as the Refinery Sector Rule.
B. These Objections to EPA’s New Rationales on Fenceline Monitoring Are Of Central
Relevance Because They Are A Core Basis for EPA’s Decision Not To Require Such
Monitoring, or to Require Corrective Action for Total Fugitive Emissions, for this
Source Category
These objections are “of central relevance” § 7607(d)(7)(B) because they fatally undercut the
“legality” of EPA’s failure to require fenceline monitoring and corrective action when total
fugitive emissions exceed a certain threshold, providing “substantial support for the argument
that the regulation should be revised.” CCAN, 952 F.3d at 320. Petitioners’ objections that EPA
must require fenceline monitoring and related corrective action for total fugitive emissions
implicate threshold legal tests that would compel the agency set such requirements. 42 U.S.C.
§§ 112(h), 112(d)(2)-(3), 112(d)(6). Further, Petitioners’ objections that EPA’s new rationales
are arbitrary, capricious and without evidence, call into question the core grounds for EPA’s
action and provide a basis for the court to reverse it. 42 U.S.C. § 7607(d)(9).
CONCLUSION
For the reasons this petition provides, EPA must convene a reconsideration proceeding
for the Final Rule. EPA must strengthen the ethylene production standards to satisfy the Act’s
requirements in § 7412(d) and (f)(2), as cited above, to ensure that communities exposed to toxic
air pollution from ethylene production facilities finally receive the emission limits and resulting
health and environmental protections that all Americans deserve and to which people near these
facilities are legally entitled.
We appreciate your thoughtful evaluation of the important issues in this petition and urge
you to act to fulfill your responsibility to protect public health. Please contact me if you have
any questions or would like additional information regarding this petition.

60

 Sincerely,
Emma C. Cheuse
Staff Attorney
EARTHJUSTICE
1001 G. St., NW, Suite 1000
Washington, D.C. 20001
echeuse@earthjustice.org
(202) 667-4500 ext. 5220
Counsel for Petitioners
RISE St. James, Louisiana Bucket
Brigade, Louisiana Environmental
Action Network, Texas
Environmental Justice Advocacy
Services, Air Alliance Houston,
Community In-Power &
Development Association, Clean Air
Council, Center for Biological
Diversity, Environmental Integrity
Project, and Sierra Club
Cc:
Associate General Counsel for the
Air & Radiation Law Office
Office of General Counsel
Mail Code 2344A
U.S. EPA
1200 Pennsylvania Ave., NW.
Washington, D.C. 20460
Mr. Andrew Bouchard,
Sector Policies and Programs Division (E143–01),
Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency,
Research Triangle Park, NC 27711
bouchard.andrew@epa.gov.
Mr. Mark Morris,
Health and Environmental Impacts Division (C539–02),
Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency,
Research Triangle Park, NC 27711
morris.mark@epa.gov
Enc:

Documents cited as Appendix attached

61