Why regulate odors? Two reasons: First, to improve public welfare. Odors from air pollution may be an occasional, minor nuisance in most neighborhoods, but not everywhere. For people living adjacent to chemical plants, refineries, and landfills, obnoxious chemical odors can be a daily occurrence that impairs their quality of life. Second, to protect public health. Reducing odors means reducing chemical emissions that cause odors. Many of those chemicals also pose health hazards. Yes, emissions of dangerous chemicals are already subject to Clean Air Act regulatory programs. But those programs are rife with gaps and limitations. The result? Kimberly A. Terrell summarized a 2022 peer-reviewed study that she published with Gianna St Julien as follows: “Numbers of cancer cases are abnormally high among Louisiana’s heavily polluted, predominantly Black neighborhoods,” many of them located in a heavily industrialized corridor known as “cancer alley.”

Can EPA legally regulate odors? Yes! In fact, both EPA and Congress anticipated that possibility. In its 1971 notice that listed nitrogen oxides as a “criteria pollutant” for which the agency would set National Ambient Air Quality Standards, the agency stated that “evaluation of other air pollutants, including fluorides, polycyclic organic matter, and odorous substances, is being conducted, and the list will be revised as the administrator [i.e., EPA] deems appropriate.” In the 1976 case affirming a district court order that required EPA to list lead for regulation, NRDC v. Train, the Second Circuit quoted legislative history that “Congress expects criteria to be issued for nitrogen oxides, fluorides, lead, polynuclear organic matter, and odors, though others may be necessary.” Congress ordered EPA in the 1977 amendments to the Clean Air Act to evaluate “whether air quality criteria or [NAAQS] should be published under the Clean Air Act for odors.” EPA published its response in 1980, concluding that “federal regulatory involvement in odor control does not appear to be warranted.” That conclusion, however, is more than four decades old. If it was ever defensible, it is out-of-date.

The CAA provides ample authority to regulate odors. The heart of the law is the “criteria pollutant” program. That program is built around two expansive commands to the agency in Clean Air Act Section 109: one, “to protect the public health” with “an adequate margin of safety,” and two, “to protect the public welfare from any known or anticipated adverse effects.” To implement those commands EPA sets “primary” NAAQS, which are “requisite to protect the public health,” and “secondary” NAAQS, “requisite to protect the public welfare.” The act’s definitional section clarifies that welfare includes “effects on economic values and on personal comfort and well-being.” The criteria pollutant program is triggered under Section 108 of the act when EPA adds to a list of pollutants that, in the agency’s judgment, “cause or contribute to air pollution which may reasonably be anticipated to endanger public health or welfare” and which are present “in the ambient air” due to emissions from “numerous or diverse mobile or stationary sources.”

Listing of a criteria pollutant triggers a series of deadlines: EPA has 12 months after the listing to issue a criteria document that “accurately reflect[s] the latest scientific knowledge” with respect to “the kind and extent of . . . effects on public health or welfare.” Simultaneously, EPA must propose primary and secondary NAAQS “for any such pollutant.” The agency then has 90 days to issue final standards. After promulgation of NAAQS, Section 110 of the act provides that states “shall” submit for EPA approval plans for “implementation, maintenance, and enforcement” of those standards within three years (a deadline that EPA can extend for 18 months for plans to attain welfare standards). At least since the 1992 Supreme Court opinion in New York v. United States, however, it has been clear that Congress “may not compel the states to enact or administer a federal regulatory program.” Thus, Congress’s provision for state implementation plans is best read as a strong suggestion. But if a state fails to submit an approvable plan, EPA must step in and issue a federal implementation plan. Under Section 172 of the act, the plan should provide for attainment of a welfare standard “as expeditiously as practicable” after EPA designates an area as nonattainment for the standard.

Why is ambient regulation of odors necessary? First, of course, obnoxious odors inarguably impact welfare. In a 2006 Federal Register proposal in the context of landfill gases, EPA recognized that odor problems create “potential for local property devaluation and poorer quality of life for local residents.” For an agency charged with safeguarding welfare “from any known or anticipated adverse effects,” no more should be required. But an important additional consideration is that protection from odors would come with a co-benefit. It would enhance protection of public health. As a practical matter, when changes are made to polluting facilities to reduce an impact from pollution, those changes generally end up reducing other impacts as well. Those reductions are termed co-benefits (or collateral benefits or ancillary benefits).

One might ask, then, why additional health protections are necessary when EPA has already promulgated primary NAAQS “requisite to protect the public health.” At times, the federal government has seemed to equate compliance with primary standards with sufficient protection of public health. During the Clinton administration, EPA’s Office of Civil Rights issued its notorious Select Steel opinion, which rejected a petition for enforcement of the agency’s environmental justice regulations. EPA’s theory in Select Steel was that without a violation of an EPA health-protection standard, “there is no affected population which suffers ‘adverse’ impacts within the meaning of Title VI” of the Civil Rights Act. How can a minority community complain of a disproportionate impact when there is no impact?

Similarly, during the Trump administration, the agency took the position that a cost-benefit analysis of a hazardous air pollutant regulation would give lesser weight to co-benefits from reductions of criteria pollutants, such as particulate matter and ozone precursors. EPA explained that particulate matter and ozone “are already addressed by the numerous statutory provisions governing criteria pollutants.” The Biden administration, however, reversed that position. EPA stated that reducing pollutant concentrations “below the levels of the NAAQS is reasonable and well-supported by scientific evidence.” This is because, for many pollutants, there is no bright-line threshold below which exposures pose no risk of injury.

Congress has recognized that the criteria program’s health-protection standards, although supposedly set with an “adequate margin of safety,” do not protect against all health impacts. Congress made that recognition explicit in the 1977 Clean Air Act Amendments—when adding the Prevention of Significant Deterioration program to keep clean air from deteriorating to the level of the primary NAAQS. In Section 160 of the act, Congress explained that one of its goals was “to protect public health and welfare . . . notwithstanding attainment and maintenance” of all NAAQS. Senator Edmund Muskie (the father of the Clean Air Act) explained in June 1977, “Even at the national primary standard level, which is the health standard, there are health effects that are not protected against.” A 1977 House report acknowledged, “The idea that the national primary standards are adequate to protect the health of the public has been belied.” Further, “The inadequacies of the standards are substantial both with regard to the pollutants which are regulated and with respect to [EPA’s] failure to regulate others.”

To date, EPA has used the criteria program to set standards for only six pollutants—carbon monoxide, nitrogen dioxide, ground-level ozone, lead, particulate matter, and sulfur dioxide. These six pollutants, however, cover a lot of territory. Ground-level ozone, for example, results from an atmospheric chemical reaction involving sunlight, oxygen, volatile organic compounds (VOCs) and oxides of nitrogen (known as NOx). Implementing the ozone standard, therefore, requires EPA and states to regulate VOCs and NOx as ozone precursors. These include a broad collection of chemicals. Another criteria pollutant, particulate matter, consists of small respirable particles which can include a variety of potentially dangerous chemicals. Criteria-program regulations, however, are not designed to abate all risks associated with the specific chemicals that fit within the categories of VOCs, NOx, or particulate matter. For example, benzene is a VOC and also a carcinogen. The criteria-pollutant program, however, only regulates benzene in its role as ozone precursor. The fact that VOC reductions may also reduce cancer risks from benzene is a co-benefit, but EPA did not calibrate its criteria-pollutant regulations for that purpose.

Similarly, the CAA’s Hazardous Air Pollutant program leaves gaps in the protection of public health. The basic hazardous air pollution standard—known as Maximum Achievable Control Technology—is an emission standard, not an ambient standard. Thus, if a neighborhood is surrounded by three major sources of hazardous air pollution, lawful emissions may be up to three times the standard. A limited exception is the MACT standard for oil refineries, which includes a “work practice standard” that generally requires refinery owners to “monitor benzene concentrations around the fenceline or perimeter of the refinery” and “maintain fenceline benzene concentrations at or below the concentration action level” of 9 micrograms per cubic meter. (The quotes are from a 2020 EPA Federal Register notice.)

For new major sources of hazardous air pollutants (i.e., those that emit at least “10 tons per year or more of any hazardous air pollutant or 25 tons per year or more of any combination of hazardous air pollutants”), EPA must set a minimum “MACT floor” that is at least as stringent as “the emission control that is achieved in practice by the best controlled similar source, as determined by” the agency. For existing major sources, EPA sets the MACT floor at “the average emission limitation achieved by the best performing 12 percent of the existing sources (for which [EPA] has emissions information).” The D.C. Circuit explained in its 2004 Mossville Environmental Action Now v. EPA opinion that “EPA need not base its standards on actual data, but could lawfully rely on estimates drawn from the regulatory data as to what the best performing 12 percent were achieving.” Further, EPA may select a standard that sources among that best performing 12 percent can meet “every day and under all operating conditions.” EPA does, however, have discretion to go beyond the MACT floors and set more stringent standards after considering “the cost of achieving such emission reduction[s], and any non-air quality health and environmental impacts and energy requirements.” For sources that do not qualify as “major” (known as “area sources”), EPA may promulgate more discretionary standards based on “generally available control technologies.”

The act’s hazardous air program also requires EPA—if necessary “to provide an ample margin of safety”—to issue an additional promulgation eight years after finalizing a MACT standard. For carcinogens, EPA has defined “ample margin of safety” to allow a lifetime excess cancer risk of between one-in-one million and one-in-ten thousand. Neither end of this “acceptable risk” range is firmly rooted in policy, but a one-in-one million risk is roughly consistent with the traditional concept of a de minimis (or trivial) risk. If there are 332 million people in the United States, however—and assuming (a big assumption) that each person faces the maximum allowable risk—we would expect any particular carcinogen regulated at the more stringent end of the range to cause 332 excess cancers within the country. And for a one-in-ten thousand risk, the number would be 33,200 excess cancers. Added to that would be risks from other hazardous air pollutants, including “synergistic risks,” which occur when combinations of chemicals pose risks greater than the sum of the risks posed by each individual chemical.

Another CAA regulatory program is New Source Performance Standards, based on “the Best System of Emission Reduction,” considering cost, any non-air quality impacts, and energy requirements. That program has served two primary purposes. NSPS standards for sources that are also regulated under the criteria program set a floor beneath which other technology-based standards (specifically, Best Available Control Technology and Lowest Achievable Emission Rate, which apply to new major sources or significantly modified major sources of criteria pollutants) cannot lawfully drop. Also, the NSPS program acts as a gap-filler, allowing EPA to regulate new and existing sources that fall at least partially between the cracks in other CAA programs, for example regulation of methane emissions from oil-and-gas production facilities.

EPA’s air quality regulatory programs are hobbled in their ability to protect public health by agency decisions to base regulatory actions on emission estimates rather than on actual emission measurements. In fact, more than five decades after the act was passed, EPA continues to designate areas of the country as “unclassifiable,” meaning that the agency believes it is unable to determine whether health protection standards are met “on the basis of available information.” In her book Next Generation Compliance, veteran environmental regulator Cynthia Giles notes, “When EPA monitors actual emissions, it finds that pollution is much worse than is being reported. . . . The net effect: a lot more pollution [goes] into the surrounding communities than . . . revealed by the estimated emissions.” Here is one example: When EPA set its action level for benzene ambient concentrations at the fencelines of oil refineries, the agency selected a concentration that no refinery would exceed as long as its estimates of fugitive emissions were accurate. Nonetheless, in 2018-2019, benzene concentrations in air around 10 oil refineries blew that limit. The offending refineries include operations by major players such as Chevron, Shell, Marathon, Valero, and BPF Energy. The result provides some insight into the efficacy of using unverified industry estimates of emissions as a basis for protecting public health.

In a 2008 case, NRDC v. EPA, the D.C. Circuit upheld the agency’s 2006 analysis of risks from the synthetic organic chemical industry’s hazardous air pollutant emissions. EPA relied on the seven-year-old results of an American Chemistry Council questionnaire, filled out by member companies, with a 44 percent response rate. In other words, EPA based its analysis that was supposed to provide the public with an ample margin of safety on old industry estimates from a minority of the regulated community. Why? The agency explained that, among other things, reliance on “industry sources is a well-established practice” and it would have been “very costly and time-consuming” for the agency to require collection and submission of data. Upholding EPA’s decision, the court explained, “We generally defer to an agency’s decision to proceed on the basis of imperfect scientific information, rather than to invest the resources to conduct the perfect study.”

Recently, the agency has sought to grapple with the excess emissions caused by “startup, shutdown, and malfunction” events, which can dwarf the emissions that are subject to permit limits. Similarly, the agency is revisiting exemptions for “fugitive emissions”; that is, emissions that “could not reasonably pass through a stack, chimney, vent, or similar opening.” Both of these emission categories are significant contributors to the air pollution problem that are not adequately accounted for in the current permitting regime.

Regulation of odors could help mitigate these deficiencies and focus EPA and state agencies on avoiding local impacts. These often fall on lower-income and minority communities. The problem has been amply demonstrated. The 2002 Third Circuit case of South Camden Citizens in Action v. New Jersey Department of Environmental Protection concerned an air emissions permit issued for a new facility in a neighborhood that already had “two Superfund sites, several contaminated and abandoned industrial sites, and many currently operating facilities, including chemical companies, waste facilities, food processing companies, automotive shops, and a petroleum coke transfer station,” not to mention permits that had been issued for “operation of a regional sewage treatment plant, a trash-to-steam incinerator, and a co-generation power plant.” The neighborhood comprised 63 percent African American, 28.3 percent Hispanic, and 9 percent white residents.

Occasionally, one hears the argument that residents of fenceline communities must have decided, voluntarily, to move close to pollution sources. Of course, Congress intended the Clean Air Act to result in healthful ambient air regardless of a neighborhood’s location. In addition, however, the assumption that residents of overburdened neighborhoods are volunteers is flawed. A 2023 environmental justice lawsuit, Inclusive Louisiana, et al. v. St. James Parish, was based on allegations that, in 2014, St. James Parish adopted a land use plan that steered industry to the parish’s 4th and 5th Districts, which are majority African American, by designating large tracts of property in those districts as “future industrial.” The lawsuit also alleged that the 2014 plan created industrial buffer zones for the parish’s white-majority churches but not Black-majority churches. (In Louisiana, a “parish” is the equivalent of a county.) The court dismissed the case on statute-of-limitations grounds but held that it “cannot say that [the plaintiffs’] claims lack a basis in fact.” Especially given the lack of specific environmental justice legislation, regulation of odors would provide a tool for mitigating the effects of such disproportionate environmental impacts.

Is it practical for EPA to set a welfare standard for odors? It is. One approach to odor regulation is the narrative standard. For example, the Virginia Administrative Code’s Standard for Odor reads, “No owner or other person shall cause or permit to be discharged into the atmosphere from any affected facility any emissions which cause an odor objectionable to individuals of ordinary sensibility.” Narrative standards are relatively common in Clean Water Act regulation. The U.S. District Court for the Northern District of Florida explained the concept in its 2012 opinion in Florida Wildlife Federation v. Jackson by way of analogy: “a state could adopt a numeric speed limit—70 miles per hour—or a narrative standard—don’t drive too fast. Or a state could adopt a combination of both—don’t drive over 70, and don’t drive too fast for conditions.”

In contrast, Colorado’s odor emission regulation takes a numeric approach. The idea is to determine how many times a sample must be diluted before becoming undetectable: “For areas used predominantly for residential or commercial purposes it is a violation if odors are detected after the odorous air has been diluted with seven (7) or more volumes of odor free air.” The approach is analogous to EPA Method 9 opacity readings, which require trained smoke readers to evaluate emission plumes. Similarly, most odor tests still require human noses, but those noses are deployed systematically, with the aid of a mechanical dilution device such as an olfactometer. The Bay Area Air Quality Management District’s “Odorous Substances” regulation provides, “a diluted sample shall be deemed odorous if during evaluation . . . at least two of the [three] subjects gave negative responses to at least 8 of the 10 odor-free or ‘blank’ presentations and affirmative responses to at least 8 of the 10 sample presentations.” The Bay Area regulation applies after 10 or more people complain to the Air Pollution Control Officer within a 90-day period “until such time as no citizen complaints have been received by the APCO for one year.”

Another approach is to develop ambient concentration standards for specific chemicals, derived from odor measurements. In a 2015 paper, “Approaches to Derive Odor-Based Values,” the Texas Commission on Environmental Quality’s Toxicology Division discusses and provides a list of publications that describe standardized methods for odor measurement. The report highlights odor threshold studies from a 1992 EPA study and a 1989 American Industrial Hygiene Association report. Under this approach, for example, a 1-hour average ambient reading of 30 parts per billion by volume of hydrogen sulfide could be a violation of an odor regulation. Of course—under an EPA welfare standard for odors—no individual source would necessarily be in violation. Instead, the state’s attainment status for the secondary odor standard would be at risk and the state, ideally, would take action to remedy the problem.

For some odors associated with some industries, surrogate odor values may be useful. Moreover—as with all environmental monitoring—technology is constantly improving and the ability to conduct widespread ambient monitoring with “electronic nose” technology may not be far off.

There are no panaceas. Just as all Clean Air Act programs suffer from gaps and mistakes, regulation of odors under the criteria program would pose frustrations. As a secondary, welfare standard, an ambient NAAQS for odors would lack a statutory attainment deadline; instead, attainment would be required on “the date by which attainment can be achieved as expeditiously as practicable after . . . such area [is] designated nonattainment.” This lack of a specific attainment deadline, of course, makes ambient regulation of odors more practical in the short-run and less scary for EPA. It also opens the door for serious foot-dragging that will require political will to overcome. But at least it would provide a handle that would allow reasonably enlightened regulators to insist on fenceline improvements.

Part of the genius of the Clean Air Act is that its overlapping programs can help make up for the shortcomings of individual programs. For example, hazardous air pollution standards result in reduction of ozone precursors that states might never have summoned the political will to mandate in discretionary parts of their state implementation plans. The ozone ambient standard helps limit volatile organic compounds (as ozone precursors) that might otherwise have only been subject to emission standards based on estimates of “acceptable” risk from individual chemicals.

Congress’s failure to enact environmental justice legislation creates a large gap in protection for some of our most vulnerable communities—those near the fencelines of industrial facilities. Regulation of ambient odors as a criteria pollutant would be one step toward addressing that gap. And with practice, and improving technology, an ambient regulation for odors could represent significant progress in realizing the Clean Air Act’s vision: “to protect and enhance the quality of the nation’s air resources so as to promote the public health and welfare and the productive capacity of its population.”