4.1. Airline Reporting Rules for Onboard Fumes/Smoke
In the U.S., commercial airlines are required to report each “failure, malfunction, or defect” that causes “smoke, vapor, toxic or noxious fumes” to accumulate or circulate in the flight deck or cabin during flight (defined as “wheels up”) (14 CFR § 121.703(a)(5)). Airlines are also required to report ground-based events if the airline is of the “opinion” that flight safety could have been “endangered” had the aircraft continued (14 CFR § 121.703(c)). There are comparable SDR rules for charter flights, too (14 CFR § 135.415). These “Service Difficulty Reporting” (SDR) rules will not capture fume events for which the causal defect is not apparent during a ground-based inspection. For example, oil fumes that contaminate the bleed air through a worn engine part subjected to thermal or physical stress during an engine power setting change inflight need not be reported if that same part does not leak under ground-based conditions during a maintenance inspection. As well, when maintenance workers overfill the oil or hydraulic fluid reservoir, the subsequent spillage into the bleed air is a recognized source of fumes [24
], but spillage is not a mechanical defect, so does not need to be reported.
In this dataset, 9.1% (n
= 1135) of the SDR reports from airlines to the FAA do not offer insight into the relevant mechanical defect or failure. It is not clear if the fault is not described or not known when the report is submitted. Technically, though, it is the defect
that is reportable, not the presence of fumes. Even though, there may not be an actual defect and, even when there is, maintenance troubleshooting can be “trial and error” [27
], which can delay defect identification.
In addition to the SDR regulation for reporting onboard fumes and smoke, U.S. airlines must also report each “interruption to a scheduled flight” caused by a known or suspected (emphasis added) mechanical difficulty or malfunction, which is not required to be reported under the SDR rule (14 CFR § 121.705). However, it is difficult to assess the degree to which airlines comply with this rule because there is no central, searchable FAA database; rather, these “mechanical interruption reports” are maintained by individual FAA Certificate Management Offices, and only for one year.
Finally, U.S. airlines must also report “occurrences which affect or could affect safety of operations”, including onboard fumes/smoke, to the FAA’s “Accident and Incident Data System” (AIDS) per Order 8020.11D (2018). The author’s 2014 FOIA request included AIDS reports with one or more of the same search terms requested for the SDR data, but U.S. airlines only submitted 2.3% as many fume/smoke reports to meet Order 8020.11D as compared to the SDR regulation (n = 366 compared to 15,885) during the same time period. It would be reasonable to assume that AIDS reports are the more serious events during which safety of operations was (or could have been) compromised. To test this hypothesis, the author compared the contents in the narrative fields of the AIDS and SDR datasets (which were selected based on the same set of search terms during the same time frame) and concluded that the nature and seriousness of the reports is comparable. It appears that the AIDS regulation is either ignored or misunderstood, and that compliance is not enforced.
In summary, then, if airline staff do not identify a mechanical source of reported fumes, or if the fumes occur before take-off, then the event is not reportable. If airline staff do identify a mechanical source of fumes, then the airline must report to the SDR database. If a mechanical source was suspected and the flight plan was affected, then the airline must report to the “mechanical interruption” database. In both of those scenarios, if safety either was or could have been compromised, then the airline must also report to the AIDS system. However, ground-based fume events can still compromise safety, fumes can occur without obvious mechanical defects, and an airline is unlikely to report the same event to two reporting systems. As a result, these reports of fumes and smoke are only a subset of the actual events, and the data that do exist get scattered, all making an accurate assessment of the relative and absolute frequencies of different sources of onboard fumes and smoke even more difficult.
In addition to the complying with FAA reporting rules, U.S. airlines must report fume and smoke events to the U.S. National Transportation Safety Board (NTSB), if the event meets the NTSB definition of either an “accident” or “serious incident” (49 CFR § 830.5, 49 CFR § 830.2). Some fume events meet the NTSB reporting criteria, such as when a crewmember is hospitalized for more than 48 h during the seven days post-flight, sustains internal organ damage, or is unable to perform normal flight duties.
Another source of fume event data for the U.S. fleet is the “Aviation Safety Reporting System” (ASRS) database, hosted by the National Aeronautics and Space Administration (NASA). ASRS catalogues aircraft incident reports for a broad range of unsafe conditions from pilots, cabin crew, maintenance workers, air traffic controllers, and dispatchers. Reporting is voluntary and confidential, with the goal of circulating information about safety issues within the industry to prevent reoccurrences. Of 341 ASRS “smoke/fire/fumes/odor” events that crewmembers reported in 2016–2017, 35% were sourced to the air supply system and 20% were sourced to electrical faults [28
4.2. Underreporting Engine Oil and Hydraulic Fume Events
Unfortunately, even within the narrowed scope of the reporting rules for fume and smoke events, there is evidence of airline underreporting. Some of the underreporting is cited for fume events in general. For example, in 2006, the FAA issued a bulletin for its maintenance inspectors, acknowledging poor airline compliance with the fume event reporting regulations. Specifically, the agency noted that “there have been concerns raised about numerous reports of [fume] events on commercial air carrier/operator aircraft. During the FAA’s analysis of this [sic] data, it appears as though there are numerous [airlines] who may not have reported these events as required by regulation” [21
There is evidence, however, that some underreporting is specific to oil and hydraulic fume events, reducing both the number of reports and the relative frequency of those events. For example, a review of fume events consistent with oil and hydraulic fluid sources at one U.S. airline over 18 months in 2009–2010 found that the airline had only reported 38% of 87 fume events documented by crewmembers to the FAA SDR system [11
]. Part of this discrepancy is explained by the airline not consistently complying with the reporting rules, but part of it is explained by regulatory language that does not require airlines to report a significant fraction of documented events, as described above.
An example of underreporting specific to engine oil fumes is a 2013 report that the FAA sent to two Congressional committees regarding a law requiring the FAA to conduct research and development into sensor and filtration technology, specific to preventing exposure to oil fumes in the bleed air on commercial aircraft [29
]. The FAA reported to Congress that officials had “mined” their fume event databases from 2002–2011, inclusive, for reports that included one or more of the search terms “fume, odor, smell, smoke, and bleed air”, and that U.S. airlines had only reported 69 oil fume events and no hydraulic fluid events during those 10 years [23
]. The FAA told Congress that, since the occurrence of these events was “extremely low”, it could not justify spending its resources to research and develop bleed air sensors and filters. However, this analysis of those same cited reports from just one of the FAA databases using the same search terms identified oil as a source of fumes in 1353 reports, hydraulic fluid as a source in 173 reports, and bleed air contaminants consistent with oil/hydraulic fluid in 1667 reports. This type of underreporting reduces the relative frequency of these reports and downplays the need for relevant regulatory control measures.
Finally, underreporting oil and hydraulic fluid fume events is not limited to U.S. airlines. In 2011, the European Aviation Safety Agency (EASA) published comments it had received in response to the question of whether a rulemaking to address oil fumes in the cabin and flight would be justified. EASA noted multiple claims by advocates for pilots, cabin crew, and passengers that European airlines underreport (a word used 53 times in the report) fume events and that the safety impacts are downplayed [30
]. Likewise, a report by the Australian Transport Safety Bureau (ATSB) noted evidence that Australian airlines underreport fume events, both to the ATSB and to the aviation regulator [31
]. An Australian Senate Enquiry into fume events on one particularly problematic aircraft type noted “strong evidence of a tendency of pilots to under-report [fume events]”, citing a combination of pressure from airline management and lack of training [32
]. The International Civil Aviation Organization (ICAO) has formally recommended that airlines “encourage flight and cabin crew members to report fume events”, referring to crew reports as “essential to assist airline maintenance technicians in identifying the root cause of an event and taking corrective action to prevent reoccurrence” [14
]. Currently, though, fume event reporting to aviation regulators is largely limited to maintenance defect reporting.
4.3. Estimates of the Frequency of Engine Oil and Hydraulic Events
The answer to the question of how often oil and hydraulic fluid fume events happen during commercial airline flights is a subject of considerable debate within the industry. It is important to get an accurate answer because, as the FAA’s 2013 report to Congress illustrates [23
], the answer will influence the regulatory agenda. The FAA itself acknowledges this reality in a 2016 bulletin in which the agency notified airlines “that some data required by [the SDR rule] is not being entered properly” and that “[b]ecause of poor data integrity, the FAA, manufacturers, and air carriers are unable to accurately detect trends necessary to proactively mitigate risk” [22
One of the first estimates of how often oil and hydraulic fluid fume events occur was published in a 2002 U.S. National Research Council (NRC) committee report about aspects of aircraft air quality associated with potentially negative impacts on crew and passenger health. The report referenced fume event data collected at three airlines “over several years”, which ranged from 0.9 events per 10,000 flights on Boeing 737 aircraft up to 39 events per 10,000 flights for the British Aerospace 146 [33
]. The NRC committee issued a series of recommendations to the FAA, including that the agency “rigorously demonstrate” the “adequacy” of its cabin air regulations. In response, the FAA acknowledged that its “rulemaking has not kept pace with public expectation and concern about air quality and…[n]o present airplane design…incorporates an air contaminant monitoring system to ensure that the air provided to the occupants is free of hazardous contaminants…” [34
]. As of this writing, the FAA has still not initiated a rulemaking on this subject.
In 2010, the FAA published an informational bulletin for airlines, stating that the agency “…continues to receive over 900 reports a year on [all sources of] smoke or fumes in the cabin and or cockpit” [35
]. Based on the number of flights in 2010 [36
], this translates into 0.9 events per 10,000 flights. The FAA noted that, “in fact, it is not unusual to receive more than one report during a 24-h period. For instance, on one day in April of 2010, five reports of smoke in the cockpit came in from one [airline]. All these incidents prompted the flightcrew to declare emergencies and divert to the nearest airport” [35
In 2013, the International Air Transport Association (IATA) reported on its analysis of operational safety reports from more than 150 airlines covering almost 31 million flights during the period 2008–2012 [37
]. Of these, 3444 reports were categorized as “smoke, fumes, and odor”, which is equivalent to about one reported event per 10,000 flight cycles. Engine-related fumes were the most commonly reported source (23%), followed by electrical-related fumes (15%). Maintenance action was the most common operational effect (25%).
In 2015, FAA-funded researchers published their review of engine oil and hydraulic fluid fume events identified in FAA, NTSB, and NASA databases from 2007–2012 and sorted them by aircraft type. They concluded that U.S. airlines reported fume events on almost every aircraft type at an average frequency of two events per 10,000 flights [38
], which translates into an average of 5.3 events per day based on the number of flights operated during that time [36
]. The authors reported “substantial variation” in the frequency of reports by aircraft type, with the maximum being eight reported events per 10,000 flights.
4.4. Recommendations to Improve Data Collection for Onboard Fumes/Smoke
Beyond the need for better compliance and enforcement of existing regulations, this review of SDR data from 2002–2011 provides insight into how fume reporting regulations could be improved.
First, SDR data would be more representative if airlines were required to report the presence of onboard fumes/smoke during ground operations, not just inflight. Industry sources confirm that the APU (which is the typical bleed air source during ground operations, whether at or away from the gate) is a source of oil and hydraulic fluid fumes [25
]. Reports confirm that APU oil fumes, which contaminate the cabin or flight deck air during ground operations, can generate reports of crew ill health [11
]. Additionally, if fumes that manifest on the ground are not resolved, they can continue inflight [25
], which can affect both the safety of operations [26
] and the health of the crew and passengers [11
]. For these reasons, ground-based fume events should be documented and prevented.
Second, reported onboard fumes/smoke for which there is no obvious mechanical fault should still be reportable. For example, overfilling an oil or hydraulic fluid reservoir is a recognized source of fumes [26
]. Additionally, some faults (such as oil leakage through a worn but not defective engine seal) that introduce fumes to the cabin air may only happen during the stress of a high-engine power setting, for example, and may not be recreated during routine ground-based troubleshooting [40
]. Maintenance staff may not identify the relevant mechanical failure even for serious events, such as those involving pilot impairment and incapacitation [6
]. SDR data would be more representative and useful if the presence of fumes/smoke caused by even a suspected defect or failure (defined to include spills and leaks) was reportable.
Third, it would be easier for SDR database users to answer the “how often” and “what type” questions regarding onboard fumes/smoke if the FAA’s report categories were more tailored to the actual sources of fumes (see Figure 1
), or if airlines could choose more than one category. For example, a pull-down menu for airlines to select standardized responses to the question of “source” would assist with subsequent analyses. The narrative field for each report can provide useful details but is too technical to be interpreted by software. Currently, when an airline submits a report, they must choose the most relevant category (“nature condition”), as defined by the FAA. One of the FAA categories is “smoke/fumes/odors/sparks”, but only 76% of the 3360 events the author classified as “oil”, “hydraulic”, or “bleed-sourced” were categorized as such. U.S. airlines categorized another 16% of those bleed air reports as “fluid loss” and assigned the remaining 8% to one of 13 additional categories. For the full dataset (n
= 12,417, all sources), airlines classified 84% of the reports as “smoke/fumes/odor/sparks” and 5.4% as fluid loss. In both cases, selecting only the “smoke/fumes/odor/sparks” category to search for fume and smoke reports seems reasonable but would underestimate the number of relevant reports.
Lastly, a crewmember reporting system for fume events to supplement the existing maintenance reporting systems is needed and must be accompanied by suitable training programs to ensure that crewmembers know how to recognize and respond to different types of onboard fumes and smoke [14
]. This point has been raised since at least 2002, starting with an NRC committee, which recommended that the FAA establish a crewmember reporting system to collect air quality event data and associated symptom reports [33
]. The following year, Congress passed a law requiring the FAA “to establish an air quality incident reporting system” [41
]. Nine years later, Congress again required the FAA to “establish a systematic reporting standard for smoke and fume events in aircraft cabins” [42
]. In 2017, the Cabin Air Safety Act of 2017 was introduced to the floor of the U.S. Senate, and the Know Before You Fly Act was introduced to the floor of the U.S. House. Again, these bills called on the FAA to establish a system for crewmembers and mechanics to report fume events on commercial aircraft, but neither bill was passed. In early 2018, the FAA recommended that airlines “assess current policy and procedures” regarding their procedures to differentiate and mitigate smoke and fumes, but the FAA did not mandate event reporting [43
]. Most recently, in October 2018, Congress mandated that the FAA “issue guidance for airline crews and maintenance technicians to report incidents of smoke or fumes on board an aircraft operated by a commercial air carrier” [44
]. In March 2021, the FAA launched a cabin air quality website that includes a link to a voluntary, deidentified reporting system operated by NASA and a link to the FAA’s maintenance defect reporting system [45
]. However, the FAA has still not implemented crewmember reporting systems for fume events on commercial aircraft.