occupancy density” when compared to other public transit vehicles (Tawfik et al. 2022, 4). For example, Shen et al. (2020) reported on a COVID-19 outbreak on that most likely resulted from exposure to a single infected individual on a bus. Twenty-four of 68 passengers (35.3%) contracted the virus, with the “recirculated air environment of the bus” being the most likely explanation for transmission (Van Dyke et al. 2022, 2). In general, bus passengers have the highest risk of exposure when they are seated close to a source (e.g., an infected passenger), and when they are in the path of airflow from the source to an outlet, such as a window or air vent (Castellini Jr et al. 2022). Increased ventilation and air filtration within the bus cabin has been shown to decrease infection risk (Kim et al. 2023) in three ways: First, it reduces the airborne presence of viral particles. Second, limiting the presence of airborne particles decreases the number of particles deposited on surfaces that can lead to fomite transmission. Third, eliminating airborne particles further reduces risks that even careful individuals will become infected due to gaps in the use of PPE.

Natural Ventilation

Using fresh air ventilation to dilute air contaminants is a common non-pharmaceutical intervention (NPI) in enclosed environments, such as occupied buildings (Van Dyke et al. 2022), and in public transportation (Luo et al. 2023). Buses are often configured with both active (e.g., mechanical heating and cooling) and passive (e.g., windows, ceiling hatches) ventilation systems. Van Dyke et al. (2022) noted such systems are relatively simple in contrast to other high-capacity forms of transportation, especially commercial aircraft, which use features such as laminar airflow (directed downward from ceiling to floor) and high efficiency filtration to improve air quality.

Research has investigated whether opening windows, as a natural ventilation mechanism, can effectively reduce bus passengers’ risks of contracting COVID-19. For example, Kim et al. (2023) calculated the amount of exposure time it took for bus passengers to reach a 10% probability of infection. In a case with no air circulation, this threshold was reached in only 4.3 minutes. However, the exposure time increased to 46 minutes in a case where all windows were open on a stationary bus, and 203 minutes for a moving bus with all windows open.

Scholars have also examined whether various configurations of opened windows can provide sufficient ventilation with mixed results. Luo et al. (2023) recommended opening the front and rear windows, noting that “the indoor air will flow like a ‘pump’” specifically “when windows are opened at locations with a large pressure difference between indoor and outdoor” (7). However, increasing natural ventilation alone does not appear to be sufficient for removing viral droplets. Kim et al. (2023) asserted that “air diffusion and escaping speed through the window was too slow,” resulting in buses’ lower ventilation efficiency when compared to cars (5). Luo et al.’s simulation found droplets had higher concentration on the side of the bus where an infected person was seated, as well as in the rear of the bus under weak airflow conditions. Even under stronger airflow conditions, they cautioned, “we should pay attention not to sit downwind of the dominant airflow” (11). Yao and Liu (2021) likewise noted that opening windows next to an infected person showed unsatisfactory results because turbulent airflow expanded the spreading of droplets, rather than reducing them. Additionally, these studies have highlighted the inconsistency of ventilation in cases where buses are moving or stationary. Kim et al. (2023) observed that while ventilation rates were consistent across rows of seats when the bus was stopped, these rates varied when the bus was in motion. Luo et al. (2023) found that when buses are relying on natural ventilation, speed is another important factor. Specifically, the slower a bus is traveling, the lower its air change rate per hour becomes, which increases risk of passenger infection.

Buses also use mechanical ventilation, or heating and air conditioning (HVAC) systems to aid natural ventilation. Using an Heating, Ventilation, and Air Conditioning (HVAC) in indoor air circulation mode moves and maintains air within the vehicle. (Tawfik et al. 2022) and increasing passengers’ exposure to outdoor air pollution without adequate air filtration (Lee et al. 2021). The use of air filtration systems is addressed in the next section.

days. Therefore, waste PPE can become a route of indirect infection” (Choi, Choi, and Rhee 2023, 2; see also Sharma, Chakraborty, and Barman 2023).

Finally, the use of PPE as a mitigation strategy is only effective when people are willing able able to use them correctly. PPR represents an administrative control aimed at reducing a hazard’s risk by influencing human behaviors and practices (Ajselv 2022). Developing engineering controls that separate people from a hazard is a more effective strategy for risk reduction (Ajslev 2022; Pu et al. 2023).

Social Distancing (and Barriers)

Social distancing is quite literally the physical act of maintaining a certain level of distance between individuals in a social setting, when possible. In the United States, social distancing was mandated at “6 feet.” The 6 feet came from a theoretical calculation of trajectory of COVID virus-laden saliva droplet can travel horizontally before hitting the ground surface. The purpose of social distancing was to minimize “crowding,” especially in locales with nonoptimal ventilation, and because of this social distancing became “a central element of public health early in the pandemic” (Dai and Taylor 2023, 5). The other reason it became so central inside the United States is that, at least at the policy level, this is a NPI which is already culturally aligned with the grand majority of the citizenry. In the public transit environment, individuals were also instructed to enter and exit through the back of public buses vehicles, and skip rows, if possible (Dai and Taylor 2023, 5; see also Cheshmehzangi et al. 2023).

Barriers feel like a logical extension to social distancing because they seem to fill in the gap that when you cannot be socially distant. In other words, you can place a barrier to imitate or replicate social distance (Cadnum et al. 2022; Dhanak et al. 2022). However, “barriers have been found to impact the airflow [...] resulting in poor air circulation [...] where particles can accumulate and/or slow removal of them” (Gnegy-Spencer, Gohlke, and Marr 2023, 384–85). More strikingly, as Gnegy-Spencer et al. (2023) note, “barriers may result in higher exposure to people on the same side of the barrier as the infected person, as with coworkers side-by-side behind a series of barriers” (402). In fact, “results indicate that in rooms with low ventilation rates, barriers may not be adequate protection for individuals within the room, as particles have the opportunity to travel around the barrier” (Gnegy-Spencer, Gohlke, and Marr 2023, 402–3). Ahmadzadeh and Shams (2022) did find that erecting “physical barriers around the head and above the passenger seats” on a bus was an effective means “to reduce the risk of infection,” but only in conjunction with also “increasing ventilation rates, selecting medium relative humidity, and increasing exhaust capacity” (28).

These first two paragraphs in this section are dealing with social distancing and barriers as a generality. In all the modes of public transportation, the true purpose of both strategies is to protect those who will be in contact with the most people (e.g., drivers/operators), as keeping them healthy is a key priority (FTA 2022). To this end, numerous recommendations have been made by the Federal Transportation Administration (2022), such as closing off the driver’s cabin, a tactic that was common even before the pandemic. Another suggestion was to close off the seats closest to the driver. However, this often cannot be done as those seats are compliant with the Americans with Disabilities Act and frequently used by the elderly, infirmed, or pregnant. The same factors arise when it comes to having passengers enter from the rear door(s) rather than the front, but this is about lessening impact, so those who can comply should, to increase the likelihood that the driver will remain healthy and uninfected (FTA 2022).

There is an unexpressed tension that Elizabeth Bays (2023) raises regarding the ripple effect of social distancing, barrier, and masking mandates and the ramifications as they relate to constitutionality. Bays’ (2023) narrow focus is on the intersections of these mitigation strategies (specifically plexiglass barriers around witness stands and cloth masks) with the fundamental rights (plural) of a courtroom defendant to confront witnesses. Importantly, she notes the court is not meant to simply defer; rather, “when the interest the government claims rests on necessity, the court requires evidence in the record supporting that claimed necessity” (259), but what happened was the courts simply became a rubberstamp (243,256-257,262). She contends this is problematic because “the CDC guidance did not even reflect the best available science” (243). As was pointed out in the paragraph prior, it is now known that “barriers may actually increase the

risk of COVID-19 transmission,” and that “cloth masks” are merely a fashion accessory which “ ” (267–69). Thus, public transit officials must consider both what is now known about these mitigation strategies as well as how they intersect with the rights of the public in deploying them over the long-term.

Surface Cleaning

When the pandemic first broke out, it was unsurprising that surface cleaning, an easy and common NPI, was one of the first implemented to stave off the spread of the virus. Almost instantaneously, the American Bus Association (ABA 2020) published standard operating procedures and best practices when it came to disinfecting and decontaminating a bus. It was a two-schedule system which involved first a “10-point critical touch” system, which was to be done multiple times throughout the day when there were no passengers, and this was then followed by a “32-point critical touch” system, which was to happen once a day when it returned to the hub (ABA 2020, 6–9). They also recommended “strict daily sanitation procedures” that were meant for those in the direct employ of rendering transportation for the public, which involved elements like requiring masks, training on how to properly disinfect, and having hand sanitizer, amongst others (ABA 2020, 9), as humans are part of the surface of the bus that needs to be cleansed.

The primary reason these procedures were so stringent is that this is one of the most effective means at controlling “fomite transmission” of a virus, and “[t]here [was] strong evidence indicating that traditional cleaning and disinfection methods [were] not adequate for infection prevention and control” (Arunwuttipong et al. 2021, 848). The two most widely used systems to clean buses are “ultraviolet germicidal irradiation (UVGI) and the hydrogen peroxide (H2O2) system.” Of these two systems, the latter is preferred to the former because it can clean the “shadows,” and, most importantly, “it decomposes spontaneously into oxygen and water, which are non-toxic byproducts.” There are two different types of hydrogen peroxide systems: “vapor (HPV)” and “aerosolized hydrogen peroxide (aHP)” (Arunwuttipong et al. 2021, 848; see also Chen et al. 2023). Of these two, the former is preferred over the latter as “studies [indicate] that HPV systems are more effective at disinfecting,” but the aHP systems are “comparable to HPV systems.” The advantages of aHP are threefold: it can be used in “rooms that were not originally designed for vapor,” it has “lower machine and maintenance costs,” and it “can be easily scaled” (Arunwuttipong et al. 2021, 849; it should be noted that hydrogen peroxide is an irritant and can damage the lungs, so that is something that must be taken into consideration before the implementation of this practice.).

Now, of the HPV systems there are two types: dry and wet. Of these two, one does not seem to be preferred over the other, as the only major difference is that during the wet process “no vapor is emitted during dwell phase, thus permitting the peroxide to dwell on any surface exposed.” The dry system passes through four phases: “dehumidification,” “conditioning, where the vapor is released until reached desire concentration,” “decontamination,” and “aeration.” The wet process also has four phases: “conditioning, gassing, dwell, and aeration.” (Arunwuttipong et al. 2021, 848–49). This process “achieved 6-log reduction in sporicidal efficiency, which is equivalent to sterilization,” and it took only “100 min” to complete the process, but it should be noted that this “decontamination procedure [...] is a supplement and cannot be used in place of cleaning” (Arunwuttipong et al. 2021, 853).

It is generally agreed upon that “effective disinfection is essential [...] during the pandemic,” but “the surging use of disinfectants and antiseptics poses risk of [...] pollution, ecological risks [...] antibiotic resistance, and biodiversity loss” (Chen et al. 2023, 24737), so caution is imperative. Importantly, it has now been demonstrated that “exposure by aerosol inhalation is 1000 times higher than by contact of contaminated surface,” and that the mishandling and overuse of “disinfectants [are] believed to be associated with misinformation, unnecessary panic, and anti-rationalism [...] as well as bureaucracy and wrong decision-making of governing agencies” (Chen et al. 2023, 24737–38) especially the World Health Organization. As noted earlier, the primary means to clean buses, hydrogen peroxide, is environmentally friendly, but many other means of mass disposal are not, such as incineration, microwaves, dry heat, and chemicals (Chen et al. 2023; Dehghani et al. 2023; Sharma, Chakraborty, and Barman 2023), which is why

a technique such as “plasma pyrolysis” is suggested as it is “an environmentally beneficial technique that converts organic waste into economically useful byproducts” (Sharma, Chakraborty, and Barman 2023, 70; see also Dehghani et al. 2023).

Non-Ridership

It is important to begin with the fact that even prior to COVID-19, public transportation has been in a continual decline “for six consecutive years” (Ammoury et al. 2023, 2; see also Owen, Arribas-Bel, and Rowe 2023; Wilbur et al. 2023). The pandemic has driven “ridership in the United States [to] a 100 year low” (Ziedan, Lima, and Brakewood 2023, 2; see also Dai and Taylor 2023). As of June 2023, “ridership has [only] recovered to more than 70 percent of pre-pandemic levels,” and bus ridership has recovered more than rail because of “work-from-home and downtown occupancy trends” (APTA 2023). This failure to recover has led experts to speculate that the “fiscal crisis” has the potential to become a “death spiral” (Davis 2023; see also APTA 2023), and this fear is predicated on the “stabilisation of these [non-ridership] patterns long after the end of the most-constraining health protection measures” (UITP 2023, 3), but they also found positive potential as “the better distribution of urban flows might enable the operation of more-efficient networks, with reduced investment requirements on fleet and staff to cover peak hours” (UITP 2023, 5; see also Ammoury et al. 2023; Borjigin, He, and Niemeier 2023; Ziedan, Lima, and Brakewood 2023).

By contrast, both New York and California were able to resolve their financial crisis without further assistance from the federal government who has been keeping the public transportation systems in most states running and relatively stable (Davis 2023). That New York was able to do this is of prime importance, as their system, in 2019, “hosted 64 percent of all U.S. mass transit trips (3.79 billion UPT out of 5.97 billion nationally).” That is “more than Chicago CTA, L.A. Metro, Boston MBTA, WMATA, SEPTA, Jersey Transit, Atlanta MARTA, Denver RTD, San Fran BART and Muni and Portland Tri-Meet combined, and that still leaves the MTA 1 billion trips ahead” (Davis 2023). What this demonstrates is that there is a way out of this financial impasse, and if the largest system – larger usually implies inflexibility – was able to do this, other urban hubs might have more outs than first thought. It also informs us of how massive bus ridership is, and how integral it is to keep society moving.

The primary reason that people stopped riding the bus, “two-thirds,” related to the “direct impacts of the spread of COVID-19” (Ziedan, Lima, and Brakewood 2023, 15). To be even more specific, it was the “executive orders” (Osorio, Liu, and Ouyang 2022, 17; see also Dai and Taylor 2023) which caused this abrupt and drastic halt. The other three major factors, accounting for “13% to 38% of the total decline in bus ridership” were “employment, telework, and population relocation” (Ziedan, Lima, and Brakewood 2023, 15). These shifts were external and indirect but affected by COVID-19. Paul and Taylor (2022) noted that the “robust regional effects on our models imply underlying metro-specific effects that our suite of independent variables did not capture” (25; see also Osorio, Liu, and Ouyang 2022). As an example, Chicago has three major public transportation hubs: CTA, Metra, and Pace, each of these can be conceived of as regions. The CTA “provides local bus and heavy rail transit in the City,” Metra “is a commuter rail service whose lines run well into the periphery of the metro area and into bordering states to the Central business District,” and Pace “is the suburban bus service which operates primarily outside of the City”(Soria, Edward, and Stathopoulos 2023, 141).

And as discussed earlier, bus ridership has begun to return at a clip significantly higher than rail, and it now becomes clearer that region might be meant polysemically, as the insidious underbelly of the role race and ethnicity has played in COVID-19 is exposed. This discrepancy “between social groups subsequently exacerbated existing health inequalities” because they were “deemed ‘essential’” and thusly “increasingly exposed to COVID-19 by being less able to adhere to [NPIs] such as working from home [...] and performing social distancing” (Owen, Arribas-Bel, and Rowe 2023, 1; Osorio, Liu, and Ouyang 2022; Paul and Taylor 2022; Soria, Edward, and Stathopoulos 2023). The ability to mitigate is a privilege “reserved only for the well-off” (Owen, Arribas-Bel, and Rowe 2023, 1). Equally troubling still is the role that gender

seems to be playing. Not only were women more likely to stop riding the bus and “less likely to return to pre-pandemic ridership,” they were also less likely to have “access to a household vehicle,” thereby leaving “women [as] particularly vulnerable to reduced access to public transportation” (Soria, Edward, and Stathopoulos 2023, 152).

Conclusion

Throughout the pandemic, public transit authorities have deployed a range of NPIs on buses to reduce COVID-19 infection risks for operators and passengers. Although many of these mitigation methods have proven to be effective in containing and reducing transmission risks, they also have other implications, such as the environmental impacts of disposable facemasks; the ecological and biological impacts of surface cleaning; the social acceptance of distancing and barriers; and the economic implications of non-ridership. Additionally, many of these strategies relied on administrative controls, which rely on the willingness and ability of individuals to modify their behaviors in the transit environment. Baselga et al. (2023) noted, “it is necessary to find long-term solutions to improve the quality of the air we breathe” in the public transit environment (8). Designing new engineering controls, which separate people from potential hazards, is essential for improved risk reduction.

Documented Exposure in Airplanes, Buses, and Trains

This section covers documented COVID-19 outbreaks in the air, buses, and rail. Public transit and public transportation were found to be stand-in terms for buses and rail, so those terms are also covered. Additionally, both outbreaks and contact tracing are defined, and there are sections regarding the data and then outbreaks proper at three levels. Finally, the last section will be the conclusion.

Outbreaks

An outbreak is “two to five or more confirmed cases of COVID-19 linked to a common location” (National Governors Association 2021, 2-3). This is the definition used by all the states’ COVID-19 dashboards, as well as all the articles except one (Luckhaupt et al. 2023, 338). This exception used a more rigid designation, “workplace-specific COVID-19 outbreak,” which is defined by the Council of State and Territorial Epidemiologists (National Governors Association 2021, 2).

Contact Tracing

Contact tracing is a “non-pharmaceutical intervention […] to control COVID-19 [and] is the process of identifying and obtaining information from individuals who have been in long enough contact with other infected individuals” (Pozo-Martin 2023, 244). This is the primary means by which data points become linked, especially between outbreaks and locations. Nearly every state participated in the collection of congregate settings data which whose vectors reside at the intersection between outbreaks and enclosed locations. The three major vectors where data was collected were corrections, healthcare, and schools, presumably because of convenience.

Contact tracing in the United States was politicized nationally, and while guidance was given by the Centers for Disease Control and Prevention (CDC), amongst other federal governmental agencies, there were no regulatory protocols in place to encourage alignment or enforcement at the state-level. Therefore, the overall approach to contact tracing in the United States was slipshod. In fact, according to Henry Bayly and colleagues (2023) “contact tracing protocols in the U.S. are unlikely to have identified more than 1.65% […] of transmission events with PCR testing and 0.88% […] with rapid antigen testing.” International data, where contact tracing was more robust, suggests the majority of COVID-19 outbreaks were linked to social establishments where people came in close contact (e.g., gyms, bars, or live music venues) and not public transit (Joselow, 2020). However, international data does establish some risk associations for transit workers. Research in Europe “identified elevated mortality risk among public transportation workers; taxi

and bus drivers were found to have the highest COVID-19 mortality rates among all occupational groups” (Heinzerling et al., 2022, 1055).

Data

The data regarding outbreaks on airplanes, buses, and trains in the United States is severely limited, as most states were collecting and collating data on a case basis, on a death-basis, and on a demographic-basis. Moreover, public transportation usage declined amid the pandemic due to “a reduced service supply” and “perception [that] transportation is riskier than private or personal means of transport because of the closer contact to other people” (Tirachini & Cats, 2020, 2). Ridership fluctuations may decrease the accuracy of assessing population-based risks for public transportation usage. Data was found on three levels: Microlevel (City/County), Mesolevel (State), and the Macrolevel (National) and was found by using search engines, both general and specialized, and by looking at all the states’ COVID-19 dashboards and datasets. In addition to population-based data, some case studies have documented specific COVID-19 outbreaks linked to public transportation.

Microlevel Outbreaks (City/County)

At the microlevel, there is an article and a dataset. The article, by Tomasi et al. (2021), focused solely on the mortality of bus union members in New York City, New York. Their total sample set was 118, and was painstakingly gathered through media scanning, which was their workaround for not being able to have access to the data they needed – they generated it themselves. The process was successful, in that a dataset was generated, but a thorough cost-benefit analysis demonstrates that this method is not cost-effective nor is it altogether clear that they were able to capture exactly what they wanted.

The dataset was gathered by the authors from the South Nevada Health District and concerns Clark County, Nevada from 01-06-20 until 05-22-23. During that time, they had a total of 137,084 cases, 3,880 which happened in the last 30 days.

This data was aggregated into three clusters: Air, Bus/Train, and Transit. Air had a total of 2,581 cases, Bus/Train had 281 cases, and Transit had eight cases, and over the last 30 days those numbers were 140, 16, and 0 cases. In relation to total cases, these three clusters equate to 2.1%, and 4.0% of the last 30 days. Mortality data was not provided, as this dataset dealt only with exposure. Additionally, this only shows individual cases, not outbreaks, but it at least grants some insight.

Mesolevel Outbreaks (State)

At the mesolevel, there is an article and three datasets. The article, Heinzerling et al. (2022), cites Tomasi et al. (2021), and has a similar focus on mortality, but they also enumerated the number of outbreaks (340), number of cases (5,641), and number of deaths (537) that occurred between 01-01-2020 and 05-26-2022 in the state of California (Heinzerling et al. 2022, 1052). What they discovered is that “[t]he largest number of outbreaks […] occurred in the bus and urban transit workplaces,” “the largest number of outbreak-associated cases occurred in air transportation,” and “the largest number of deaths […] occurred among workers in transportation support services” (Heinzerling et al. 2022 1053). In relation, to the overall totals of outbreaks, cases, and deaths – all public transportation services consisted of 1.0% of all outbreaks, 1.0% of all cases, and 3.0% of all deaths, so they were right, with Tomasi et al. (2021), to be concerned with this particular sector.

The first dataset was gathered by the authors from the California Department of Public Health and concerns the state of California from 01-01-21 until 05-05-23. During that time, they had a total of 48,006 outbreaks and 585,784 cases.

These data are aligned with the work of Heinzerling et al. (2022), as the most outbreaks occurred on the Bus, and the most cases occurred in the Air. In relation to the total number of outbreaks and cases, Air/Rail/Bus make up 0.0 % of all outbreaks and 5.0% of all cases, so this dataset aligns with Tomasi et al. (2021) and Heinzerling et al. (2022) regarding its significance.

particles are removed within several seat rows from a source” and recirculated air passes through high-efficiency filtration (HEPA) systems, which are highly efficient in removing contaminants and viruses (Bennett et al. 2022, 1). Unfortunately, it is still possible for viral transmission to occur before contaminated air is recirculated (Wang et al. 2023). This risk, particularly when coupled with high passenger density (such as in economy class) and length of flight time, can influence infection probability (Wang et al. 2023).

Because COVID-19 and other viruses are primarily transferred with airflow, researchers have focused evaluating the design of cabins and ventilation systems to identify air distribution characteristics and infection risks. Wang et al. (2023) compared COVID-19 infections probability in commercial airplanes of different sizes and seat configurations during a four-hour flight. Average infection probabilities were similar across the varying aircraft, with the exception of a 3-3-3 seat configuration. Wang et al. (2023) noted “the airflow pattern in this configuration could bring the virus to the lower part of the cabin and exhaust it through the outlet, or could restrict the dispersion of the virus” (10).

Effectiveness

The two primary methods used for studying issues related to COVID-19 transmission on airplanes are experimental measurements and computer simulations. Wang et al. (2022) noted that although experimental measurements “can provide realistic and reliable information,” they can also be “prohibitively expensive to build mockups that represent various aircraft models and to produce the distribution of full-size particles” (2). CFD simulations, by contrast, offer a more inexpensive and flexible method for studying ventilation, airflow, and viral transmission as boundary conditions can be easily changed within the programs.

It is not surprising, then, that CFD was by far the most common methodology used in the literature reviewed for this project. Researchers used CFD to compare the infection probability of COVID-19 in differently-sized commercial airplanes (Wang et al. 2023), analyze the airflow and ventilation rates of different kinds of ventilation systems (Bilir et al. 2022), explore the impact of different flight intervals (e.g., takeoff, climb, cruise, descent) on ventilation and contaminant dispersion (Elmghraby et al. 2018), and simulate transmission of COVID-19 under various flight conditions (Wang et al. 2022). Other researchers integrated CFD with other data sources. For instance, Bennett et al. (2022) coupled tracer particle data reported by U.S. Transportation Command with CFD simulations reported by Boeing and National Institute for Occupational Safety and Health (NIOSH) data to build non-linear regression models to compare airborne viral exposure based on seat occupancy (middle seats vacant or occupied) and when passengers wore surgical masks. Although using CFD is a widely accepted methodology, researchers do caution that the number of assumptions made in building simulations necessitates CFD findings to be validated with experimental data (Wang et al. 2022, 2023; see also Jianhong et al. 2022).

In a unique field study, Gameiro da Silva et al. (2023) used a multiprobe device to analyze the indoor climate on 25 flights that occurred between May 2019 and February 2020. The device measured variables including “operative temperature, relative humidity, illuminance, atmospheric pressure, CO₂ concentration and an Indoor Air Quality (IAQ) index based on the concentration of volatile organic compounds (VOCs)” (2). Measurements were analyzed to verify if the indoor environment of the airplane cabins during flights were in accordance with EN16798-1, a European standard covering indoor environmental criteria for both residential and non-residential structures. Although the standard was achieved in most cases, they contend there remains room for improvement “since the predominance of values in the best comfort category is still marginal, both in terms of thermal comfort and IAQ” (10).

To frame the above measures, it is also key to consider the explicit delimitations of the practices. First and foremost, “test conditions do not mimic the messy reality of human beings on aircraft” (Anderson 2022, 94). To demonstrate the absurdity of this point, Anderson (2022) points to a highly cited study “commissioned by the United States Transportation Command [,] United Airlines, Boeing, and some sensor and defense companies” where their “mannequins sat still and faced forward without talking, eating, or drinking[!]” (94; see also Norvihoho et al. 2023). The unrealism of these experimental conditions is further confounded with the deployment of CFD modeling because they “do not incorporate [...] cabin realities”

such as “a service cart entraining air as it moves,” “people leaving their seat,” or, more damningly, “someone in the economy section who wears a cloth mask that they remove to communicate [...] and then keep it off while they eat and drink” (94; see also Norvihoho et al., 2023).

Techniques

The most basic technique used to move air on an aircraft, operating on the individual level, is the overhead gaspers. These devices merely control airflow, and, on their own, are ineffective at accomplishing anything but aiding the individual passenger’s thermal comfort. Unsurprisingly, personalized ventilation systems (PVS) have become a hotbed of innovation. Jianhong et al. (2022) found that “PVS combined into the neck pillow is more effective than the PVS integrated into the seat at preventing cross-infection” (125), while Kurec et al. (2023) speak on both the “Airshield,” whose aim is “to utilize the air from the nozzle-optimized gaspers to create an air barrier (shield) around each passenger” and the “Airbubble” which “is an air purifier mounted close to the headrest and can protect either the [pilots] or each passenger” (3). Importantly, as McElheny et al. (2022) make clear, “the most hazardous time on a plane is during taxiing when the vents are turned off,” and so, to this end, the “vents should be open and cleaned with alcohol wipes before the team enters the plane” (534).

The next level, the collective level, involves the ECS, which is the primary means by which air is controlled inside the aircraft. The first of these collective techniques involves how the air is vented. Liu et al. (2022) compared a novel displacement ventilation system (DV) with a mixing ventilation system (MV), commonly installed in commercial airplanes to evaluate contaminant transmission and thermal comfort performance. Results were mixed for thermal comfort, with both systems offering advantages and disadvantages. Researchers noted the infection risk “under DV was lower than that under MV” at the beginning of the epidemic, but such risks were reduced in later stages of the pandemic by mask wearing. Scaffolding on this work, Li et al. (2023) conducted an experiment to ascertain which ventilation system was the most efficacious. They compared six systems:

• MV- BR (mixing ventilation with the return outlets installed in the bottom of the side wall)
• MV-CR (mixing ventilation with the return outlets located at the ceilings)
• SV-SR (stratum ventilation with the return outlets installed in the side wall)
• SV-CR (stratum ventilation with the return outlets located at the ceilings)
• ICV-SR (interactive cascade ventilation with the return air outlets installed in the side wall)
• ICV-CR (interactive cascade ventilation with the return air outlets located at the ceilings) (3).

After their analysis, Li et al. (2023) found that ICV-SR can “create a healthier human microenvironment,” “has superior contaminant removal efficiency and cross-contamination risk control ability,” and that it “seems to be more feasible and superior in terms of cross-infection risk control during the epidemic” (11-12).

Additionally, MV “is inefficient because clean air is supplied from diffusers placed far from occupants, and it has time to mix with the polluted ambient air before it is inhaled” (Danca et al. 2022, 3). The evidence instead recommends that personalized ventilation “in conjunction with MV, provides better protection against airborne infection than systems relying solely on MV” (Danca et al. 2022, 3; see also Liu et al. 2023).

This is vitally important because it has been demonstrated that lowering relative humidity “less than 40%” ensures that “droplets [evaporate] faster, leading to a reduction in the mass of droplets, which remain floating” (Danca et al. 2022, 3; see also Liu et al. 2023). Relative humidity thus is a crucial threshold when it comes to controlling for a contagion COVID-19, and this is most easily accomplished through the ECS and venting – the balance is both delicate and precarious.

Outside of controlling the air explicitly, there are indirect means by which the air can also be controlled. Bennett et al. (2022) estimated airborne virus exposure in relationship to middle seat occupancy and the

use of surgical masks by passengers. When middle seats were vacant, exposure reductions averaged 54 percent for the seat row where an infectious passenger was located and 36 percent for a 24-row cabin containing one infectious passenger. Universal masking of passengers also reduced exposures by 62 percent.

The work of Elmghraby et al. (2018, 2019) points to the need for further research on how airflow and air quality are impacted by the body forces exerted on aircraft in various flight intervals, such as takeoff, climb, and cruise. Though Elmgrhaby et al. (2018) found no significant differences in contaminant exposure between steady level (cruise) and descent intervals of a flight, “the concentration of the contaminant surrogate . . . increased substantially during the climb leg from the steady level flight” (6). Elmgraby et al. (2019) conducted an experiment on the effectiveness of airflow design and source control strategies to reduce cough-released airborne contaminants in the cabin of a Boeing 767-300 during the climb leg of a flight. They tested altering supply airflow direction and moving the coughing passenger to different cabin locations. Changing the airflow supply angle from the ceiling only had mixed results, with a 30degree angle leading to the lowest passenger exposure. Other angles (20- and 60-degrees, respectively) resulted in similar and higher exposures. In terms of source control, “the airflow patterns in the cabin, the body forces on the aircraft during the climb, and the existence of walls and/or surfaces near the cougher all have confounding effects on the resulted contaminant dispersion” (Elmgraby et al. 2019, 9). Ultimately, their findings demonstrate the need to further investigate “the potent effect body forces have on the airflow behavior and contaminate dispersion” in airplane cabins and the need for “ventilation design remedies” to mitigate health risks (Elmgraby et al. 2018, 7).

Best Practices

Rather than living in a “post-pandemic era”; researchers suggest the emergence of a “new era of pandemic” (Cheshmehzangi et al. 2023, 8; see also Ahmad et al. 2023; Bijukchne et al. 2023; Charostad et al. 2023; Matsee et al. 2023, and, especially, Yadav 2023). This calls for a new worldview, or what Speakman et al. (2022) and Lu et al. (2023) have termed a “harmonization” of the “global legal framework” as it relates policies, procedures, and protocols between governments, multinational organizations, industry sectors, regulatory agencies, trade associations, membership bodies, and professional associations. In a time of crisis, there should not be squabbling and squawking for power, but rather a rapid deferment and dissemination of evidence and expertise. This first and foremost calls for a revaluation of the underlying philosophical frameworks which guide the design practices of engineers.

Inherently safer design (ISD), a set of principles introduced by Kletz in 1978, focuses on eliminating and mitigating hazards during the design process (Gonzales-Cortes et al. 2022; Pu et al. 2023). An underlying principle of this approach is that risk reduction can be addressed more effectively when “risk assessment is introduced early in the process design and engineering stages to avoid or eliminate hazards, instead of controlling them” (Gonzales-Cortez et al. 2022, 276). The ISD approach has been applied across a broad spectrum of disciplines to reduce risks from environmental health hazards to occupational health and safety concerns. Most relevant to this research, the philosophy of Inherently Healthier Processes focuses on the prevention of “work-related injuries and illnesses” by eliminating hazards at the source, “instead of depending on PPE and management measures” (Pu et al. 2023, 5). Taking a similarly proactive risk reduction approach, the NIOSH has developed a comprehensive national strategy for eliminating workplace hazards and exposure (Pu et al. 2023). Known as Prevention through Design, or Design for Safety, this approach incorporates risk and risk mitigation in the concept and design phases of a project (Taubiz & Contos 2023).

A final piece of the philosophical framework involves continued consideration of the risks for COVID-19 infection even after attempting to eliminate them in the design process. Two philosophies that draw from occupational health and safety literature may be informative in this regard. First, the Hierarchy of Controls (HoC) approach, proposed by Manuele, identifies five main options of risk reduction in descending order of effectiveness (Ajslev et al. 2022, Pu et al. 2023). For instance, engineering controls seek to separate

workers from hazards (e.g., detection systems) and administrative controls involve “instructions, training, education, or procedures to affect the behavior, reactions, or practice of people” (Ajslev 2022, 4). A limitation of focusing on risk controls (at the lower levels of the HoC) is that they tend to be less effective than hazard reduction strategies (at the higher levels of the HoC). However, they can still inform active and passive safety measures, as well as procedural safety strategies. Second, Floyd (2023) argued for a focus on residual risks, or the amount of risk that remains after a task or process’ inherent risks have been reduced by acceptable risk controls methods, or compliance with minimum standards and regulations. Taking such a risk-based approach, Floyd (2023) contends, “promotes continual improvement in hazard identification and risk reduction” and “can help address the gap between minimum compliance and reducing risk to as low as reasonably practicable” (67) (to see such holism in action, see WRI India 2020).

During most of the pandemic, “many of the solutions adopted at most airports are detached solutions rather than integrated or interconnected solutions” (Ammoury & Salman 2023, 12). With regards to long-term and more sustainable solutions, the aviation industry has begun to shift toward “a proactive and predictive” model (Lu & Sun 2023, 22). There is evidence of this by recalling the slipshod nature at the beginning of the pandemic to now seeing recognizable clusters and packages of safety (see Lu & Sun 2023, 27-28). The aircraft best practices consist of:

• Replacing cabin air filters frequently
• Installation of HEPA air filters
• Blocking middle seats
• Reducing load factor for reasonable on-board social distancing
• Cancellation of food service
• Cabin disinfection and cleaning (Lu & Sun 2023, 28).

Another list, provided by Sanguinetti et al. (2023; see also McElheny et al. 2022), is separated into methods and methodologies. The methods they recommend are physical distancing, physical separation, divergent orientation, reduced occupancy, symptom screening, reduced exposure time, increased air exchange, air cleaning, separate air spaces, strategic air low, avoid surface contact, and surface hygiene, and the methodologies are seating configuration, pathways, barriers, ventilation and air circulation, air filtration and cleaning, on-board surface sanitization, hygienic materials, hygienic construction, touchless technology, PPE and supply provisioning, communication and monitoring, and multimodal support (see also Kamga & Eichenmeyer 2021; Luo et al. 2022; Tirachini & Cats 2020).

Notably missing, from these best practices is the time which is considered the most dangerous: “taxiing when the vents are turned off” (McElheny et al. 2022). Therefore, an additional process which should be added is that the “vents should be open and cleaned with alcohol wipes before the team enters the plane” (534).

A last, and most important, best practice is the control of passengers, specifically those who are aggressive. Anderson (2022) recommends that the sale of alcohol, both on the plane and off, should be regulated, perhaps to the point of being eliminated. Additionally, the process of reporting aggression should be made more fluid, so “that law enforcement meet the flight at the gate,” and the disruptive passenger(s) can “be escorted off first, serving as an example,” (97). Moreover, the consequences which follow should include the potential for “a system-wide air travel ban” (97). This is becoming a key issue, as the International Civil Aviation Organization noted back in 2021 that “the COVID-19 pandemic is also proving to be a psychological contamination” (cited in Cahill et al. 2023, 76), and that current “research highlights the need for a system level/human factors response to the management of well-being and mental health for aviation workers,” especially if they “consider their accountability across the triple bottom line (i.e., economic, society and ecological)” (Cahill et al. 2023, 105).

All of these best practices likely feel obvious now, especially after having read sections 3.1, 3.2, and 3.3. Yet, questions linger surrounding the long-term feasibility and viability of some practices: For example,

how many airlines are still blocking, social distancing, or cancelling? You can only lose money for so long, which leads to what Lu et al. (2023) term “P.R.I.M.” (Prevention, Retrenchment, Innovation, and Long-term Management) (35-36). The airline industry is currently between the I and the M. This leads us to the most cited best practice which has not yet been taken up by the majority of airports, although, all of the data states that they should, and it is what Ammoury & Saman (2023) call the “seamless passenger journey.” In short, a passenger should not have to touch anything from the beginning of their journey to the end that has not already been touched by them (see also Lu et al. 2023; Pilipenco 2023; Simoes e Silva et al. 2023a, 2023b).

The reason this innovation is so vital is that it has now been demonstrated that fomite transmission of COVID-19 is near zero, especially if the surfaces are “rubber” – if the surfaces are “glass, stainless steel, [or] hard-plastic,” then one merely needs to follow the protocols already laid out within this literature review (Pilpenco 2023, 8). To this end, Simoes e Silva et al. (2023a) have crafted what they call the “smart cabin concept,” extending the touchless smart interfaces into the cabin environment of an airplane. Examples of this concept include: the “Recaro Self-Cleaning Seats, which have disinfectant stored within them;” “A smarter faucet,” which is simply a faucet you do not have to touch to interact with, and the “Cabin Cleaning Light,” which is a cart of ultraviolet light which has been shown to kill “bacteria, viruses, and superbugs” (Simoes e Silva et al. 2023b, 4).

Finally, researchers have noted an increasing awareness of air cleaning as potentially promising technology that could be applied to the airplane cabin environment. Gameiro da Silva et al. (2023) stated that “far-UVC light (207-22nm wavelength) may efficiently inactivate viruses and bacteria” and was being applied in large buildings (9). However, the safety and economics of this technology and its application for mass transit have not yet been thoroughly studied.

Conclusion

Although aircraft cabins are “served by high-performance ventilation systems,” the literature reviewed for 3.3 indicates that more research is needed to ensure such systems reduce the risk of disease transmission to the extent possible (Bennett et al. 2022, 1). Moreover, the literature indicates the importance of multileveled strategies (e.g., personal, aircraft, the overall organization of travel experiences) to address risks that can be eliminated through design, as well as to control for residual and unanticipated risks. Because there are lingering questions surrounding the long-term viability of several best practices developed amid the COVID-19 pandemic, designing sustainable solutions must remain a forefront concern.