Practical Implications

• Virus-containing emissions from the respiratory tract are usefully grouped into three size ranges based on diameter: small particles (0.1–5 µm), large particles (5–100 µm), and ballistic drops (>100 µm).
• Speaking (and other vocalizations) by an infectious person when unmasked indoors appears to be especially important as a source of airborne virus-containing particles and consequently an important contributor to viral transmission of SARS-CoV-2.
• Indoor transmission at room scale is most likely a consequence of inhaling airborne particles in either the small or large size range; mask wearing, ventilation, filtration, and limiting occupant density and duration can all contribute to attenuating the transmission risk.
• Proximity-scale transmission risk is probably influenced by inhaling large particles, such contributions to transmission risk can be attenuated by mask wearing and social distancing.
• Available evidence is insufficient to eliminate from concern other transmission pathways, including some that have not received much attention, such as respiratory particle deposition onto and subsequent release from clothing fabrics.

1 INTRODUCTION

The COVID-19 pandemic has revealed inadequacies in our collective understanding of the transmission of respiratory viral diseases.1 As an illustration of this point, consider the conflicting views of prominent scholars about the role of aerosols in the transmission of SARS-CoV-2, the causative agent of COVID-19. Heneghan et al.2 wrote that “the lack of recoverable viral culture samples of SARS-CoV-2 prevents firm conclusions over airborne transmission.” Greenhalgh et al.3 responded “there is consistent, strong evidence that SARS-CoV-2 spreads by airborne transmission.”

Disagreements such as this are much more important than a purely academic concern. As of mid-November 2021, almost 2 years into the COVID-19 pandemic, the reported diagnosed cases total 250 million globally with 5.1 million deaths. The diagnosed case rate represents about 3% of the world’s population. The case fatality rate (ratio of deaths to diagnosed cases) is 2%. These disease outcomes have occurred notwithstanding major efforts to manage the spread of the pandemic that have been broadly and deeply disruptive. Understanding modes of transmission and their relative importance is central to the design of effective public health interventions.

In historical context, COVID-19 is only one among several important respiratory viral diseases. Others of note include measles, SARS, the common cold, and influenza. Knowledge about the transmission of these diseases, albeit incomplete, contributes to our understanding of how SARS-CoV-2 transmission occurs and how nonpharmaceutical interventions may contribute toward limiting the spread. Conversely, scientific and public health efforts motivated to control the COVID-19 pandemic might improve our ability to control the incidence of other respiratory viral infections in the future.

An important factor contributing to the inadequate state of understanding regarding the transmission of SARS-CoV-2 is the multidisciplinary complexity of salient processes. Important scientific and technological disciplines that would study these processes include virology, immunology, respiratory physiology, aerosol dynamics, environmental fluid mechanics, building ventilation systems, and human behavior. A sufficient future understanding will require the synthesis of contributions from multiple fields of knowledge that do not have an adequate shared history of collaboration or even cross-disciplinary communication.

Available evidence suggests that a large majority of the transmission of SARS-CoV-2 occurs indoors.4, 5 Evidence also strongly supports a view that virions emitted along with respiratory fluid from infectious persons is a major component in the chain of transmission. Such emissions are transported, diluted, and transformed during the time of travel between source and receptor. The transfer of viral materials to a susceptible person’s sensitive tissues is pivotal in initiating new infections.

This article aims to contribute to an improved state of understanding about SARS-CoV-2 transmission by highlighting key elements from the disciplinary perspective of indoor aerosol science and technology. Key elements that are pertinent to the airborne transmission of SARS-CoV-2 are reviewed, including emissions from the respiratory tract, environmental transport and fate, and uptake by a susceptible person. There is a strong emphasis on aerosol science, elements of which transcend the three major processes in the chain of transmission.

The review is structured as follows. After this introduction, Section 2 defines important terms and summarizes several key concepts. Section 3 presents pertinent information from COVID-19 outbreak investigations. Section 4 reviews four pivotal aspects of the indoor dynamic behavior of particles emitted from the respiratory tract: how transport distance varies with particle size, the rate of deposition to indoor surfaces, the role of water evaporation inducing particle shrinkage, and particle motion attributable to inertia. Section 5 provides a qualitative and quantitative summary of particle emissions from the respiratory tract, emphasizing the size distribution of particles emitted along with where in the respiratory tract the particles originate. This section also stresses the distinctions between emissions from illness symptoms, such as coughing and sneezing, versus asymptomatic generation activities, such as talking and quiescent breathing. Section 6 addresses particle deposition in the respiratory tract upon inhalation, considering especially the influence of particle size on regional deposition. Sections 7 and 8 assess the source-to-receptor transmission process, respectively, considering transmission that occurs with room-scale separation between the infectious and susceptible persons and transmission in close proximity (as could occur, for example, in face-to-face conversation). Among nonpharmaceutical interventions to reduce the disease spread are several that interrupt aerosol transmission: masking, increased ventilation rates, and the use of room air filtration systems. These are described in Section 9. The final two sections provide a synthesis of the state of knowledge (Section 10) and concluding thoughts (Section 11).

I emphasize that this article has an indoor aerosol science focus, which represents only a portion of the important aspects of the COVID-19 pandemic. The review is written from the perspective of one person, possessing expertise in aerosol science, indoor environmental quality, and environmental engineering science. These areas of expertise emphasize a fusion of empirical evidence with mechanistic, process-oriented descriptions, along with a particular focus on source-receptor relationships. Another orientation of these subject areas is the need for science to guide the development and application of technology for improving the human condition even when the state of knowledge is incomplete. An overarching goal in preparing this review is to contribute toward more complete illumination of what is known and what we do not yet know about the indoor aerosol science aspects of SARS-COV-2 transmission. My perspective is limited by the scope of my experience and imagination, permitting only partial insight into important aspects of the complex process of indoor aerosol transmission of SARS-CoV-2.

2 KEY CONCEPTS AND TERMS

Clear communication is an important challenge in making progress toward understanding transmission of respiratory infections. With contributions needed from multiple disciplines that have developed independently there is considerable risk of misunderstanding because of differences in the denotation and connotation of terms. Such difficulties are highlighted in a table in Tang et al.6 that describes the different understandings of terms such as “airborne,” “aerosol,” and “particle” by clinicians, aerosol scientists, and the public. One purpose of this section is to define the language used for essential elements of this paper.

Along with exhaled air, a person with a viral respiratory tract infection can emit finely divided aqueous particles and drops that contain virus. The suspended aqueous materials include mucus originating from the lining of the respiratory tract airways and saliva from the oral cavity. The site of origin of particles and drops in the respiratory tract can vary according to the generating activity and associated particle-production mechanism.

2.1 Regions of the respiratory tract

It is convenient to divide the respiratory tract into three regions. The passages in the head can be referred to collectively as the naso-oro-pharyngo-laryngeal (NOPL) region. This zone will also be referred to as the “head region” of the respiratory tract. Beginning with the trachea and proceeding through many generations of bifurcation are the conducting airways known collectively as the tracheobronchial (TB) region. Gas exchange occurs in the most distal portions of the lung known as the pulmonary (P) or alveolar region.

2.2 Particle and drop size classification

Aqueous particles emitted from the respiratory tract span a vast range of sizes, with diameters varying from less than 1 µm to more than 1 mm. A three-order-of-magnitude range of diameters corresponds to a factor of 1 billion (109) difference in volume or mass. Dynamic behavior varies markedly for airborne particles across this size range. At the smaller end, particles can remain airborne for long periods (hours) and, except when very close to surfaces, their movement is mainly associated with the prevailing motion of the air in which they are suspended. At the upper end of the size range, the drops behave ballistically, with motion dominated by their initial momentum and attenuated by air resistance. Milton7 describes four particle size categories, incorporating information about respiratory tract deposition into the classification. Respirable particles are smaller than 5 µm in diameter. They can reach the pulmonary region of the respiratory tract and deposit there. At the other extreme are ballistic drops, larger than about 100 µm in diameter, which are not substantially inhaled. The intermediate size categories described by Milton are termed thoracic (≤10–15 µm, capable of entering the TB region and depositing there), and inhalable (≤100 µm). Pöhlker et al.8 have also provided a carefully described set of terms for the condensed-phase emissions from the respiratory tract relevant to infectious disease transmission. For this paper, I will use language that reflects a hybrid of the terms from Pöhlker et al. and from Milton. Small particles will refer to those with diameters less than 5 µm. Large particles will be those with diameters in the range 5–100 µm (or 5–50 µm in some cases). Small particles are respirable; large particles are inhalable. Ballistic drops (or simply drops) are larger than 100 µm in diameter and are generally not inhalable.

2.3 Generation of respiratory particles and drops

Many activities produce airborne emissions of respiratory particles and ballistic drops. For SARS-CoV-2, much of the transmission occurs from individuals who are asymptomatic.9 Consequently, even quiet breathing and talking by people who are not evidently ill should be considered as having the potential to initiate transmission. Historically, much attention has focused on emissions resulting from respiratory symptoms of illness, such as coughing or sneezing. That focus appears misdirected for understanding the bulk of community transmission of SARS-CoV-2.

Several articles discuss the mechanisms of particle and drop production in different regions of the respiratory tract.8, 10, 11 Relevant mechanisms include the bursting of bronchiolar fluid films during quiescent breathing, the closing and opening of folds in the vocal cords during talking and other vocalization activities, and the shear-induced instability of airway lining fluid in the tracheobronchial region and the mouth. Section 5 of this paper explores in detail the number and size of particles generated in different regions of the respiratory tract by different activities.

2.4 Particle shrinkage caused by water evaporation

An important feature of respiratory emissions is that the aqueous particles undergo evaporation, losing water, and consequently shrinking. The speed of evaporation is rapid for small particles and for the smaller portion of the large-particle mode.12 The extent of shrinkage depends on the ambient humidity and on the proportion of nonvolatile solutes in the emitted particles. Evidence suggests that the diameter of equilibrated particles after shrinkage is in the range 25%–50% of the emitted particle diameter.8, 12, 13 For the larger portion of the large-particle mode and for ballistic drops, the evaporation rate can be sufficiently slow to not strongly influence airborne behavior prior to deposition.14 Section 4.3 presents more information about evaporation-associated particle shrinkage.

2.5 Modes of transmission

In the classical view, there are three potential modes for the transmission of respiratory viral infectious agents.15 One mode combines droplet spray and direct contact. Transmission by this mode occurs at close range and either involves direct physical contact between infectious and susceptible persons or the direct transfer via ballistic drops of viral material. A second mode is fomite mediated, in which a surface becomes contaminated with virus-containing emissions, a susceptible person acquires the virus by hand contact with the surface, and then transfers the virus to susceptible tissues in the eyes, mouth, or nose. The third mode is termed airborne or aerosol transmission; it critically entails inhaling the infectious agent contained in suspended particles. An important source of confusion in the literature concerns the role of exposure via inhalation when the infectious and susceptible persons are in proximity. Some have conflated transmission in proximity with the predominance of the droplet spray and direct contact route. However, others have clarified: exposures are elevated when in proximity not only via droplet spray but also by aerosol inhalation. Consequently, identifying proximity as a factor in disease transmission does not reliably confer understanding about which transmission mode is dominant. Aiming to improve clarity, Li has separated the transmission scale from the transmission mode, indicating that, when in proximity, transmission can occur by any of the three modes of spray, inhalation, and touch, whereas at a distance (room scale), only inhalation and touch (via fomites) can occur.16

2.6 Transmission scales

For indoor exposure to a respiratory virus, the transmission scale can be key to understanding and mitigating risk. The finest transmission scale is termed short-range,11 near field,8 close contact,16 or source proximate.17 Although not precisely defined, the relevant separation scale between the infectious and susceptible persons would typically be less than 2 m. Interpersonal distances for comfortable indoor conversation or seating separation at a meal would be examples of the short-range scale. Key features of short-range scale include the potential for droplet spray transmission and for aerosol transmission at high efficiency in an exhaled plume.

A second indoor transmission scale applies to a room or to all the rooms that are well connected through open doorways or recirculating air-handling systems in a small building. For convenience, we will term transmission in these circumstances to be at room scale. Bond et al.17 termed the processes operating on this scale as confinement effects. Other authors have termed this scale long range11 or distant.16 Those latter terms are avoided here because of potential confusion about transmission over scales much larger than a room.

This third indoor transmission scale applies when transport occurs over larger dimensions within a single building or even from one building to another. This larger scale has not been prominently evident in SARS-CoV-2 transmission, although there are examples of transmission seeming to have occurred between different dwelling units in an apartment building.18 Transmission from one dwelling to another via open windows was investigated in Hong Kong apartment buildings for SARS.19

This review substantially addresses transmission at room scale (Section 7) and when source and receptor are in close proximity (Section 8). The article does not specifically address transmission at the third or largest indoor scale, as defined in the previous paragraph.

2.7 Indoor air motion and ventilation rates

Indoor air motion and ventilation or air-change rates can influence the risk of transmission. Air motion is characterized by the velocity field, the time-dependent pattern of spatially varying air speeds and directions. Information on indoor velocity fields is limited, especially for circumstances that might be important for community spread of respiratory viruses. Available evidence indicates that a median indoor air speed of ~10 cm/s might be typical with some common conditions producing speeds as low as 1 cm/s or as high as 1 m/s.20-22 Ventilation rate as used here will represent the supply of outdoor air to an indoor space. A common unit of measure for highly occupied spaces is volume flow rate per occupant, such as liters per second per person (L/s pers−1). Typical values specified in guidelines and standards are in the range 5–10 L/s pers−1. A poorly ventilated space might have a ventilation rate of <3 L/s pers−1 whereas a highly ventilated space might have a ventilation rate >25 L/s pers−1. An indicator of ventilation rate is the increment of CO2 indoors above the outdoor concentration. Persily and de Jonge23 reviewed the evidence regarding carbon dioxide generation rates for building occupants, reporting central tendency values in the range 0.0025 L/s per person (18 g/h per person) for a child’s bedroom to 0.0055 L/s per person (39 g/h per person) for a lobby. Using 21 g/h per person (0.0030 L/s per person) for a classroom of pupils in the age range 5–8 years, outdoor air ventilation rates of 3, 10, and 25 L/s pers−1 would correspond to steady-state increments in indoor CO2 level of 1070, 320, and 130 ppm, respectively (assuming p = 1 atm, T = 293 K). For residences, outdoor air ventilation is commonly reported in terms of the air-change rate, which is the outdoor air ventilation rate normalized by the indoor volume. A typical value for residential air-change rates is 0.5 h−1. Air-change rates at a particular time in a particular residence might commonly vary across about two orders of magnitude, 0.05–4 h−1.24

2.8 Infectious dose

In mechanistic investigations of infectious disease transmission, it would be useful to know the quantity of virus needed to initiate a new infection. That information is generally not accessible, however. Some clues may be available from laboratory studies such as the tissue culture infective dose, or the dose needed to initiate a reaction in a laboratory animal. Quantitative modeling of disease transmission often utilizes the concept of infectious quanta. When this approach is used, whether for SARS-CoV-2 or for other infectious agents, the probability of a susceptible person becoming infected (p) is related to the number of infectious quanta inhaled (q) according to the expression p = 1−e−q, so that if one quantum is inhaled, the likelihood of a new infection is 1 − e−1 = 63%. Examples are available for applying this concept to the case of SARS-CoV-2.25, 26

2.9 Disease isotropy

When considering potential pathways of infection, it is important to recognize that the infectious dose may be related to the receptor target. In other words, the quantity of virus necessary to initiate infection might be different if the virus deposits in the pulmonary region of the respiratory tract versus on the mucus membranes of the nose or mouth. To address such ideas conceptually, Milton27 introduced the concept of disease isotropy. An isotropic infection is one that would be “transmitted with equal effectiveness and virulence by all routes, whether aerosol, large droplet, or direct contact ….” Milton identified smallpox as an anisotropic infection, “most effectively and virulently transmitted by fine particle aerosols.” Influenza is also classified as anisotropic, “with aerosolized virus infectious at lower doses and more likely to result in ‘typical influenza-like disease’ (fever plus cough) than intranasal inoculation.”28 Whether COVID-19 is similarly anisotropic is not yet known.

2.10 Superspreading transmission

A substantial proportion of SARS-CoV-2 transmission is believed to occur in superspreading events.29-31 In such events, a sizeable number of new infections occurs, commonly from a single infectious person. The circumstances of a gathering can contribute to superspreading events, as has been documented in the case of a choir practice.26 Other evidence points to “supershedding” as a possible factor in superspreading events. For example, Bueno de Mesquita et al.32 reported considerable variability among individuals in the rate of shedding of the influenza virus. Asadi et al.33 found that some individuals were particularly high emitters of respiratory particles during speech. The concept of superspreaders is well established for other respiratory viral infections, including the common cold caused by coxsackievirus A21,34 measles,35 SARS,36 and influenza.37

3 OUTBREAK INVESTIGATION RESULTS HIGHLIGHT THE IMPORTANCE OF INDOOR ENVIRONMENTS

Case reports from outbreaks of COVID-19 provide important clues about how SARS-CoV-2 is transmitted. The most extensive such study assessed 318 outbreaks in China, each with three or more cases.5 The authors found that “all identified outbreaks of three or more cases occurred in indoor environments.” The authors also reported that among “7324 identified cases in China with sufficient descriptions, only one outdoor outbreak involving two cases occurred.” All these cases occurred during the first winter of the pandemic, January-February 2020. Among the 318 outbreaks, only three involved 10 or more cases. Homes and transport environments were the dominant location category. The study relied on municipal case report investigations, which varied in quality, and which tended to lack some important information needed to fully understand transmission pathways.

More detailed investigations have been conducted of several specific outbreaks, including in a call center,38 in fitness centers,39 at a restaurant,40 and at an apartment building18 in South Korea; in a nursing home in the Netherlands41; in a choir rehearsal in the USA26; in a restaurant in China42; on board the Diamond Princess cruise ship43; and in a courtroom in Switzerland.44

Madewell et al.45 undertook a systematic literature review and meta-analysis of publications reporting household transmission of SARS-CoV-2. They identified 54 relevant studies aggregately reporting almost 78,000 household secondary transmissions. The estimated overall household attack rate was 17%, was higher among spouses (38%) than other family contacts (18%) and was higher for adult contacts (28%) than for child contacts (17%). The authors highlight that households are “closed spaces, where family members may crowd and be in close contact with conversation. There may be reduced use of personal protective equipment relative to other settings.” Potentially important information such as the degree of crowding and the ventilation conditions of the dwellings was not available.

Bulfone et al.4 reviewed the published evidence available through August 12, 2020 regarding outdoor transmission of SARS-CoV-2. They concluded that “existing evidence supports the wide-held belief that the risk of SARS-CoV-2 transmission is lower outdoors but there are significant gaps in our understanding of specific pathways.”

In aggregate, these studies support a view that indoor environments are important in the transmission of SARS-CoV-2. However, they do not provide clear evidence about the relative importance among different modes of transmission.

It is also important to recognize that the outbreak investigations cumulatively represent a very small fraction of SARS-CoV-2 infections. The proportion is almost certainly less than 10−3 (<0.1%). The conditions that allow outbreaks to be investigated are not necessarily statistically representative of all conditions in which transmission occurs. The metaphor of the tip of the iceberg comes to mind, but in the case of an iceberg, the visible portion above water is about 10% of the total. The other relevant metaphor is lamppost science, with the caution that we should not only be looking where the light is good.46 The main points are to accept what clues can be gleaned from this evidence but to recognize the limitations and to resist the temptation to generalize from investigations in which the outcomes depend on peculiar circumstances.

4 PARTICLE AND DROP DYNAMIC BEHAVIOR

The SARS-CoV-2 virion is approximately 0.1 µm in diameter. Zhu et al.47 reported a range of 0.06–0.14 µm, based on electron micrography of negative-stained viral particles. In the environment, respiratory viruses would not commonly exist as free entities. In air, they would be associated with aqueous suspensions of mucus and saliva emitted from the respiratory tract. Evaporation of water would cause the emitted particles to shrink to some extent, with the ultimate degree of shrinkage limited by the nonvolatile substances in the emitted particles and potentially influenced by the local relative humidity.

Particles and drops emitted from the respiratory tract span a diameter range of more than three orders of magnitude, from less than 1 µm to more than 1 mm. It is worth repeating that a 1 mm drop contains a billion times more material than a 1 µm particle. On the other hand, there are orders of magnitude more small particles emitted than ballistic drops, so the contribution of small particles to the total quantity of material emitted can be substantial.

The size of particles and drops containing potentially infectious virions is centrally important in the disease transmission process. Size pertains to where in the respiratory tract the emissions originate.10 The viral load or virion concentration of respiratory fluids can vary with location in the respiratory tract. Combining these two factors, respiratory particles and drops of different sizes could have different virion concentrations. In general, emissions that originate from deep in the respiratory tract tend to be small particles, whereas large particles and especially ballistic drops are generated in the NOPL region. Environmental transport, dynamic behavior, and fate of airborne particles and drops are strongly influenced by size. Whether and where virion-containing particles and ballistic drops deposit on the exterior surface of a susceptible person or within his or her respiratory tract is also highly dependent on size.

4.1 Transport distance varies with particle size

The literature on respiratory viral disease transmission is permeated with a profoundly significant error regarding particle size. As one example, Seto48 wrote, “it is apparent now that only small particles of <5 µm … will result in airborne transmission potentially over longer distances because these particles can remain suspended in the air for prolonged periods. Most lung infections result in droplet transmission whereby the larger particles from the cough are transmitted for <1 m and do not remain suspended in the air.” This statement contains some truth: large (ballistic) drops do not travel far. But it is seriously wrong in suggesting that the particles larger than 5 µm cannot travel beyond 1 m distance. In describing the history of such misunderstanding, which is widespread, Randall et al.49 emphasize that the pattern of deposition in the respiratory tract became inappropriately conflated with airborne travel distance. The facts are these. First, 5 µm is an appropriate estimate of the upper-bound size for particles that penetrate to and deposit in the pulmonary region of the respiratory tract.50 Second, the minimum diameter of a ballistic drop that is likely to travel no further than 1 m from the source is about 100 µm.1, 6

It is difficult to overstate the importance of this error. Particles in the size range 5–100 µm do not fall to the floor within a meter of the emission source, but rather can travel room-scale distances. Furthermore, these particles can be inhaled and deposit in the respiratory tract. Because emissions in the diameter range 5–100 µm can be substantial and because particles in this size range are inhalable, there is a great risk of misunderstanding the potential for inhalation exposure to respiratory viruses by mistakenly believing that all particles larger than 5 µm fall within 1 m of the source.

To elaborate, consider the information presented in Figure 1. The data points and connecting line segments display the settling velocity of water drops and particles when the two main forces—gravity and drag—are balanced.51 The three horizontal lines, marked at 1, 10, and 100 cm/s, respectively, represent approximate lower-bound, central tendency, and upper-bound values for indoor air speeds. The vertical lines divide the particle or drop size into zones, as follows. For particles smaller than about 20 µm in diameter, particle motion is more strongly controlled by air flow than by gravitational settling under all common indoor conditions. Conversely, for drops larger than about 300 µm in diameter, gravitational settling dominates, independent of the prevailing indoor air motion. In the intermediate region, the movement of particles and drops reflects a balance between the influence of indoor air currents carrying suspended particles and the influence of gravity pulling the particles downward. For indoor air speeds that are relatively still, 1–10 cm/s, particles in the range of 20–60 µm diameter can experience motion that is about equally balanced between the effects of air currents and gravitational settling. For indoor air speeds that are moderately high, 10–100 cm/s, the corresponding size range for these balanced flow conditions is approximately 60–300 µm.

Data points and upward sloping line represent the terminal settling velocities of water droplets as a function of particle and drop diameter.51 The three horizontal lines indicate approximately the minimum, central tendency, and maximum indoor airspeeds commonly encountered in occupied spaces

A common distance for a drop to settle from the mouth or nose of an infectious person to an upward surface is on the order of 1 m. That distance might be as high as about 1.5 m from a standing person to the floor. It might be less than about 0.5 m from a seated person to a table. In a magnitude sense, the noteworthy comparison would be the competition for an airborne particle between settling through 1 m versus being transported by air currents a horizontal distance of 1 m. The assessment is a magnitude comparison in part because the relevant fall distance is variable, as noted above, and because of variability in both the speed and direction of indoor air currents.

Regarding the transport range of expired drops and particles, the evidence presented in Figure 1 can be summarized as follows. Drops larger than about 300 µm will fall within 1 m of the source largely independent of the indoor air flow conditions. If the indoor air motion is moderately strong (ie, with speeds of 10–100 cm/s), then some portion of the particles in the size range 60–300 µm will also fall to an upward surface within 1 m of the source, with the proportion being progressively larger as particle size increases. Particles smaller than 60 µm will remain suspended beyond 1 m from the source to a substantial extent under these conditions. For weak indoor airflow conditions (speeds of 1–10 cm/s), large particles can fall substantially to upward surfaces, with the possibility that the proportion that settles is greater than the proportion that travels beyond 1 m down to particle diameters of approximately 20 µm. Under any airflow conditions, most emitted particles that are smaller than 20 µm in diameter will remain suspended long enough to travel well beyond 1 m from the emission source.

In addition to background air motion and gravitational settling, the motion of respiratory drops can be influenced by the impetus provided with their emission. Bourouiba52 investigated the fluid mechanics of exhalations, sneezes, and coughs. She reported that “given various combinations of an individual patient’s physiology and environmental conditions, such as humidity and temperature, the gas cloud and its payload of pathogen bearing droplets of all sizes can travel 23–27 feet (7–8 m).” Xie et al.14 likewise reported that “expelled large droplets are carried more than 6 m away by exhaled air at a velocity of 50 m/s (sneezing), more than 2 m away at a velocity of 10 m/s (coughing) and less than 1 m away at a velocity of 1 m/s (breathing).” The travel distance of emissions is substantially reduced if the mouth and nose are covered.53-55

4.2 Particle deposition to room surfaces

The deposition of airborne particles to room surfaces commonly dominates the total loss rate of large particles from indoor air. Across the size range of particles emitted from the respiratory tract, the influence of deposition to room surfaces as a removal mechanism is both highly variable and potentially important. For particle diameters larger than about 10 µm, this loss process has not been well studied.

One relevant experimental investigation was reported by Thatcher et al.56 In a (small) room-sized experimental chamber (V = 14 m3), the size-resolved particle deposition rate was determined following pulsed injection of polydisperse particles spanning the approximate diameter range 0.5–10 µm. The experiments were conducted for three distinct furnishing levels (empty, carpet only, and fully furnished) and for four airflow rate conditions (fans off and mean air speeds induced by fan of 5, 14, and 19 cm/s). Results were reported in terms of size-resolved deposition rates for the different conditions of indoor airflow and furnishing level. Considering that densely occupied spaces would tend to have high surface areas and high air movement, the fully furnished condition with the highest airspeed is selected for analysis here.

Figure 2 displays the experimental measurement results for deposition loss rate from Thatcher et al.56 for the 19 cm/s average airspeed in a fully furnished room. The experimental data only span the range of 0.5–9 µm, so extrapolations are presented to extend the ranges downward to 0.1 µm and upward to 50 µm diameter. For particles larger than 1 µm, Thatcher et al. found that the effect of higher interior air motion was substantial, with an average deposition loss rate 2× as high for an airspeed of 19 cm/s compared with fan off conditions (with airspeed <2 cm/s) in the fully furnished chamber. The influence of furnishing level was smaller, with the average loss rate for the fully furnished case being 23% higher than for carpet only (in both cases with an average airspeed of 19 cm/s).

Deposition loss rate to room surfaces as a function of particle diameter. The experimental data are for a small furnished room with an average air speed of 19 cm/s, as measured by Thatcher et al.56 The dashed lines represent extrapolations, with an estimated deposition loss rate of 0.3 per h for particle diameters of 0.1–0.5 μm. The large-particle extrapolation is based on a linear regression of log(loss rate [h-1]) against log(diameter [μm]) using the six measurements with largest diameters. The slope and intercept of the regression line are 1.27 and −0.125, respectively. Note that for particles at the lower end of the size range and smaller, the loss rate would tend to increase with diminishing size, owing to the increasingly strong influence of Brownian diffusion for smaller particles

4.3 Respiratory particles shrink through water evaporation

A noteworthy feature of respiratory emissions is that the particles and drops shrink owing to evaporative loss of water.12 Upon leaving the respiratory tract, aqueous particles emitted indoors encounter an environment with lower temperature and lower water vapor content. Since water is the dominant component of the emitted particles, its evaporation can be sufficient to cause particle size to shrink. The shrinkage phenomenon has several complex aspects, including heat and mass transfer kinetics and the influence of dissolved solutes and suspended mucins. In brief, the key elements are these (a) small particles evaporate so rapidly that they reach their equilibrium state essentially instantaneously; (b) ballistic drops may fall to the floor before much evaporation occurs; and (c) the nonvolatile impurities may be sufficient in abundance so that shrinkage is limited to a diameter decrease typically in the range of about 2–4×. This last point means that what was classically referred to as droplet nuclei,57 that is, the persistent, airborne dried residue of emitted droplets with a maximum size of about 5 µm, would have been emitted with a diameter of less than 20 µm. Conversely, drops emitted at 100 µm diameter or larger would not shrink to become droplet nuclei as classically understood. Indeed, these larger drops deposit by settling more rapidly than they attain their equilibrium size by means of evaporation. Being limited to about a factor of 4 effect on diameter, the role of evaporation is much less important for airborne particle dynamics than the variability in the sizes of emitted particles and drops, whose diameters can span a factor of 1000 or more. Nevertheless, evaporation-associated shrinkage can meaningfully influence airborne particle dynamics, such as the loss rate by deposition to indoor surfaces, and the pattern of deposition if inhaled.

The state of knowledge about the chemical composition of emitted respiratory particles and drops is surprisingly incomplete. Nicas et al.12 modeled the emissions as a combination of normal saline with glycoproteins and estimated that a fully desiccated particle would shrink to 44% of the diameter of the aqueous emission. Liu et al.58 assumed a model solution of 150 mM of NaCl in predicting a dried droplet nuclei size of 32% of the original diameter. Walker et al.59 concluded that a drop of 100 µm emitted into air at 50% relative humidity would have an equilibrium diameter of 28 µm for artificial saliva and 30.4 µm for artificial deep lung fluid, respectively. Investigating the seasonality of influenza infections, Marr et al.60 proposed a novel physicochemical interpretation in which the evaporation of respiratory emissions influenced droplet physics and chemistry in such a way as to modulate virus survival and transmission. Recent investigations of the evaporation of respiratory droplets include a detailed modeling study61 and single-particle levitation experiments.62

To illustrate the extent of shrinkage quantitatively, consider the case of normal saline, an aqueous solution with 9 g/L of NaCl. The density of dry NaCl is 2.2 g/cm3, so the volume ratio of this solution would be 4.1 cm3 of dry NaCl per 1 L of normal saline solution, corresponding to a volume ratio of 0.41%. An emitted particle would, if fully dried, shrink to a diameter that is (0.0041)1/3 = 16% of the initial diameter. This value is somewhat lower than the range expected for respiratory droplets because it does not account for the contribution of nonvolatile mucins. Also, depending on the ambient humidity, the emitted particle may not become completely desiccated.12

Because of limitations in the available evidence, the effects of particle shrinkage through water evaporation are not explicitly incorporated into the emissions, fate, and exposure calculations reported subsequently. Refinements to the calculations to incorporate the influence of evaporation would be warranted as information improves about the composition of emitted respiratory particles.

4.4 Inertial drift is a short-range phenomenon

Another primary consideration for the airborne behavior of respiratory emissions is inertial drift. Airborne particles tend to follow accelerating, decelerating, or bending streamlines only to the extent that the drag exerted on them by air is sufficient to overcome their momentum’s tendency to continue moving at a constant speed in a straight line. Inertia is relatively weak for ordinary indoor airflows and small particles; however, for large particles and especially for ballistic drops inertia becomes increasingly important. A characteristic measure of the strength of inertia is the stopping distance, a calculated result indicating how far a particle would travel before coming to rest if injected into stagnant air at an initial velocity Uo.63 In this calculation, the only force opposing particle motion is drag. For the purposes of illustration, assume that drag is described by Stokes law, for which it is necessary that the particle Reynolds number is much smaller than 1. For ballistic drops moving at high speeds, this assumption does not hold. The drag would be higher and the stopping distance smaller than predicted by equation 1).

Equation (1) evaluates the stopping distance, S, for a particle of diameter, dp, and density, ρp, injected with initial velocity, Uo, into stagnant air of viscosity, µ. Table 1 summarizes the stopping distance for water droplets across a range of particle diameters and initial velocities.
(1)

TABLE 1.
Stopping distances (S) for particles and drops of different sizes as a function of initial velocity

	dp =1 µm	dp =10 µm	dp =100 µm
Uo =1 cm/s	3 × 10−6 cm	3 × 10−4 cm	0.03 cm
Uo =10 cm/s	3 × 10−5 cm	0.003 cm	0.31 cm
Uo =1 m/s	3 × 10−4 cm	0.03 cm	3 cmb

Several important qualitative features are evident in Table 1. First, inertial drift is not important for transporting small particles. Second, inertial drift is generally a short-range phenomenon, operating on scales of a few cm or less for inhalable particles at ordinary indoor air speeds. Third, inertial drift can be meaningful for transport close to surfaces for large particles (or, by inference, for ballistic drops), especially at higher air speeds.

5 EMISSIONS OF RESPIRATORY PARTICLES AND DROPS

5.1 Overview

Respiratory particle emissions have been investigated for different types of activities: breathing, talking (and other vocalization activities), coughing, and sneezing. This research has been concerned with application to infectious disease transmission, including not only viral diseases but also bacterial diseases such as tuberculosis64 and whooping cough.65 Earlier studies characterizing emissions emphasized coughing and sneezing. More recent investigations have emphasized tidal breathing and vocalization.

For each type of important respiratory activity, one would like to know the pattern in terms of the time rate of emission of size-resolved numbers of particles and drops. One would like to have this information in relation to influencing variables, such as the pattern of breathing or the intensity of speech. One would especially want to know how the infectious agents are distributed among the respiratory emissions. Also, one would ideally have such information not only across populations of individuals but also longitudinally over time as the period of infectiousness progresses. The pattern of how an infectious agent is distributed across the size of respiratory emissions almost certainly varies from one infectious agent to another. And, depending on how the infectious agent reproduces in and is shed from different regions of the respiratory tract, that distribution may even vary with circumstances (e.g., with the time course of infectiousness) for a given illness. In combination, these circumstances are scientifically and technologically challenging. Available evidence provides important clues; however, a complete picture is not yet accessible of the aspects of respiratory emissions that are pertinent for understanding respiratory disease transmission.

Several cautions should be recognized when considering the available evidence. First, many published studies only report a portion of the emitted size range of particles and drops. Older studies relied on manual sampling and analysis technologies that were effective in capturing the larger emissions but not so reliable in determining size segregated smaller particles.66, 67 Conversely, many recent studies rely on automated instruments for measuring airborne particle size distributions that are unable to characterize ballistic drops.33, 68

Second, most empirical studies have characterized emissions without regard to their viral (or bacterial) content. A common approach used to estimate size-resolved virion emissions combines size-resolved particle and drop emissions with information about the viral load of respiratory tract fluids. This approach is vulnerable to large errors if the part of the respiratory tract where particles are generated has a different abundance of virions than the sampled respiratory fluids. This point is elaborated below.

Third, the role of particle shrinkage through water evaporation is not consistently addressed. Specifically, reported size distributions are sometimes based on conditions at the mouth, where the particle composition would approximately match that of the respiratory fluids. However, measurements are often made under conditions in which particles will have shrunk to their equilibrated size after water evaporation.

5.2 Size distribution of emitted particles and drops

Pöhlker et al.8 recently provided a thorough review and assessment of particle and drop emissions from the respiratory tract in the context of infectious disease transmission. That report effectively addresses two major concerns about emission patterns. First, the authors consider a full range of particle and drop sizes, from less than 0.1 µm diameter (smaller than the SARS-CoV-2 virion) to drops much larger than 100 µm. Second, they interpret the available evidence in a way that associates the emitted particles and drops with sites of origin in the respiratory tract. This latter point is missed in much of the prior literature and is pivotal for ultimately associating respiratory emissions with infectivity.

Pöhlker et al.8 highlight three primary mechanisms leading to particle generation in four regions of the respiratory tract. One mechanism is bronchiolar fluid film bursting, which is believed to be the primary means of particle generation in the pulmonary region.69 A similar process may occur in the moving folds of the vocal cords during vocalization activities. A second mechanism is associated with high-speed airflows that cause strong interfacial shear and can produce particles and drops through fragmentation of stretched mucus or saliva. This second mechanism could occur in the trachea and in the upper respiratory tract (oral and nasal cavities) and would be especially pronounced for vigorous expiratory actions, such as coughing or sneezing. A third mechanism involves the motion of mouth, lips, and tongue through which drops are formed mechanically from saliva. The four relevant regions for particle generation are the small bronchioles in the distal airways, the trachea and main bronchi, the larynx and associated vocal cords, and the oral cavity.

Consolidating and distilling the available empirical evidence, Pöhlker et al.8 present parameters of lognormal distributions that are intended to represent the central tendency of particle and drop emission for three regions of the respiratory tract and for three different respiratory activities. The three regions are the bronchiolar (B), larynx and trachea (LT), and the oral cavity (O). The respiratory activities characterized are breathing, speaking, and coughing.

Table 2 presents a further quantitative interpretation of the work of Pöhlker et al., in which their reported lognormal representations are integrated over size sections that correspond to small particles (0.1–5 µm), large particles (5–100 µm), and ballistic drops (100 µm–1 mm). Results are presented separately for the three sites of origin (B, LT, and O) and for the three reported respiratory actions.

TABLE 2.
Particle and drop emission rates (EN, by count, and EV, by volume) from the respiratory tract during breathing, talking, and coughing

	Breathing		Speaking		Coughing
	EN (h−1)	EV (µm3/h)	EN (h−1)	EV (µm3/h)	EN (cough−1)	EV (µm3/cough)
B modes
Small PM (0.1–5 µm)	8.2 × 105	2.6 × 104	2.0 × 106	64 × 104	7.9 × 104	3.1 × 102
Large PM (5–100 µm)	1.4	147	3.5	360	0.0	0.2
LT mode
Small PM (0.1–5 µm)	—	—	8.2 × 105	2.0 × 106	4.5 × 103	1.2 × 104
Large PM (5–100 µm)	—	—	4.8 × 103	7.2 × 105	46	8.3 × 103
O modes
Small PM (0.1–5 µm)	—	—	2.5 × 103	7.7 × 104	190	6.0 × 103
Large PM (5–100 µm)	—	—	6.0 × 104	8.0 × 109	1.6 × 103	4.9 × 107
Drops (100 µm–1 mm)	—	—	4.2 × 104	3.0 × 1011	370	4.7 × 109

A few details of the analysis supporting Table 2 should be highlighted. The minimum particle size in the calculation was selected as 0.1 µm because that is the approximate size of the SARS-CoV-2 virion. Smaller particles cannot contain the virus. The results are reported in emission rates or as emission factors, both in terms of count (EN, number or particles or drops per time, for example) and in terms of volume of the condensed-phase respiratory emissions (EV). The primary reporting by Pöhlker et al. is in terms of concentration in undiluted exhalations; I have used exhaled volumes as reported by Pöhlker et al. for breathing (360 L/h), speaking (700 L/h), and for a single cough (1.5 L) to obtain the emission rates and emission factors reported in Table 2. The assumed flow rate associated with breathing corresponds to a sedentary or passive condition. The speaking air flow is based on an estimated rate of 120 words per minute. The frequency of coughing is highly variable, so emissions are expressed on a per-cough basis, rather than per time.

Sneezing is another type of respiratory emission event that can contribute to infectious disease transmission. Particle and drop emissions from sneezing have been characterized experimentally by Duguid66 and Han et al.70 In their review, Pöhlker et al. reported that the Han et al. data were presented in such a way that “no mode-specific particle number and volume concentrations could be retrieved for the further steps of our analysis.”

Some important qualitative features emerge from Table 2. This discussion emphasizes particle and drop volume emissions, as viral content seems more likely to vary with that indicator rather than with emissions by number. Emissions from the deep lung (B modes) are primarily in the small particle size range. The estimate for speaking is of similar magnitude but somewhat higher than for quiescent breathing, mainly because the exhalation flow rate for speaking is assumed to be higher (by about 2×). Considering all small-particle emissions for speaking, the LT mode is the largest, more than an order of magnitude bigger than the B and O modes. Large-particle emissions from the oral mode appear potentially important. Particles in the diameter range 5–100 µm are emitted with speaking at a volumetric rate that is several orders of magnitude higher than are small-particle emissions. These facts are all potentially important aspects for transmission of respiratory diseases: that emissions of large particles are high during speech, that such particles can travel room-scale distances, and that they can be inhaled and deposit within the (upper portion) of the respiratory tract. The emitted volume associated with ballistic drops is even larger, but these emissions generally do not travel more than a few m from the source. Ballistic drops also tend not to be inhaled, so additional steps in the source-to-receptor transfer process would be needed to complete the chain of transmission in this case.

5.3 Asymptomatic emissions

Evidence from the spread of COVID-19 suggests that a considerable proportion of the transmission occurs without symptoms, that is, without illness-induced sneezing or coughing.9, 71 Even for influenza, Lindsley et al.72 have suggested that emissions from illness symptoms may not dominate. “Because individuals breathe much more often than they cough, … breathing may generate more airborne infectious material than coughing over time.”

5.4 Incorporating viral load into emission estimates

Efforts to model the transmission of SARS-CoV-2 using bottom-up approaches have combined emission information of the type reported in Table 2 with viral load data.73 Viral load data regarding the number of viral RNA copies per volume of respiratory fluid have been reported for sputum samples,74, 75 for “posterior oropharyngeal saliva,”76 and from swabbing, reported as “throat samples.”75 Reviewing such evidence, Buonanno et al.73 concluded that “the concentrations of viral load in the mouth can reach values of 109 RNA copies/mL and occasionally up to 1011 RNA copies/mL during the course of the [Covid-19] disease.” For purposes of illustration, assume that 109 RNA copies/mL (=10−3 RNA copies/µm3) applies to fluid throughout the respiratory tract, and if one also assumes that the particle and drop emissions rates in Table 2 apply, then the emissions rate from speaking would be 2100 RNA copies per hour in the small particle size range (<5 µm), 8 million RNA copies per hour in the large-particle size range (5–100 µm), and 300 million RNA copies per hour in the drop mode (100 µm–1 mm). The significance of these numbers regarding the transmission of a SARS-CoV-2 infection would depend on the fate of the particles and drops after emission, the consequential exposures, and the resulting doses and susceptibility of specific sites for a receptor.

A large study that was undertaken during autumn 2020 at the University of Colorado, Boulder, provides important information about the viral load of SARS-CoV-2 saliva.77 During a 3.5-month period that predated the availability of vaccines and the spread of the delta variant, 72 500 saliva samples were collected on the university campus from individuals who were self-assessed to be asymptomatic. Using a quantitative RT-PCR assay, 1405 of these samples (2%) were found to be SARS-CoV-2 positive. Among the positive samples, the viral abundance varied quite broadly, over many orders of magnitude. Among the positive samples, 12% exceeded a viral load of 108 RNA copies/mL; 4% exceeded 109 RNA copies/mL; 1.5% exceeded 1010 RNA copies/mL; and 0.5% exceeded 1011 RNA copies/mL. Remarkably, the probability distribution of viral loads among the positive samples from this asymptomatic population was similar to the distribution constructed from literature reports for symptomatic (hospitalized) persons.

Information available about the spatial distribution of viral load in the respiratory tract is sparse. Sputum samples might reasonably represent the fluids lining the trachea and therefore correspond reasonably well to particles emitted in the LT mode. It is an open question to what extent swab sampling from the upper respiratory tract reflects the particles and drops emitted from the mouth during speaking or coughing. For the pulmonary region, sampling technologies are available to collect samples via bronchoalveolar lavage.78 However, this is a highly invasive procedure that would not be applied without medical necessity. That restriction would likely make inaccessible the direct determination of viral load in the respiratory fluid of the pulmonary tract for asymptomatic SARS-CoV-2 infections.

Emphasizing viral shedding from the upper respiratory tract may be appropriate. Gandhi et al.9 remarked on the “high level of SARS-CoV-2 shedding in the upper respiratory tract,” noting that “live coronavirus clearly sheds at high concentrations from the nasal cavity even before symptom development.” Hou et al.79 reported “a striking gradient of SARS-CoV-2 infection in proximal (high) versus distal (low) pulmonary epithelial cultures. … These findings highlight the nasal susceptibility to SARS-CoV-2 with likely subsequent aspiration-mediated virus seeding to the lung in SARS-CoV-2 pathogenesis.”

5.5 Emissions variability

Another important feature of the emissions evidence is the large variability. Regarding respiratory particles and drops, the information in Table 2 is intended to reflect average conditions. Emissions from speech increase strongly with loudness. “Furthermore, a small fraction of individuals behaves as ‘speech superemitters’, consistently releasing an order of magnitude more particles than their peers.”33 Viral load data indicate even greater variability, both across the population of subjects sampled and with time for any individual.74-76 He et al.71 studied the time course of viral shedding in 94 patients with COVID-19. They “inferred that infectiousness started from 12.3 days … before symptom onset and peaked at symptom onset.” The large variability in viral load and in respiratory emissions would contribute greatly to an uneven pattern of transmission risk.

6 PARTICLE DEPOSITION IN THE RESPIRATORY TRACT

Among the various factors influencing airborne transmission of respiratory diseases, deposition of inhaled particles in the respiratory tract may be the best understood. This topic has broad relevance and has been studied in many contexts, including radiological protection, air pollution toxicology, industrial hygiene, and drug delivery. The phenomenon is affected by a combination of the morphometry of the respiratory tract, the physiology of respiration, the fluid mechanics of breathing airflow, and the physics of airborne particle motion.

For ordinary breathing, it is entirely reasonable to assume that an inhaled particle will deposit if it strikes the lining of the respiratory tract. For particles in the size range of concern, from 0.1 µm to 100 µm in diameter, the primary means of airborne motion are advective flow, gravitational sedimentation, and inertial drift. The smallest particles in this size range are also influenced by Brownian motion (diffusion).

In industrial hygiene, particles are broadly classified by size as inhalable (<100 µm), thoracic (<~10 µm), and respirable (<~4 µm). Inhalable particles can be brought into the respiratory tract during normal inhalation. For particles in the vicinity of 100 µm or larger, only a portion is brought into the respiratory tract during inhalation because of inertial separation processes. Aspiration efficiency accounts for the size-dependency of inhalable particles, as displayed in the upper frame of Figure 3.80, 81 Thoracic particles can penetrate past the head into the tracheobronchial tree. Respirable particles can penetrate to the alveolar or pulmonary region of the lung.

Collection efficiency of particle samplers as a function of diameter specified for instruments that capture inhalable (upper), thoracic (middle), and respirable (lower) size fractions of airborne particulate matter, as defined by the American Conference of Governmental Industrial Hygienists92

With respect to airborne disease transmission, three key conceptual points can be inferred from Figure 3. First, particles up to about 100 µm in diameter can be brought into the respiratory tract through inhalation. Second, particles smaller than about 5 µm are respirable and therefore can penetrate to the alveolar region. Third, the different size categories are not sharply divided at their respective boundaries.

For particles in the diameter range 0.1–10 µm, Figure 4 displays modeled overall deposition efficiency in the respiratory tract along with subdivided efficiencies for three different respiratory tract zones: in the head or upper respiratory tract, in the tracheobronchial tree, and in the alveolar or pulmonary region. Overall deposition efficiency (upper trace) exhibits a minimum of about 13% for diameters in the range 0.3–0.5 µm. Smaller particles deposit more efficiently because of their higher diffusivity. Larger particles deposit more efficiently because they are more strongly influenced by inertial drift through the bending flow paths in the respiratory tract, and because of the influence of gravitational settling if they reach the deep lung. Where virus-containing particles deposit in the respiratory tract is important in relation to the susceptibility of these different regions to the initiation of an infection.

Deposition fraction as a function of particle size in the overall respiratory tract (upper), and in three different regions: in the head (2nd frame), in the tracheobronchial tree (3rd frame), and in the pulmonary region (bottom). For each frame, the normalization is based on the abundance of particles in air as it enters the respiratory tract. So, for example, for 10 μm particles, the low deposition in the pulmonary region occurs because such particles are deposited with high efficiency as air passes through the upper portion of the respiratory tract (the head) before reaching the pulmonary region. These results reflect predictions for a seated adult breathing through the nose, using a semiempirical model of the International Commission on Radiological Protection (ICRP) as reported by Hofmann.50 Note: read the vertical axes using left-hand scales for the top and 3rd frames, and right-hand scale labels for the 2nd and bottom frames

There is a noteworthy complementarity between regional emissions and regional deposition of particles in the respiratory tract. Large particles primarily emitted from the head will mainly be deposited in the head if inhaled. Small particles emitted from the pulmonary region have the potential to penetrate and deposit deep within the lungs of those subsequently exposed. If a viral infection is established regionally, with subsequent viral shedding from that region, the associated particles emitted from the infectious person can preferentially deposit in the same region of the respiratory tract of a susceptible individual. A part of this process is governed by size-dependent penetration and deposition patterns in the respiratory tract. Specifically, any large particles that might be generated in the lower respiratory tract have a low probability of escaping because inertial drift and gravitational settling will tend to cause them to deposit as they traverse from the distal airways to the mouth or nose.82

The total rate at which airborne particles are deposited in the respiratory tract can be estimated as the product of three terms: the airborne concentration, the volumetric breathing rate (also known as the minute volume), and the deposition fraction. Because concentrations and deposition fractions are size dependent, accurate analysis requires that this product be computed for each (typically narrow) particle size interval separately and then summed to determine the total. Handbook values from exposure studies can be applied for the volumetric breathing rate. For example, the USEPA’s Exposure Factors Handbook83 recommends using a mean inhalation rate of about 16 m3/d for adults, which corresponds to an average of 11 L/min. Values of about 5 L/min apply for sleep and sedentary/passive conditions, rising to about 12 L/min for light intensity activity and 25–30 L/min for moderate intensity. These specific values are somewhat age dependent and reflect averages for men and women combined.

7 ROOM-SCALE AIRBORNE TRANSMISSION

7.1 Overview

In outbreak investigations, evidence clearly points to indoor environments being much more important than outdoor settings for the transmission of SARS-CoV-2 infections.5 However, some key issues are not yet resolved. How important is large-particle inhalation compared with small particle inhalation? How important is inhalation near a source (within a few m) compared with room-scale transmission? Does the deposition of ballistic drops make a meaningful contribution to the spread of SARS-CoV-2 infections? These questions are important because the answers help guide the nonpharmaceutical public health interventions that are key to limiting the adverse consequences of the pandemic but are also disruptive to normal life. This section explores what indoor aerosol science can contribute to understanding airborne transmission of SARS-CoV-2 at the room or small building scale. In the next section, issues related to transmission in proximity will be explored. In these two sections, all indoor occupants are assumed to be unmasked and unvaccinated. The effects of interventions, such as masking, enhanced ventilation, and room air filtration, are considered in Section 9.

To frame the discussion, a few additional points should be made. First, some factors that affect transmission are highly variable. For these, the dominant portions of transmission may occur under conditions that are more favorable for spread, rather than under average conditions. So, for example, in assessing SARS-CoV-2 transmission risk, one might reasonably emphasize circumstances in which ventilation rates are within common ranges, but lower than average. Second, community spread of an airborne viral infection would commonly be associated with spaces where people congregate and socially interact. The evidence summarized in Table 2 illustrates that emission rates are much higher during speaking than from quiet tidal breathing. Respiratory particle emissions increase with speech loudness.33 Consequently, circumstances in which many people gather unmasked in crowded indoor environments and engage vigorously in speaking or other vocalization would be of particularly strong concern.

Two key limitations in the state of knowledge regarding indoor aerosol science should also be acknowledged. First, most aerosol science research has been focused on particles smaller than 10 µm in diameter. With public health and public policy emphasizing fine particles, there has been even greater attention devoted to the particles smaller than 2.5 µm in diameter. The state of understanding regarding the dynamic behavior of coarse particles (2.5–10 µm diameter) is less developed than for fine particles (<2.5 µm). The state of knowledge about large particles (5–100 µm) is even weaker. Second, quantitative assessments of room-scale transmission rely on the indoor air being relatively well mixed, so that the airborne concentrations are close to uniform spatially, at least when assessed on a time-averaged basis. The state of knowledge about air mixing in indoor environments is substantially incomplete,84-86 another fact that contributes to uncertainty in any assessment.

7.2 Wells-Riley model of infection risk

One widely used approach for modeling spread of airborne infection at the room or small building scale is based on the Wells-Riley model35 for assessing infection risk associated with inhaling droplet nuclei (diameter <5 µm). In this model framework, emissions are parameterized in units of infectious quanta, with inhalation of one quantum yielding a 63% (=1 − 1/e) chance of initiating an infection. Airborne concentrations of quanta are determined from a material balance model that typically assumes that the indoor air is well mixed. In the original model formulation, the material balance was based on assumed steady-state conditions. Removal of droplet nuclei from indoor air occurred by means of ventilation potentially augmented by particle filtration. Subsequent applications that have built on the Wells-Riley approach have allowed for time-varying indoor concentrations87 and have incorporated loss of droplet nuclei from indoor air by mechanisms other that ventilation and filtration, such as deposition and airborne inactivation of the infectious agent.25, 26 One of the substantial challenges in applying this approach is to determine quanta emission rates. Efforts in the case of SARS-CoV-2 include bottom-up calculations that combine respiratory particle emission data with viral load information73 and retrospective assessments of outbreaks.25, 26

The Wells-Riley modeling approach was devised and has mainly been applied for assessing transmission risk involving the inhalation of small particles (droplet nuclei). It is not well suited for evaluating infection risk associated with the inhalation of large particles. In common applications, well-mixed indoor conditions also have been assumed. That approach has not been used to address proximate exposure conditions.

7.3 Transfer efficiency approach

An alternative approach for investigating the transmission of infectious agents emphasizes transfer efficiency for respiratory emissions between infectious and susceptible persons. Broadly, such an approach is built around answering this question: what proportion of the material emitted from the respiratory tract of someone who is infectious is subsequently inhaled by a susceptible person? That proportion, combined with information on emission rate, would yield an estimate of the quantity inhaled. If combined with information about fractional deposition in the respiratory tract (as in Figure 4), then the quantity of infectious particles deposited in different regions of the respiratory tract could be computed.

This transfer efficiency approach has been applied to infectious disease transmission and to other indoor air quality concerns. One quantitative metric is the intake fraction, which in this context would represent the fraction of an emission that is inhaled by all exposed persons.88 Other closely related efforts have been published. Rudnick and Milton89 explored exposure to metabolic CO2 as a means of quantifying the rebreathed fraction, that is, the proportion of air inhaled that would have come from the exhalations of other building occupants. Melikov et al.90 used the intake fraction concept to quantify the benefits of intermittent occupancy for reducing the risk of airborne transmission. Bond et al.17 introduced the idea of the effective rebreathed volume and explored how that measure could specifically address the differential influence of particle size on infection risk.

Table 3 presents parameters for three example categories of indoor spaces that are commonly occupied and have either long occupant duration (residences) or high occupant density (classroom and restaurant). For each category, data are presented for a base-case condition and for a higher-risk condition. The parameter values were selected to represent conditions that were common in the United States prior to the pandemic. These parameter values will be used in the assessments to follow as illustrations of the influence of occupant density and ventilation rate on transmission risk. The emissions considered will emphasize the asymptomatic activities of breathing and speaking, as described in Table 2.

TABLE 3.
Indoor environmental parameters influencing airborne disease transmission risk

Environment (Condition)	Occupants	Air-change rate (h−1)	Vent. rate (L/s pers−1)
Residence (Base case)	4	0.5	8.7
Residence (High exposure)	12	0.2	1.2
Classroom (Base case)	25	2.5	6.7
Classroom (High exposure)	25	1	2.8
Restaurant (Base case)	60	4.3	5.0
Restaurant (High exposure)	60	1	1.2

7.4 Intake fraction for inhalable particles

For well-mixed indoor environments, with either sustained occupancy or under steady-state conditions, the aggregate intake fraction, iF, satisfies the following expression91(2)

Here, N is the number of occupants of the indoor space. In calculations to follow, it is assumed that one person is infectious and that all others are susceptible, so N − 1 represents the number of susceptible persons. The parameter Qb is the average volumetric breathing rate of each exposed person, λtot is the total pollutant removal rate constant, and V is the interior mixed volume. The total removal rate constant would reflect the sum of contributions from ventilation, deposition, filtration, and any inactivation of the airborne virus. For the purposes of evaluating iF, the possibility of resuspension following deposition is not considered. For the explorations considered here, no contributions from inactivation are included in the analysis; the total loss-rate coefficient is taken to be the sum of the air-change rate, plus the deposition loss-rate coefficient, plus the rate of loss by air filtration, when present. The per-person average volumetric breathing rate is assumed to be Qb = 10 L/min. The air-change rates are those values reported for the six cases in Table 3. The deposition loss rate coefficients are as presented in Figure 2. The analysis only considers particle sizes up to 50 µm diameter, because of the high degree of uncertainty about deposition and mixing for large particles in the diameter range 50–100 µm.

Figure 5 displays the estimated intake fractions for the six cases. The strong particle size dependence for large particles (diameter ≥ ~5 µm) is attributable to the high deposition loss rate coefficient for the bigger inhalable particles. Focusing on particles in the submicron mode, the intake fractions vary over an order of magnitude across exposure conditions, from about 9 per thousand (0.9%) for the base-case conditions in a residential setting to about 110 per thousand (11%) for the high-exposure case in a restaurant.

Aggregate intake fraction as a function of particle diameter for emissions of respiratory particles in three different settings (restaurant, classroom, and residence) and for two different exposure conditions (base case and high exposure). The vertical axis in each case represents the estimated proportion of the emitted particles that are cumulatively inhaled by the susceptible occupants, with a value of 10 (per thousand) indicating that 1% of the emitted particles would be inhaled. See Table 3 for assumed parameter values. The steep decrease in intake fraction with increasing particle diameter is attributable to the much higher deposition rate of the larger particles, as displayed in Figure 2

Equation (2) and the results presented in Figure 5 do not take into account aspiration efficiency, that is, that only a fraction of large particles can be inhaled (see upper frame of Figure 3). Given the orders-of-magnitude variability in intake fraction with particle size as displayed in Figure 3, the effect of aspiration efficiency is relatively small: estimated relative reductions in intake would be 23%, 42%, and 48% for 10 µm, 30 µm, and 50 µm particles, respectively.92

7.5 Estimating infection risk from intake fractions

To obtain an estimated quantity that is closer to infection risk, the intake fraction results can be combined with emission factors to yield total intake rates. Table 4 presents a summary of such estimates, in which the intake fractions as displayed in Figure 5 are combined with emissions data from Table 2. Two emission cases are considered, one with the infectious person only breathing and the second case with the infectious person speaking. The time scale of 1 h is selected to represent, in magnitude, the duration of a gathering in a home (the high-exposure case for a residence), the duration of a university lecture, or the time spent at a meal in a restaurant.

TABLE 4.
Estimated inhalation intake rates (µm3/h) of respiratory particles from an infectious emitter for several illustrative indoor environmental conditions

Exposure conditions	Susceptible persons (N − 1)	Particles from breath		Particles from speaking
		Small (0.1–5 µm)	Large (5–50 µm)	Small (0.1–5 µm)	Large (5–50 µm)
Residence (base)	3	180	0.13	6.6 × 103	4.5 × 104
Residence (high)	11	1020	0.48	3.0 × 104	1.6 × 105
Classroom (base)	24	480	0.81	2.5 × 104	3.5 × 105
Classroom (high)	24	960	0.95	4.1 × 104	3.5 × 105
Restaurant (base)	59	740	1.7	4.4 × 104	8.3 × 105
Restaurant (high)	59	2360	2.3	10.0 × 104	8.7 × 105

Several features of the results merit discussion. First, although not displayed in Table 4, it is important to note the emission modes associated with the small and large particles. As seen in Table 2, only the B modes contribute to emissions from quiet breathing. These modes represent particles originating in the distal airways from bronchiolar fluid film bursting. Even on a volume basis, these modes are dominated by small particles. For speaking, most particles in the small size range originate from the LT mode, representing the larynx and tracheal region, with most of the remainder associated with the B modes. Almost all the large particles from speaking are associated with the O modes, originating in the oral cavity. For emissions from speaking, there are substantial contributions to inhalation intake in both the small- particle and large-particle size ranges. The intake rate in both size ranges is much larger than when emissions only originate from quiet breathing.

A second important feature is revealed in pairwise comparisons of the high-exposure versus base-case conditions. For small particles, the high-exposure circumstances produce substantially higher aggregate exposures for all pairs, with high:base ratios ranging from 1.6× (small particles from speaking in a classroom) to 5.7× (small particles from breathing in a residence). By contrast, for the large-particle exposures, there is little difference between high and base-case conditions for the classroom and for the restaurant. The dominant removal mechanism for large particles is deposition, which is assumed to be consistent across indoor environments (although strongly dependent on particle size). Consequently, the influence of ventilation rate on large-particle concentrations is small. With no change in assumed occupancy, the inhalation intake of large particles is not strongly influenced by the high versus base-case assumptions. By contrast, a difference is seen in the large-particle intake between the high and base-case results for residences; that difference is attributable to the higher assumed occupancy for the high-exposure case.

Another important point to stress regarding the risk of transmission is not fully evident from Table 4. Occupancy levels can have a second-order or quadratic influence on the risk of airborne transmission. The Table 4 data illustrate one factor: the volume of air inhaled by susceptible persons scales linearly with the number of occupants. A second factor not displayed is that the likelihood of an infectious person being present would also tend to scale in proportion to the number of occupants. Consequently, by reducing occupancy levels by 50%, the risk of transmission would tend to be reduced by 4×, for conditions in which the well-mixed indoor air model apply.

The evidence in Table 4 suggests that, depending on circumstances, the cumulative inhalation rate of respiratory particles that are emitted from one occupant can be in the approximate range 103–106 µm3/h. Is this rate high enough to make a meaningful contribution to the transmission of SARS-CoV-2? Recall that in their bottom-up modeling calculations, Buonanno et al.73 found that “the concentrations of viral load in the mouth can reach values of 109 RNA copies/mL.” This viral load was exceeded in 0.08% of the saliva samples (4% of the positive samples, which were 2% of those collected) collected from asymptomatic subjects in the University of Colorado study.77 A unit conversion shows that 109 RNA copies/mL corresponds to 10−3 RNA copies/µm3. Combining the results, emissions from breathing only might yield a cumulative inhalation intake of ~1 RNA copy per hour for the exposure conditions represented in Table 4, whereas speaking could produce cumulative intakes in the approximate range 10–1000 RNA copies per hour. Buonanno et al.73 cited evidence from studies of SARS-CoV to suggest that an infectious quantum might be equivalent to 10–100 RNA copies. Consequently, this exploration indicates that speaking could meaningfully contribute to the risk of infectious spread of SARS-CoV-2 indoors, whereas quiet breathing would generally not.

An important caution must be stressed. The finding of higher intake rates for large particles compared to small particles for emissions from speaking does not automatically imply that the large particles dominate the risk of infection. Large particles originate mainly in the mouth and will deposit predominantly in the head when inhaled. Small particles originate from throughout the respiratory tract and can also deposit throughout, including in the pulmonary region. The source location can influence the viral load in respiratory fluids and therefore the abundance of virus in emitted particles. Although not known, SARS-CoV-2 might exhibit infection isotropy such that the risk of infection for a given amount of deposited viral RNA varies with location in the respiratory tract.

8 AIRBORNE TRANSMISSION IN CLOSE PROXIMITY

8.1 Overview

A key unresolved topic regarding airborne transmission of SARS-CoV-2 is the importance of proximity. How much does it matter that a susceptible person is within conversational distance (<1–2 m) of an infectious person? A classical view of airborne transmission involving aerosol inhalation assumes that it occurs over room scale, incorrectly conflating near-field transmission with droplet or contact transmission mechanisms.6 Part of the misunderstanding arises from misclassifying the size that divides inhalable airborne particles from ballistic drops, as has already been discussed. Exposure to ballistic drops, that is, those larger than about 100 µm, would occur predominantly near the source. Most ballistic drops do not travel very far before depositing on surfaces because of their high inertia and high settling velocities. One should also expect some near-field enhancement of small (0.1–5 µm) and large (5–100 µm) particles, simply because of the time required for advective and turbulent transport to distribute the emissions throughout the indoor space. But evidence is not sufficient to fully discern the extent of concentration enhancements for inhalable particles when a susceptible person is close to an infectious source.

Chu et al.93 conducted a systematic review and meta-analysis of the evidence regarding physical distancing and the risk of viral transmission. They concluded that “transmission of viruses was lower with physical distancing of 1 m or more, compared with a distance of less than 1 m …; protection was increased as distance was lengthened ….” Two aspects of their assessment should be highlighted. First, the scope of their investigation included studies of SARS-CoV and MERS-CoV in addition to SARS-CoV-2. Second, most of the studies reviewed were from healthcare settings rather than reflecting community spread. The authors did indicate that “the association was seen irrespective of causative virus … [and also irrespective of] health-care setting versus non-health-care setting.”

In this section, relevant mechanistic evidence about the influence of proximity in airborne viral transmission is assessed. Evidence regarding person-to-person transmission from experiments and from numerical modeling is presented and critically evaluated. This section also considers important qualitative features of the fluid mechanics of airflow from exhalations, the buoyant thermal plume of building occupants, and general indoor air motion. The role of proximity influencing risk from ballistic drops is undisputed: such drops pose their primary impact when a susceptible person is within 2 m of an infectious source. However, the cumulative evidence is not conclusive regarding transmission via large or small particles. The evidence to be reviewed suggests that enhanced near-field exposure via inhaling large particles can make a meaningful contribution to the overall transmission risk. Conversely, the evidence does not point clearly toward a strong excess contribution of proximity for the transmission risk associated with inhaling small particles.

8.2 Experimental investigations of person-to-person transfer of respiratory exhalations

Direct experimental investigations of tracer gas flows from person to person provide valuable insight into the potential for airborne transmission. Olmedo et al.94 reported one such study, employing two breathing thermal manikins in a 35-m3 chamber operated at an air-change rate of 5.6 h−1, which corresponds to a per-manikin ventilation rate of 27 L/s, that is, a well-ventilated condition. Air motion in the room was also influenced by natural convection: a 488 W heat load was provided by a combination of a radiator (300 W) plus the two manikins (94 W each). The mean room air temperature was 22°C, and the air supply temperature was 16°C. One experimental configuration used a common mixing ventilation design, with a supply register at the ceiling and two return registers at the top of one wall. With cool air supply and a high rate of heat dissipation in the room, mixing would occur rapidly. Indeed, the concentration of tracer gas in the breathing zone of the receptor manikin was only slightly higher than the concentration in room exhaust air (difference of <5%) for a case with the manikins standing face to face and separated by distances in the range 0.35–1.1 m. The receptor manikin was breathing at a rate of about 10 L/min. Given spatially uniform concentrations, the intake fraction for a nonreactive tracer would be the ratio of the inhalation flow rate to the ventilation exhaust flow rate, or 10 L/min/(54 L/s × 60 s/min) = 3.0 per thousand (0.3%).

When considering exposure in proximity, an important feature is that, at least for direct transmission, the transport time scale (on the order of seconds) from source to receptor is very much shorter than the time scale for particle loss by deposition. It is worthwhile to make an upper-bound estimate of intake considering the possibility that deposition losses between source and receptor might be negligible when the two persons are in proximity.

Table 5 compares intake rates of inhalable particles for two conditions based on the Olmedo et al. experiments with mixing ventilation. In the upper-bound case, it is assumed that there are no depositional losses. In the lower-bound case, a well-mixed model representation of the room is assumed with deposition losses incorporated as in the previous section. Note that the difference between the two bounds is big for the large-particle case (16× increase) whereas it is only moderate for intake of small particles (45% increase).

TABLE 5.
Bounding estimates of the inhalation intake rate (µm3/h) of respiratory emitted particles for mixing ventilation configuration investigated by Olmedo et al.

Exposure conditions	Particles from speaking
	Small (0.1–5 µm)	Large (5–50 µm)
Upper bound (no deposition)	6400	1.54 × 106
Lower bound (well-mixed)	4400	9.8 × 104

Another noteworthy feature of the findings of Olmedo et al.94 is their sensitivity to specific experimental conditions. As one example, in addition to measuring exhaled tracer gas in the breathing zone of the exposed manikin, they also measured concentrations at the chest and in the thermal plume 10 cm above the head (termed “C10”). The concentrations at chest level were similar to those in the breathing zone, as were C10 values when the spacing between manikins was ≥0.8 m. However, for closer spacing, the C10 value moderately increased, reaching 35% above that in the breathing zone with 0.35 m spacing between manikins.

Much larger effects were seen when a displacement ventilation configuration was used instead of mixing ventilation. In this case, cool ventilation supply air was provided at the floor and air was exhausted from the ceiling. Strong thermal stratification was established, with buoyant warm air aloft and dense cooler air below. Such stratification impedes vertical mixing, so the exhaled air from the source manikin could be effectively trapped in a horizontal band of air close to the emitted height. Under these conditions, for some source-receptor placements, the breathing zone concentration of tracer gas at the receptor manikin was as much as an order of magnitude higher than the concentration in the room exhaust air. Substantial spatial gradients with enhanced breathing zone levels were also observed in a test with ventilation only through an open doorway (no mechanical system operating). The enhancements were substantial when the manikins were spaced at 0.35 or 0.5 m distance, but not when the spacing was 0.8 or 1.1 m.

Chen et al.95 reviewed the work of Olmedo et al.94 along with similar experimental studies of exposure to exhaled contaminants in mechanically ventilated rooms. That assessment specifically aimed to differentiate what they termed as “direct exposure” from “indirect exposure.” Direct exposure occurred when the “exhaled jet from the infected person directly enters the breathing zone of the target person.” Their review identified 191 experimental cases cumulatively reported in 10 research papers of which 133 (70%) used mixing ventilation conditions and 58 used displacement ventilation. As in Olmedo et al.,94 the results were quantified in terms of a concentration ratio of tracer, dividing the breathing zone level (Cexp) by the concentration in the room exhaust (CR). Chen et al.95 defined a threshold concentration ratio of Cexp/CR ≥ 1.2 as indicating a significant contribution from direct exposure. Note that at the lower bound this condition represents only a 20% increment in exposure above well-mixed conditions. Using this threshold, they reported that 28/133 (21%) of cases with mixing ventilation had significant direct exposure as did 30/58 (52%) of the cases with displacement ventilation. In the reviewed studies, the distance between source and receptor varied in the range 0.3–3 m. Air-change rates tended to be high, mostly in the range of 4–12 h−1. Per-person ventilation rates were consistently high across all the reviewed studies, in the range 20–110 L/s pers−1. At the 1.0 m scale, a clear effect of interpersonal spacing was reported: “When the inter-person distance was greater than or equal to 1.0 m, the normalized exposure for most of the cases was lower than 1.1.” Conversely, “when the inter-person distance was less than 1.0 m, there were many cases with a normalized exposure significantly larger than 1.0 (as high as 13.0).”

Although substantial, this body of experimental work is not sufficient to characterize the importance of proximity influencing inhalation exposure to respiratory particle emissions. The experimental configurations all featured high ventilation rates and high air-change rates that are typical for well-designed healthcare environments but are less representative of community conditions where people congregate and socialize. The experiments were conducted typically with just two breathing thermal manikins to represent occupants. The manikins were stationary. The fluid mechanics of emission events were represented by breathing only (except for several experiments simulating coughing in Liu et al.96), most commonly with exhalation through the mouth rather than the nose. Some of the investigations explicitly targeted healthcare environments, such as assessing the transmission risk between patients in a two-bed ward.97, 98 None of these studies can be taken as specifically representative of conditions that would prevail for people sharing a meal at a table in a restaurant, or while socializing in a crowded nightclub. The high degree of variability in the results suggests a need for caution in extrapolating beyond the conditions studied.

A caution also must be expressed about interpreting a concentration ratio, such as Cexp/CR, as a performance figure for assessing exposure risk. In investigations with high ventilation rates, the denominator in the ratio becomes small, leading to a perceived enhancement of risk associated with close proximity. However, these chamber experiments also show that high ventilation rates produce relatively small concentrations except under some quite close conditions. The highest Cexp/CR ratios, roughly 10 for some cases with proximity at a scale of ~0.5 m, indicate that the very near-field exposure is 10× the value that would prevail in a well-mixed room. However, with a high per-person ventilation rate as in these experiments, the well-mixed room condition might be as much as 10× lower than in a poorly ventilated space. Consequently, another interpretation of the proximity effect from these experiments could be this: being very close to an infectious person in a well-ventilated space has comparable risk to sharing occupancy in a poorly ventilated room, even when separated by 2 m or more.

The first detailed study of person-to-person transfer using breathing thermal manikins was reported by Bjørn and Nielsen.99 They focused on assessing displacement ventilation systems, which were being implemented with the idea of improving the overall efficiency of pollutant removal in occupied spaces and thereby improving the energy efficiency of ventilation systems. That study contains important qualitative insights. For example, the authors note that “exhalation does not necessarily follow the boundary layer flow close to the body but is able to ‘break free’ and penetrate the breathing zone of other persons.” This effect would be enhanced in the presence of thermal stratification associated with displacement ventilation. However, Bjørn and Nielsen99 go on to say that “this is probably not a problem for most practical ventilation applications. People rarely stand 0.4 m apart, breathing through the mouth directly into each other’s faces. … The most common breathing mode is through the nose, while sitting apart at some distance, perhaps facing each other’s backs, e.g., in an auditorium or concert hall, in public transportation, etc. The experiments show that these situations are not critical. At mutual distances larger than 1 m, the phenomenon is losing its importance….”

All experiments reviewed by Chen et al.95 used a tracer gas (N2O) to mark the exhaled breath from the source manikin. In a complementary effort, Liu et al.96 used a numerical modeling approach to assess the behavior of particles up to 100 µm in diameter. A limitation is that the assessment still applied only for the fluid mechanics of emissions by breathing. Recall, as summarized in Table 2, one expects particles of 100 µm diameter and larger to be emitted from the mouth during speech, but not from quiet breathing. In each simulation, the positions of 1600 particles were numerically tracked after simulated release from the source. For the 100-µm particles, at manikin spacings of 0.5 m, 1.0 m, and 1.5 m, respectively, the total number depositing anywhere on the receptor manikin surface were 99 (6.2%), 91 (5.7%), and 13 (0.8%). The proportion of those depositing that landed on the face was in the range 1–10% and none specifically deposited on exposed mucosae. The mean ± standard error of numbers of particles inhaled at the three distances was, respectively, 3.0 ± 0.5, 9.0 ± 2.8, and 0.3 ± 0.3. This evidence supports the idea that a proximity effect enhances exposure when the spacing between source and receptor is about 1 m or less. Note that the inferred intake fraction by inhalation for the three spacings would be approximately 0.2%, 0.6%, and 0.02%, respectively. For comparison, assuming well-mixed conditions and no proximity effect, Nazaroff91 reported, “that a typical inhalation intake fraction associated with release of a nonreactive contaminant into a US residence would be ~4000 per million [0.4%].” Note that 100-µm particles settle very rapidly. Deposition loss strongly attenuates exposures for this particle size in a well-mixed room. In sum, the simulations by Liu et al.96 suggest that proximity might contribute materially to an increase in inhalation exposure to large particles when source-receptor separation distances are less than about 1 m.

8.3 Airflow conditions influence proximate transmission

The potential influence of proximity on exposure depends crucially on fluid flow conditions. Three interacting aspects merit attention: the fluid flow induced by the expiratory event from the infectious source; the airflow in the vicinity of the susceptible person; and the background airflow in the indoor environment. Details of such flows exhibit complexity that is well beyond the scope of this article. So, the emphasis will be to highlight some major features with pointers to the literature as a starting point for the reader interested in deeper study.

The detailed investigation of fluid flows associated with exhalation and other expiratory events is a recent development, with most of the relevant studies having been published in the last few decades. Murakami100 provided an early review of the application of computational fluid dynamics to the study of “the microclimate surrounding the human body.” His analysis documented the asymmetry of airflows induced by breathing. Exhalation produces a cone-shaped jet that can travel tens of cm with airspeeds attenuating as more air is entrained. Inhalation produces a more nearly uniform hemispherical converging inflow field with smaller extension from the face. Such asymmetry is also discussed by Abkarian et al.101

8.3.1 Airflow conditions in relation to emissions

Schlieren and shadowgraph techniques have been applied for qualitative investigation of airflow fields induced by emissions from the respiratory tract, including coughing, sneezing, breathing, whistling, laughing, and talking.53, 54, 102 These investigations were motivated by interest in the spread of respiratory infectious diseases. Based on their imaging work, Tang et al.54 remarked that “typical conversation at distance on the order of 1 m apart appears to be safe for much of the time, but … individuals talk in very different ways with a large variety of airflow patterns, even when speaking the same words.” Xu et al.103 extended these imaging techniques to generate quantitative assessments of “turbulent exhaled airflow from 18 healthy human subjects whilst standing and lying.” They reported peak velocities during exhalation that were measured at 3 cm from the nose or mouth. Individual values (depending on subject, whether mouth or nose breathing, and whether standing or lying) spanned about an order of magnitude, from 0.3 to 3 m/s. Mean values across subjects for the different modes of breathing and body positions spanned about a factor of 2, from 0.8 to 1.8 m/s. The authors also reported that the exhaled centerline airspeed diminished to below 0.1 m/s at a distance of ~0.4 m from the subject. At this speed, background room air motion would commonly begin to dominate.

Gupta et al.104 measured the basic flow properties of coughs. Their study subjects were 25 healthy volunteers (12 female, 13 male). The researchers found considerable variability across the subjects and concluded that “cough flow characteristics from a subject cannot be used to represent the whole population.” Ranges of reported values are reproduced here, separately for male (M) and female (F) subjects: cough peak flow rate (3–8.5 L/s M; 1.6–6 L/s F) and cough expired volume (0.4–1.6 L M; 0.25–1.25 L F). A subsequent study used similar methods to assess airflows from breathing and talking.105 For a subject reading a passage, over a 2-min period, the volume inhaled and exhaled was about 27 L (13.5 L/min), only modestly higher than the average flow rate from breathing quietly (12 L/min for that subject). Significantly, while speaking, although 86% of the volume inhaled was through the nose, 85% of the exhaled air passed through the mouth. Even for habitual nose breathers, much of the exhaled air when speaking would be from the mouth, a feature that would influence near-field airflow characteristics. The evidence from this study also demonstrates that speech produces highly irregular exhalation flows, with short-term rates up to 1–2 L/s.

Bourouiba et al.106 evaluated the flow fields induced by uncovered coughs and sneezes. Their study included evaluations of the dynamic behavior of suspended aqueous particles, as influenced by momentum, drag, and evaporation. Chen et al.55 studied the fluid flow aspects of coughing with the mouth covered. They reported that “covering a cough with a tissue, a cupped hand, or an elbow can significantly reduce the horizontal velocity and cause the exhaled particles to move upward with the thermal plumes generated by human bodies.”

The complexity of airflow fields induced by breathing was studied by Xu et al.107 For exhalation through the mouth, the peak velocity averaged across many subjects, measured at 3 cm distance, varied between about 0.6 and 1.0 m/s. The influence of exhaled air on CO2 concentration could be observed out to about 35 cm from the mouth. Similarly, the influence of exhalation on airspeed was discernable out to a range of about 30–40 cm from the mouth. A later study103 included nose breathing, which generated somewhat higher average peak airspeeds (at 3 cm distance), varying between 1.1 and 1.8 m/s.

Abkarian et al.101 recently reported on detailed airflow investigations associated with breathing and speaking, referring to this subject as “linguistic aerodynamics” or “aerophonetics.” Their work highlights the importance of sequential plosives in conversational speech. They find that “exhaled materials reach 0.5–1 m in 1 s during normal breathing and speaking.” Furthermore, “airflow speeds at 1- to 2-m distances from a speaker are typically tens of centimeters per second.” They caution, however, that their work does not account for “movement of the head or trunk of the speaker and the influence of background motions of the air due to ventilation.”

8.3.2 Airflow conditions in relation to inhalation exposure

In considering the risk of airborne transmission of a respiratory virus when source and receptor are in proximity, airflow conditions around the susceptible person are also important. A major feature of indoor air flow near people is the convective boundary layer, established because of metabolic heat generation combined with the buoyancy of warm air.

A standard heat generation measure is the metabolic unit, or met, with 1 met = 58 W/m2 corresponding to the condition of being seated and relaxed. With a typical body surface area for an adult of 1.8 m2, the at-rest metabolic energy production would be 104 W. That energy is dissipated through a combination of radiant heat transfer, evaporation of water, and convective heat transfer. Worth noting is that a 2000 (kilo)calorie per day diet corresponds to an average power dissipation rate of 8.4 MJ/d = 97 W.

Substantial air flow is associated with the personal convective boundary layer. Licina et al.108 cited previous work to suggest that as much as 60 L/s of volumetric flow was entrained into the plume above a standing person. The airspeed profile reported by Gena et al.109 above a thermal manikin suggests a similar value. Craven and Settles110 reported thermal plume flow rates above a human to be in range 20–35 L/s for a linearly stratified room environment. The peak air speed above the head can be as high as 23–25 cm/s.100, 109 Note that the volumetric flow rate in the convective boundary layer is higher than typical outdoor air ventilation rates (generally on the order of 10 L/s per person). It is very much higher than the volumetric flow rate associated with breathing. In near-field exposure conditions, emissions from the respiratory tract of an infectious person will be diluted both by entrainment of airflow in the exhaled jet and by the movement of air in the convective boundary layer of the susceptible person. Sun et al.111 reviewed the personal convective boundary layer (which they term the “human thermal plume”), in the context of airborne transmission of SARS-CoV-2. The interaction of the personal convective boundary layer with room air flows induced by ventilation systems has been explored experimentally112 and numerically.113

8.3.3 Significance of room airflow conditions for proximate transmission

Although substantial research efforts have aimed to understand indoor airflow conditions, these studies have tended to stress idealized circumstances, for example, evaluating ventilation system concepts or modeling system performance in small rooms with simple boundary conditions and low occupancy levels. As observed in a recent review,114 “air flow patterns within a space are crucial for determining the distribution, transport and fate of any airborne contaminants. Predicting these flow patterns is extremely challenging since they depend critically on both the boundary conditions … and on the internal dynamics of the fluid, particularly buoyancy forces associated with temperature differences.” Bhagat et al.114 address the influence of people on indoor air motion in relation to the transmission of airborne infectious diseases. Themes discussed include the exhaled jet, the thermally buoyant plume, and the wake flow behind a moving person. To this latter point, the authors note that “wake velocity is approximately 80% of the person speed, implying [that] flows behind a person [are] of the order of 0.8 m/s [and] are possibly the largest in a space.” With walking, the body sheds its thermal plume, so that forced convection replaces natural convection in removing the metabolically generated heat. In their summary, Bhagat et al.114 write that “room flows are ‘turbulent’ in the sense that spatiotemporal variations of the flow are larger than the mean flow. They take place in complex geometries where the placement and sizes of inlets and outlets determine overall flow patterns, superimposed on which are significant perturbations associated with often transient events such as the movement of occupants, the opening and closing of doors and … variations in the external conditions.” In the context of airborne infectious disease transmission, the complexity of airflow conditions, from the source, near the receptor, and throughout the indoor environment represents only one important axis in a system with multidimensional complexity.

8.4 Possible transmission involving two successive airborne stages

Not widely appreciated in the contemporary literature concerned with infectious disease transmission is the possibility that viral transfer from source to receptor could involve two successive airborne stages. To envision such a process for SARS-CoV-2, consider two important points. First, the virus can remain viable on surfaces for periods of hours to days.115, 116 Second, particles that deposit on surfaces can become resuspended. This latter point has been substantially investigated for abiotic particles.117-119 Regarding microbes, Duguid and Wallace120 highlighted the “bacterial contamination of air produced by liberation of dust from the skin and personal clothing during bodily movement.” They reported “Experiments with nasal carriers of Staphylococcus aureus showed that the air was infected with this pathogenic organism more regularly and to a greater degree by the liberation of dust from clothing than by sneezing.” Licina et al.121 have reviewed the role of clothing as an intermediary contributor to microbial exposures. Prior to onset of the COVID-19 pandemic, Stephens et al.122 reviewed fomites more broadly, including their role in transmission of infectious viral agents. Asadi et al.123 recently demonstrated that viral particles shed from the body of a guinea pig could infect a susceptible partner in a separate cage through airborne transmission. This evidence points to the plausibility of this scenario for SARS-CoV-2 transmission: deposition of ballistic drops or large particles on clothing, followed by subsequent release from the clothing through ordinary movement, transport to the breathing zone assisted by the personal convective plume, inhalation intake, and deposition in the upper respiratory tract. If ballistic drops are involved in this process, one should recognize that the initial deposition and persistence on a clothing surface would allow for drying and shrinkage to full equilibration with local environmental conditions. Also plausible is that the movement of fabric fibers would cause fragmentation of the drop residue, producing inhalable particles from larger emissions. The process of respiratory drops becoming fractured and released from surfaces after depositing has not been specifically studied. However, at least at the level of “proof of principle” there may be parallels with the phenomenon of “thunderstorm asthma,” which involves the environmental rupture of pollen grains to produce respirable allergenic fragments.124

9 MITIGATING AEROSOL TRANSMISSION

Technical measures to reduce indoor exposure to airborne particles commonly target one of three control points: emissions, airborne transport, and receptors. These three control points provide opportunities for mitigating the aerosol transmission of SARS-CoV-2. Masks are used both to reduce respiratory emissions and to attenuate inhalation intake. Air filtration and enhancing ventilation rates to reduce indoor concentrations of infectious aerosols can contribute to reducing the transmission risk. This section describes these control measures, presents illustrative quantitative examples, and highlights benefits and limitations.

9.1 Masks

The wearing of masks in public, already common in some countries, became a worldwide phenomenon during the COVID-19 pandemic. For protecting against community transmission of infectious diseases, the scientific foundation for their use is incomplete. Many articles have been published recently that report on new evaluations. This section presents some highlights on the current state of knowledge about the degree to which masking is effective in limiting SARS-CoV-2 transmission. The section includes quantitative assessments of speech-associated emissions reduction as well as intake reduction in classroom settings.

By way of background, in an editorial published a decade ago, Li outlined fundamentally important points about masking.125 He also described some important limitations to our understanding. That editorial was inspired by the experience in Hong Kong and elsewhere during the 2003 SARS epidemic and the 2009 H1N1 influenza pandemic. The editorial is remarkably prescient for issues relevant to the COVID-19 pandemic.

Three important factors determine the efficacy of masking for limiting the spread of respiratory infectious agents such as SARS-CoV-2. The first and second factors are that masks remove particles from airstreams, reducing their release from infectious persons in exhaled air and reducing their intake from air inhaled by susceptible persons. A key feature in the control benefit on both the source and receptor side is that filter efficiency is a strong function both of particle size and of mask quality. This aspect interacts with the important yet inadequately understood element regarding the sizes of particles that are responsible for airborne transmission. The third key factor is that the mask “stops the exhalation puff of the wearer from being directly injected into air, instead redirecting it into the body’s thermal plume.”125 This third aspect is potentially important for protecting against transmission when the infectious and susceptible persons are proximate. Visualization studies nicely demonstrate the influence of masking in attenuating the airspeeds of exhalations.53, 54, 126

Three tiers of face coverings and masks have been commonly used in community settings during the SARS-CoV-2 pandemic. The most highly protective are N95 respirators or their close relatives (such as KN95 and FFP2). The second tier is the surgical mask. The third tier is a face covering in which one or more layers of fabric (or other porous or permeable materials) are adapted in such a way as to cover the mouth and nose of the wearer. Prior to the SARS-CoV-2 pandemic, published research on the performance of masks focused on N95 respirators and surgical masks and emphasized particle sizes smaller than 5 µm diameter. From this earlier work, good examples of the laboratory evaluation of mask efficacy have been published.127, 128 Cowling et al.129 reported on a systematic review of the efficacy of mask wearing to limit influenza transmission in the wake of the H1N1 pandemic. They concluded that “further studies in controlled settings and studies of natural infections in healthcare and community settings are required to better define the effectiveness of face masks and respirators in preventing influenza virus transmission.” Face coverings are also used for protection against exposure to particulate air pollution. Shakya et al.130 assessed the efficiency of several types of cloth masks and a surgical mask against particles in the size range 0.03–2.5 µm.

Leung et al.131 tested the effectiveness of surgical face masks on the emission of viral particles from subjects with acute respiratory illness. Respiratory emissions were collected for 30-min sampling periods during quiet breathing; natural coughing was allowed and recorded. The sampling was size segregated into small (<5 µm) and large particles, without a clearly specified upper bound for the large particles. For subjects without a mask and infected with a coronavirus (but not SARS-CoV-2), viral RNA was identified in large particles for 3 of 10 (30%) subjects and in small particles for 4 out of 10 (40%) subjects. The researchers “did not detect any virus [in either large or small particles] from participants wearing face masks.” The effect of masking in reducing emissions was smaller for subjects infected with influenza virus, and no significant differences were observed with masking for emissions from those with rhinovirus infections. An important experimental design detail should be noted in the context of community transmission of SARS-CoV-2: these tests did not involve the subjects speaking.

Three contemporary studies motivated by the COVID-19 pandemic demonstrate progress but also illustrate challenging aspects in advancing knowledge about the efficacy of masks and face coverings. Asadi et al.132 used human subjects to measure outward emission of respiratory particles with and without masks during speaking and coughing. They tested N95 and KN95 respirators, surgical masks, and several cloth face coverings. They measured size-distributed emissions in the approximate diameter range 0.3–20 µm. Their primary reporting was in terms of the change in particle number concentration sampled, summed across the entire size range measured. That choice might indicate lower performance than would a volume-weighted effectiveness measure. The sampling approach collected only a subset of the emitted particles, those projected forward into a sampling cone. A higher proportion of emitted particles may have escaped detection during experiments with masks because of the altered flow field. This study also revealed the interesting finding that movement of the jaw against mask material generated airborne particles. The authors attributed this observation to “friable cellulosic fibers in homemade cotton-fabric masks.” Earlier literature, as reviewed by Licina et al.,121 has shown that the movement of fabric against skin can liberate skin flake particles into the air.

Konda et al.133 used a two-chamber laboratory apparatus to measure size-resolved particle penetration through various fabrics, alone or in combination. The test sample had an area of 59 cm2, and two constant flow rates were tested, 35 L/min and 90 L/min. These flow rates are substantially higher than the at-rest volumetric breathing rates that would apply in common social settings. The particle diameters measured in this study spanned a range from 10 nm to 6 µm. The lowest order of magnitude in size is not directly relevant to assessing the transmission risk for viruses that are ~100 nm in diameter alone and are emitted in association with the nonvolatile components of respiratory fluids.

Pan et al.134 used a test chamber outfitted with two manikin heads to assess the performance of various masks and face-covering fabrics both for control of exhaled particles and for protection on inhalation. They used four different particle generators to produce polydisperse particles spanning an overall diameter range of 0.04 µm to >100 µm. Airborne particles were measured with an aerodynamic particle sizer (range 0.3–20 µm). Glass slides were mounted on the face of the receiver manikin to optically measure the deposition of larger particles. Among the key findings was that “the fit of the mask was important.” For high-quality protection, the authors recommended “a three-layer mask consisting of outer layers of a flexible, tightly woven fabric and an inner layer consisting of a material designed to filter out particles. This combination should produce an overall efficiency of >70% at the most penetrating particle size and >90% for particles 1 µm and larger if the mask fits well.”

The COVID-19 pandemic has also inspired several synthesis studies that assemble and interpret the literature on mask wearing and its effectiveness. Chu et al.93 undertook a systematic review and meta-analysis on the effect of mask wearing on the transmission of respiratory viral infections. They concluded, albeit with low certainty, that “medical or surgical face masks might result in a large reduction in virus infection.” Altogether, they identified 39 relevant studies for their analysis; however, only five of these were undertaken outside of healthcare settings. The review by Howard et al.135 effectively highlights an important point that is underappreciated: conditions in healthcare environments are considerably different than in the broader community. Transmission risks and mitigation measures can be rationally differentiated between these circumstances. Specifically, with regards to wearing masks, Howard et al.135 recommend an “increasing focus on a previously overlooked aspect of mask usage: mask wearing by infectious people (‘source control’) with benefits at the population level, rather than only mask wearing by susceptible people, such as health care workers, with focus on individual outcomes.”

Quantitative estimates of the expected efficacy of masking are possible with knowledge of the size-dependent removal efficiency of mask materials. Figure 6 presents an example of such data, using theoretical fits to empirical data.136 The modeled efficiency plots show markedly different efficiency among the three mask categories for small particles (diameter <5 µm). Conversely, all three mask types effectively filter large particles (diameters of 5–100 µm).

Modeled filtration efficiency of three different mask types as a function of particle diameter. The efficiency represents the fractional removal of particles from air passing through the mask. The traces reflect equations that combine theory with empirical data, as reported in equation (S4b) of Wagner et al.136

The influence of masking on limiting respiratory emissions depends on the combination of size-dependent emissions and size-dependent removal efficiency. This point is illustrated for the case of speaking in Table 6. Here, the fractional reductions of emissions from different modes, as reported in Table 2, are summarized, both considering the number of particles as well as the volume of respiratory droplets, and separately analyzing for small particles (0.1–5 µm), large particles (5–100 µm), and drops. All mask types are effective in substantially reducing emissions of large particles and drops. Modeled control of fine particle emissions becomes progressively better with mask quality improvement from cloth to surgical and then to respirator. An important qualitative point is illustrated by the results in Table 6: High-quality masks are required for protection against small particles whereas a wide range of face coverings can provide considerable benefits in reducing emissions of large particles and drops.

TABLE 6.
Effect of masking in proportionally reducing emission rates of respiratory particles and drops when speaking

	Cloth mask		Surgical mask		Respirator mask
	Number	Volume	Number	Volume	Number	Volume
B modes
Small PM (0.1–5 µm)	33%	44%	65%	88%	86%	97%
Large PM (5–100 µm)	94%	95%	95%	95%	99%	99%
LT mode
Small PM (0.1–5 µm)	45%	65%	90%	95%	98%	99%
Large PM (5–100 µm)	95%	95%	95%	95%	99%	99%
O modes
Small PM (0.1–5 µm)	77%	84%	95%	95%	99%	99%
Large PM (5–100 µm)	95%	95%	95%	95%	99%	99%
Drops (100 µm–1 mm)	95%	95%	95%	95%	99%	99%

Because masks can reduce both emissions and inhalation intake, the net effect of consistent mask use indoors can be remarkably large in lowering overall intake rates of respiratory particles. That point is illustrated in Table 7, which reports the modeled intake rates of respiratory particles from an infectious emitter who is speaking in the base-case simulation of a classroom setting, as reported in Table 4. For small particles, the rate of inhalation intake is reduced by amounts ranging from 83% for cloth masks to 99.98% for respirator masks. Mitigation effectiveness is even better for large particles, with intake reductions considerably better than 99% for all mask types.

TABLE 7.
Effect of masking on inhalation intake rates (µm3/h) of respiratory particles from an infectious emitter who is speaking in the classroom base-case simulation

Exposure conditions	Small (0.1–5 µm)	Large (5–50 µm)
Unmasked	25 400	345 200
Cloth mask	4400	860
Surgical mask	110	860
Respirator mask	6	35

The simulation results in Tables 6 and 7 indicate that masks can be highly effective in attenuating the risk of indoor airborne transmission of SARS-CoV-2. However, some important limitations should also be recognized. First, the efficiency plots in Figure 6 assume good mask fit. Any leakage that leads to bypass of exhaled or inhaled airflow would reduce the overall efficiency and effectiveness. Second, masks are not a consistently suitable intervention in all community settings. In indoor environments where people eat and drink, such as in restaurants and nightclubs, many people would be necessarily unmasked for much of their duration of occupancy.

9.2 Ventilation and filtration

Particles are primarily removed from indoor air by a combination of three processes: ventilation to outdoor air, filtration, and deposition to indoor surfaces. A technical mitigation measure to reduce the transmission of airborne SARS-CoV-2 is to increase the rate of removal by means of enhanced ventilation and/or improved filtration. Through such means, the airborne concentration of infectious agents can be reduced. The extent of reduction scales directly with the degree to which the total removal rate (by means of ventilation plus filtration plus deposition) is increased.

To illustrate the potential of this intervention, let’s return to the base-case simulation of a residence with conditions summarized in Table 4 for the circumstances in which an infectious person is speaking. To isolate the influence of ventilation and filtration, let’s further assume that all occupants are unmasked. In the base-case condition, ventilation of the 250 m3 residence occurs at a volumetric rate of 125 m3 h−1, there is no filtration, and deposition occurs at a size-dependent rate as displayed in Figure 2.

To this base case, four levels of ventilation and filtration enhancement are considered. In each case, the enhanced volume flow rate is five house volumes per hour (10× the base air-change rate), that is, 1250 m3 h−1. In the first intervention case, that air flow is free of all infectious particles, either because it is ventilation supplied from outdoor air or because it is passed through a perfect (100% efficient) filter. In the other three cases, the air is recirculated internally through air-handling filters with ratings of MERV 12, MERV 10, and MERV 6, respectively. The assumed size-dependent single-pass particle removal efficiencies for these filters are displayed in Figure 7. For each of the mitigation cases, ventilation or air filtration is provided with continuously operating flow. The filtration cases could be realized either by a stand-alone recirculating fan-filter unit or by incorporating improved filters into a central air-handling system. In the latter case, it is assumed that the fan is configured to operate continuously, rather than under control of a thermostat.

Size-dependent single-pass efficiency for three grades of filters used in ventilation systems. The efficiency curves are as reported in Table S2 of Azimi et al.142 (The MERV 12 filter is for entry “12 #2”.)

Table 8 displays simulation results, reporting cumulative inhalation intake rate of small (0.1–5 µm) and large (5–50 µm) particles for the base-case conditions and for each of the mitigation configurations. Comparing the entries for mitigation conditions to the base-case conditions, one observes that the effectiveness of these interventions is poor (almost negligible) for large particles and in the range minor to moderate for small particles. For example, the fan-filter configuration with MERV 12 filter reduces small particle intake by 69%, but only reduces large-particle intake by 7%.

TABLE 8.
Effect of filtration and enhanced ventilation on inhalation intake rates (µm3/h) of respiratory particles from a speaking infectious emitter in the residential base-case simulation

Conditions	Small (0.1–5 µm)	Large (5–50 µm)
Base case	6600	44 600
Add 5 h−1 ventilation	1890	41 300
Add 5 h−1 MERV 12	2030	41 300
Add 5 h−1 MERV 10	2400	41 400
Add 5 h−1 MERV 6	4050	42 100

Two qualitative features of these results merit further attention. First, the much poorer performance for large particles as compared to small particles is a consequence of the very high baseline deposition rate of large particles to indoor surfaces. To illustrate, the regression reported in Figure 2 suggests that 30 µm particles are lost by deposition to room surfaces with a loss rate coefficient of about 56 h−1. Adding removal by filtration at a rate of 5 h−1 would increase the total removal rate to 61 h−1. The corresponding percentage reduction in modeled intake would be the same as the percentage increase in total removal, which would be about 8% for this case. The somewhat higher effectiveness for small particles reflects the lower baseline removal rate by deposition. Considering only emissions of small particles (0.1–5 µm), the volume-weighted median size is approximately 1.7 µm, for which the deposition rate would be about 1 h−1 (see Figure 2). The total removal rate for this size particle in the base-case residential simulation would be 1.5 h−1. Adding 5 h−1 of removal by increased ventilation would lower the intake by 5/6.5 = 77%, similar to the small particle control effectiveness indicated for enhanced ventilation in Table 8.

A second important qualitative point is revealed by comparing results between Tables 7 and 8. Masking is much more effective than ventilation and filtration in mitigating transmission risk. That finding is a consequence of masking being a closed-path intervention whereas ventilation and room air filtration are open-path interventions. A key distinction is that with an open-path air-cleaning configuration, the opportunity exists for the contaminant to travel from the source to the receptor without passing through the treatment device, whereas in a closed-path configuration, air traveling from source to receptor must traverse the treatment.137

10 SYNTHESIS

This section summarizes important aspects of the current state of knowledge about aerosol transmission of SARS-CoV-2 in community settings, emphasizing three key steps in direct airborne transmission: emission of virus-containing particles from an infectious person, transport in the indoor environment between source and receptor, and inhalation intake and deposition in the respiratory tract of a susceptible person. Before proceeding, three major challenges are highlighted.

First, this system is remarkably complex across several dimensions. These include where in the respiratory tract the virus replicates and sheds, differences in how and where particle emissions occur in the respiratory tract according to the type of activity (breathing, speaking, coughing, etc.), the broad range of inhalable particle sizes that are emitted, the widely varying size-dependent dynamic behavior of airborne particles indoors, the influence of indoor environmental factors such as ventilation and filtration rates, the variability of particle deposition probability and location within the respiratory tract, and the relative importance of source-receptor proximity as a key factor affecting airborne transmission risk.

A second broad concern is the pervasive error in the literature regarding respiratory infectious diseases about the size that differentiates airborne particles from ballistic drops. Starting from an incorrect assumption that particles larger than 5 µm will not travel more than 2 m from the emission source has produced downstream errors in interpreting the evidence. Such errors have clouded understanding and impeded progress.

A third feature is the novelty and virulence of SARS-CoV-2, especially the recently emerging delta variant. COVID-19 is the most severe pandemic of the past century. Although caused by nominally similar viral agents, the transmission characteristics of SARS-CoV-2 are remarkably different from those of SARS-CoV, the coronavirus responsible for the 2003 SARS epidemic. An appropriate response to these circumstances demands a fresh look at the evidence.

A useful organizing principle for thinking about airborne transmission of SARS-CoV-2 is an array with 3 × 3 elements. One axis represents size, clustered into small particles (0.1–5 µm diameter), large particles (5–100 µm), and ballistic drops (>100 µm). The other axis is the stage in the transmission process: emissions from the infectious source, transport in the indoor environment, and uptake by the susceptible person. In the descriptions that follow, one should be mindful that the boundaries among categories are not sharply defined, especially regarding particle size. The top end of the small particle size range can exhibit similar behavior as the lower end of the large-particle range. Similarly, the smallest of the ballistic drops behave similarly to the biggest of the large particles. Despite the fuzziness of the boundaries, clustering is useful because the bulk of each grouping has distinct features.

Ballistic drops (≥100 µm) are emitted from the nose and mouth. Although considerable attention historically emphasized symptomatic events, such as coughing and sneezing, it is also apparent that talking (and other vocalization activities, such as laughing and singing) can be a substantial source of ballistic drops. Although the number of ballistic drops emitted is relatively small, these drops contain a substantial proportion of the volume of emitted respiratory liquids. To the extent that virus is generated and shed in the upper throat, mouth, and/or nasal passages, these ballistic drops could be a significant contributor to total transmission risk. Once emitted, the ballistic drops remain airborne for only a short time and generally travel only a short distance from the source. Exceptions to the travel distance can occur for events such as uncovered sneezes,106 but for SARS-CoV-2 such events appear to be outliers, rather than predominant features of total emissions. During the short period that they are airborne, ballistic drops do not evaporate much, so that their size when deposited is close to the size as emitted. Indoor environmental conditions normally would have relatively little influence on the fate of ballistic drops. Indoor airflow speeds are generally small enough (typically well below 1 m/s) to not strongly affect the trajectory of ballistic drops. The deposition of ballistic drops on the body envelope of a susceptible receptor could occur by a combination of inertial impaction and gravitational settling if the receptor is proximate to the source. If sufficient viral deposition occurs directly on exposed mucus membranes, then a new infection could be initiated. Deposition of ballistic drops onto other surfaces could contribute to exposure via fomites. That is typically thought of in terms of transfer from a contaminated surface to the mucus membranes via hand contact. For example, if sufficient virus-containing ballistic drops deposit on the clothing of a susceptible person through a substantial period of exposure near an infectious person, as in conversation, then it is conceivable for an infection to be initiated because of hand-mediated transfer of the virus from the clothing surface to the mucus membranes of the susceptible person. The same process could apply with the initial deposition occurring onto exposed skin, or onto an inanimate surface, such as a table during a shared meal.

Large particles (5–100 µm) are generated in the upper respiratory tract. Emission sites include the larynx and mouth during speech as well as the trachea, oral, and nasal cavities during coughing and sneezing. Two prominent features of their behavior distinguish large particles from the two other size modes. First, as compared with small particles, large particles settle rapidly and therefore have relatively short indoor air residence times. The short residence time has several important consequences. (a) The fraction of emitted large particles that is inhaled is attenuated by the rapid deposition-associated removal. (b) The opportunity to control exposure through improved filtration and ventilation is diminished for large particles. (c) To the extent that indoor air lifetimes are on the same scale or shorter than mixing time scales, the proximity effect becomes especially pronounced for large particles. To elaborate on point (c), because of the high rate of removal by deposition of large particles to indoor surfaces, there would be a tendency for the near-source concentration to be more substantially enhanced for large particles than would be expected for small particles. Second, as compared with ballistic drops, large particles can be inhaled, which is a more efficient mode of transfer to mucus membranes than impaction restricted to the small areas of open eyes, nostrils, or mouth. When inhaled, large particles would tend to be deposited primarily in the head and secondarily in the tracheobronchial tree; they do not penetrate to the pulmonary region of the respiratory tract.

Small particles (0.1–5 µm) can be generated in the deep lung through the bronchiolar fluid film bursting mechanism.69 Vocalization, including speaking and singing, also generates small particles, which may be produced at and near the vocal cords.10 Small particle dynamic behavior has been extensively studied in indoor air and is relatively well understood.138, 139 Indoor air concentrations resulting from a given time pattern of emissions are influenced mainly by rates of ventilation, filtration, and deposition. The well-mixed model of an indoor environment provides a reasonable approximation of reality. When inhaled, small particles deposit throughout the respiratory tract, with the bigger of the small particles depositing significantly in the head. There is substantial probability of tracheobronchial and pulmonary deposition across the small particle size range.

Large and small particles shrink after emission through the evaporative loss of water. This process only causes a loss of volatile components. The residual nonvolatile components are sufficient in abundance so that the diameters of equilibrated particles are likely to be 20–40% of the emitted diameters.60 The kinetics of evaporation are rapid both for small particles and for the lower end of the large-particle size range such that the equilibrium state can be assumed to be attained instantaneously. For the bigger of the large particles, mass and heat transfer limitations may cause the airborne dynamic behavior to depend on time-varying size.14

Is it possible to assemble all the evidence and present a coherent picture of the primary modes of indoor airborne transmission of SARS-CoV-2? Perhaps. Key features of the evidence are these. In less than 2 years since the onset of the pandemic, COVID-19 has spread to every community. Notwithstanding enormous social effort entailing widespread use of nonpharmaceutical interventions to stem the spread, more than 250 million diagnosed cases have occurred (as of mid-November 2021). The infections appear to be mainly spread indoors rather than outdoors. Much of the transmission originates from infectious persons who are not particularly symptomatic at the time. Superspreading events are common and contribute to the overall rapid spread of the disease. Community spread is dominant. Social distancing and masking are at least moderately effective control measures. From a mechanistic perspective, emissions are considerably larger from speaking and other vocalization activities than from quiet breathing. Emissions from speaking span a broad range of sizes with substantial contributions to small particles, large particles, and ballistic drops. In combination with information on viral loads in respiratory fluids and on anticipated infectious doses, emissions from speaking appear sufficient to initiate new infections in circumstances in which the transfer efficiencies from source to receptor are high. Such circumstances could occur when the infectious and susceptible persons are in proximity and engaged in unmasked conversation. They could also occur at room scale, especially under poor ventilation conditions.

The significance of superspreading events points in the direction of important room-scale transmission mediated by small particles. Transmission at room scale provides the means for all occupants to share in the exposure. The emphasis on small particle contributions is indicated by such particles persisting sufficiently in indoor air to be present at high concentrations, especially if the removal rate is not sufficient, that is, because of low ventilation rates and nonexistent or ineffective air filtration. Surgical masks and especially cloth face coverings would not eliminate the transmission risk associated with small particles. However, absent N95-level respiratory protection, a combination of these factors would contribute to effectively lowering risk: low occupant density, short occupancy intervals, sufficient outdoor air ventilation, effective room air filtration, and face coverings.

Indications that masking and proximity are important factors influencing the transmission risk suggests contributions from large-particle inhalation and/or ballistic drop deposition. Ballistic drops can only make a sizeable contribution to exposure in the near-field range (within a meter or two from the source). Large particles can travel farther, but because of their high rate of deposition, the probability of high exposure is enhanced through direct exposure to the plume of emissions. For proximate exposures, masking is likely to be an effective mitigation measure because large particles can be effectively removed from airstreams on both the emissions and inhalation intake sides. Also, a mask on an infectious source who is speaking will attenuate the exhaled air plume from the mouth, limiting the efficiency of large-particle transport across meter-scale distances.

The descriptions in the previous two paragraphs might seem contradictory. However, the available evidence does not suggest that they are mutually exclusive. Instead of thinking about routes of exposure as “either/or,” the high rates of community spread strongly suggest that “both/and” would apply.

The relative dominance of indoor versus outdoor environments as settings for disease transmission provides additional clues about transmission mechanisms. First, transmission that is modulated by the indoor confinement of small particles indoors does not have a direct analog in outdoor air. That is, to the extent that small particle, room-scale transmission is a dominant contributor to the overall spread, the concomitant dominance of indoor settings over outdoor venues is clearly consistent. Less clear is to reconcile that proximity is a major risk factor indoors, but not outdoors. On the one hand, it is plausible that proximity is a risk factor outdoors but that it does not emerge strongly from outbreak investigations because sustained proximity cannot involve large numbers of susceptible persons within a meter or so of an infectious person. For example, a minimum of three new cases was one of the inclusion criteria for the outbreak investigations reported by Qian et al.5 Outdoor transmission in proximity that may have infected one or two persons per event would not have been detected. It is also plausible that the altered airflow conditions between outdoor and indoor environments attenuate the near-field risk outdoors. To elaborate, indoors a typical airspeed is 0.1 m/s. Outdoors, routine monitoring data measured at 10 m height for meteorological and climatological purposes show that typical airspeeds are in the range 2–8 m/s (https://www.ncdc.noaa.gov/societal-impacts/wind/). Even when diminished to a fraction of this range for breathing heights of 1–2 m, outdoor air speeds would commonly be much higher than those indoors. The higher outdoor airspeeds would tend to reduce the near-field concentrations by speeding the rate of dilution. That effect would likely be more pronounced for small and large particles than for ballistic drops, whose motion is less affected by background air speeds. So, considering the indications that transmission risks in proximity are higher indoors than outdoors, this point of comparison might argue in favor of the close-proximity indoor transmission being dominated by large particles rather than ballistic drops. However, all this evidence about the indoor and outdoor transmission relationships is at best suggestive. Systematic investigations are needed of transmission and fate of large particles and ballistic drops in relation to source-receptor proximity under different airflow conditions.

11 CONCLUSION

In the legal system of the United States, different standards of proof apply, depending on the type of decision to be made.140 For conviction in a criminal trial, the evidence presented must persuade the decision maker to a level that is beyond a reasonable doubt. In the case of a civil suit seeking financial damages, the decision must be supported by a preponderance of the evidence, with “preponderance” indicating “more likely than not.” An intermediate standard, clear and convincing evidence, also applies in certain circumstances. In working toward an understanding of the modes of transmission of SARS-CoV-2, these qualitative descriptors about standards of proof can help clarify the strength of evidence and its consistency in supporting interpretations. Community transmission of infectious diseases in general, and of COVID-19 in particular, occurs through processes that are understood at a broad conceptual level but are sufficiently complex in detail to defy reliable quantitative assessment. Important public health and public policy decisions should be made using the best available scientific understanding, even if we cannot meet an empirical scientific standard of statistically significant based on randomized controlled trials. This concluding section summarizes key elements in the transmission of SARS-CoV-2 along with my qualitative judgements about the strength of the supporting evidence. There is clear precedent for providing such interpretations in the case of infectious disease transmission indoors, as illustrated in this quote from a multidisciplinary systematic review: “There is strong and sufficient evidence to demonstrate the association between ventilation, air movements in buildings and the transmission/spread of infectious diseases such as measles, tuberculosis, chickenpox, influenza, smallpox and SARS.”141

Beyond a reasonable doubt, in terms of influence on global public health and well-being, COVID-19 is the most impactful pandemic of the past century. In less than 2 years since the onset of the pandemic, the disease has spread globally and infected about 3% of the human population. In the United States, which has seen the largest death total of any country, COVID-19 was the third leading cause of death in 2020, behind only cardiovascular disease and cancer.

Beyond a reasonable doubt, SARS-CoV-2 transmission occurs primarily in community settings, both in households and in environments where people from different households gather. Predominantly, the transmission process is initiated by the emission of the SARS-CoV-2 virus from the respiratory tract of an infectious person, in particles and drops that are mainly comprised of respiratory fluids. Transmission occurs much more commonly indoors than outdoors. Many people are infectious without experiencing strong symptoms of illness.

The evidence is clear and convincing that superspreading events make major contributions to the overall community spread of COVID-19. Clear and convincing evidence also supports a view that several nonpharmaceutical interventions can contribute to reducing (but not eliminating) the risk of transmission in the community: mask wearing, maintaining social distance (≥1–2 m), and adequate ventilation and/or air filtration.

A preponderance of the evidence suggests that speaking and other vocalizations are important means by which the virus is emitted. It is more likely than not that inhaling either or both small particles (0.1–5 µm diameter) and large particles (5–100 µm) containing the SARS-CoV-2 virus constitutes the dominant exposure mechanism for a susceptible person. A preponderance of the evidence supports a view that transmission can occur either at room scale or when a susceptible person is proximate to an infectious person. Mitigation of room-scale transmission would emphasize providing sufficient ventilation and/or filtration, limiting occupancy, and limiting event duration. Protecting against room-scale transmission also points toward the importance of masks that are efficient in filtering small particles. Room-scale transmission is the more likely contributor when superspreading events occur. Mitigating transmission in proximity would emphasize mask wearing and maintaining social distance. Proximity can amplify the source-to-receptor transmission efficiency especially of large particles. Face coverings can reduce the risk of transmission at the proximate scale not only by filtering particles but also by attenuating the range of the exhaled jet of air from the infectious source. At close range, ballistic drops may also contribute materially to exposure risk through deposition onto a susceptible person, not only directly onto sensitive mucus membranes, but also onto other surfaces such as exposed skin and clothing. The final portion of the source-to-receptor transmission chain for ballistic drops and large particles that deposit onto a susceptible person’s skin and clothing has not been adequately studied. There is enough evidence to at least suggest a concern that the resuspension of viral particles from such surfaces followed by entrainment into the human convective boundary layer and inhalation could contribute to overall transmission risk. Mitigation strategies for transmission by ballistic drops overlap substantially with the mitigation of proximate-scale transmission by means of large-particle inhalation. Mask wearing and maintaining social distances are key.

A remarkable feature of the COVID-19 pandemic has been the collective involvement across humankind in rising to face a newly emergent challenge. Individually, and collectively, we are challenged to make decisions about our actions that matter substantially in ways that were not relevant 2 years ago. Loevinger’s140 essay about the foundations of proof in science and law closes with a statement that expresses an apt ambition for these circumstances: that application of the “disciplines of analysis, synthesis, and explication will result in sufficient agreement among a large enough segment of the population to get the world’s work done in a tolerably satisfactory manner and to produce observable benefits from the process.”

REFERENCES

• 1Wang CC, Prather KA, Sznitman J, et al. Airborne transmission of respiratory viruses. Science. 2021; 373(6558):eabd9149.
• 2Heneghan CJ, Spencer EA, Brassey J, et al. SARS-CoV-2 and the role of airborne transmission: a systematic review. F1000 Research. 2021; 10: 232.
• 3Greenhalgh T, Jimenez JL, Prather KA, Tufekci Z, Fisman D, Schooley R. Ten scientific reasons in support of airborne transmission of SARS-CoV-2. BMJ. 2021; 397: 1603– 1605.
• 4Bulfone TC, Malekinejad M, Rutherford GW, Razani N. Outdoor transmission of SARS-CoV-2 and other respiratory viruses: a systematic review. J Infect Dis. 2021; 223: 550– 561.
• 5Qian H, Miao T, Liu L, Zheng X, Luo D, Li Y. Indoor transmission of SARS-CoV-2. Indoor Air. 2021; 31: 639– 645.
• 6Tang JW, Bahnfleth WP, Bluyssen PM, et al. Dismantling myths on the airborne transmission of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2). J Hosp Infect. 2021; 110: 89– 96.
• 7Milton DK. A Rosetta Stone for understanding infectious drops and aerosols. J Pediatric Infect Dis Soc. 2020; 9: 413– 415.
• 8Pöhlker ML, Krüger OO, Förster J-D, et al. Respiratory aerosols and droplets in the transmission of infectious diseases. arXiv. 2021: 2103.01188v4.
• 9Gandhi M, Yokoe DS, Havlir DV. Asymptomatic transmission, the Achilles’ heel of current strategies to control Covid-19. N Engl J Med. 2020; 382: 2158– 2160.
• 10Johnson GR, Morawska L, Ristovski ZD, et al. Modality of human expired aerosol size distributions. J Aerosol Sci. 2011; 42: 839– 851.
• 11Wei J, Li Y. Airborne spread of infectious agents in the indoor environment. Am J Infect Control. 2016; 44: S102– S108.
• 12Nicas M, Nazaroff WW, Hubbard A. Toward understanding the risk of secondary airborne infection: emission of respirable pathogens. J Occup Environ Hyg. 2005; 2: 143– 154.
• 13Vejerano EP, Marr LC. Physico-chemical characteristics of evaporating respiratory fluid droplets. J R Soc Interface. 2018; 15: 20170939.
• 14Xie X, Li Y, Chwang ATY, Ho PL, Seto WH. How far droplets can move in indoor environments — revisiting the Wells evaporation-falling curve. Indoor Air. 2007; 17: 211– 225.
• 15Moriyama M, Hugentobler WJ, Iwasaki A. Seasonality of respiratory viral infections. Annu Rev Virol. 2020; 7(1): 83– 101.
• 16Li Y. Basic routes of transmission of respiratory pathogens — a new proposal for transmission categorization based on respiratory spray, inhalation, and touch. Indoor Air. 2021; 31: 3– 6.
• 17Bond TC, Bosco-Lauth A, Farmer DK, et al. Quantifying proximity, confinement, and interventions in disease outbreaks: a decision support framework for air-transported pathogens. Environ Sci Technol. 2021; 55: 2890– 2898.
• 18Hwang SE, Chang JH, Oh B, Heo J. Possible aerosol transmission of COVID-19 associated with an outbreak in an apartment in Seoul, South Korea, 2020. Int J Infect Dis. 2021; 104: 73– 76.
• 19Mao J, Gao N. The airborne transmission of infection between flats in high-rise residential buildings: a review. Build Environ. 2015; 94: 516– 531.
• 20Matthews TG, Thompson CV, Wilson DL, Hawthorne AR, Mage DT. Air velocities inside domestic environments: an important parameter in the study of indoor air quality and climate. Environ Int. 1989; 15: 545– 550.
• 21Baldwin PEJ, Maynard AD. A survey of wind speeds in indoor workplaces. Ann Occup Hyg. 1998; 42: 303– 313.
• 22Zhai Y, Zhang H, Zhang Y, Pasut W, Arens E, Meng Q. Comfort under personally controlled air movement in warm and humid environments. Build Environ. 2013; 65: 109– 117.
• 23Persily A, de Jonge L. Carbon dioxide generation rates for building occupants. Indoor Air. 2017; 27: 868– 879.
• 24Nazaroff WW. Residential air-change rates: a critical review. Indoor Air. 2021; 31: 282– 313.
• 25Buonanno G, Morawska L, Stabile L. Quantitative assessment of the risk of airborne transmission of SARS-CoV-2 infection: prospective and retrospective applications. Environ Int. 2020; 145: 106112.
• 26Miller SL, Nazaroff WW, Jimenez JL, et al. Transmission of SARS-CoV-2 by inhalation of respiratory aerosol in the Skagit Valley Chorale superspreading event. Indoor Air. 2021; 31: 314– 323.
• 27Milton DK. What was the primary mode of smallpox transmission? Implications for biodefense. Front Cell Infect Microbiol. 2012; 2: 150.
• 28Nguyen-Van-Tam JS, Killingley B, Enstone J, et al. Minimal transmission in an influenza A (H3N2) human challenge-transmission model within a controlled exposure environment. PLoS Pathog. 2020; 16:e1008704.
• 29Althouse BM, Wenger EA, Miller JC, et al. Superspreading events in the transmission dynamics of SARS-CoV-2: opportunities for interventions and control. PLoS Biol. 2020; 18:e3000897.
• 30Lau MSY, Grenfell B, Thomas M, Bryan M, Nelson K, Lopman B. Characterizing superspreading events and age-specific infectiousness of SARS-CoV-2 transmission in Georgia, USA. Proc Natl Acad Sci USA. 2020; 117: 22430– 22435.
• 31Majra D, Benson J, Pitts J, Stebbing J. SARS-CoV-2 (COVID-19) superspreader events. J Infect. 2021; 82: 36– 40.
• 32Bueno de Mesquita PJ, Noakes CJ, Milton DK. Quantitative aerobiologic analysis of an influenza human challenge-transmission trial. Indoor Air. 2020; 30: 1189– 1198.
• 33Asadi S, Wexler AS, Cappa CD, Barreda S, Bouvier NM, Ristenpart WD. Aerosol emission and superemission during human speech increase with voice loudness. Sci Rep. 2019; 9: 2348.
• 34Buckland FE, Bynoe ML, Tyrrell DAJ. Experiments on the spread of colds. II. Studies in volunteers with coxsackievirus A21. J Hyg. 1965; 63: 327– 343.
• 35Riley EC, Murphy G, Riley RL. Airborne spread of measles in a suburban elementary school. Am J Epidemiol. 1978; 107: 421– 432.
• 36Riley S, Fraser C, Donnelly CA, et al. Transmission dynamics of the etiological agent of SARS in Hong Kong: impact of public health interventions. Science. 2003; 300: 1961– 1966.
• 37Glass LM, Glass RJ. Social contact networks for the spread of pandemic influenza in children and teenagers. BMC Public Health. 2008; 8: 61.
• 38Park SY, Kim Y-M, Yi S, et al. Coronavirus disease outbreak in call center, South Korea. Emerg Infect Dis. 2020; 26: 1666– 1670.
• 39Bae S, Kim H, Jung T-Y, et al. Epidemiological characteristics of COVID-19 outbreak at fitness centers in Cheonan, Korea. J Korean Med Sci. 2020; 35:e288.
• 40Kwon K-S, Park J-I, Park YJ, Jung D-M, Ryu K-W, Lee J-H. Evidence of long-distance droplet transmission of SARS-CoV-2 by direct air flow in a restaurant in Korea. J Korean Med Sci. 2020; 35:e415.
• 41de de Man P, Paltansing S, Ong DSY, Vaessen N, van Nielen G, Koeleman JGM, et al. Outbreak of coronavirus disease 2019 (COVID-2019) in a nursing home associated with aerosol transmission as a result of inadequate ventilation. Clin Infect Dis. 2021; 73. 170– 171.
• 42Li Y, Qian H, Hang J, et al. Probable airborne transmission of SARS-CoV-2 in a poorly ventilated restaurant. Build Environ. 2021; 196: 107788.
• 43Xu P, Jia W, Qian H, et al. Lack of cross-transmission of SARS-CoV-2 between passenger’s cabins on the Diamond Princess cruise ship. Build Environ. 2021; 198: 107839.
• 44Vernez D, Schwarz S, Sauvain J-J, Petignat C, Suarez G. Probable aerosol transmission of SARS-CoV-2 in a poorly ventilated courtroom. Indoor Air. 2021. 10.1111/ina.12866
• 45Madewell ZJ, Yang Y, Longini IM Jr, Halloran ME, Dean NE. Household transmission of SARS-CoV-2: a systematic review and meta-analysis. JAMA Network Open. 2020; 3:e2031756.
• 46Collins FS. No longer just looking under the lamppost. Am J Hum Genet. 2006; 79: 421– 426.
• 47Zhu NA, Zhang D, Wang W, et al. A novel coronavirus from patients with pneumonia in China, 2019. N Engl J Med. 2020; 382: 727– 733.
• 48Seto WH. Airborne transmission and precautions: facts and myths. J Hosp Infect. 2015; 89: 225– 228.
• 49Randall K, Ewing ET, Marr LC, Jimenez JL, Bourouiba L. How did we get here: What are droplets and aerosols and how far do they go? A historical perspective on the transmission of respiratory infectious diseases. Interface Focus. 2021; 11. SSRN: 20210049.
• 50Hofmann W. Modelling inhaled particle deposition in the human lung — a review. J Aerosol Sci. 2011; 42: 693– 724.
• 51Beard KV. Terminal velocity and shape of cloud and precipitation drops aloft. J Atmos Sci. 1976; 33: 851– 864.
• 52Bourouiba L. Turbulent gas clouds and respiratory pathogen emissions: potential implications for reducing transmission of COVID-19. JAMA. 2020; 323: 1837– 1838.
• 53Tang JW, Liebner TJ, Craven BA, Settles GS. A schlieren optical study of the human cough with and without wearing masks for aerosol infection control. J R Soc Interface. 2009; 6: S727– S736.
• 54Tang JW, Nicolle ADG, Pantelic J, et al. Qualitative real-time schlieren and shadowgraph imaging of human exhaled airflows: an aid to aerosol infection control. PLoS One. 2011; 6:e21392.
• 55Chen C, Lin C-H, Jiang Z, Chen Q. Simplified models for exhaled airflow from a cough with the mouth covered. Indoor Air. 2014; 24: 580– 591.
• 56Thatcher TL, Lai ACK, Moreno-Jackson R, Sextro RG, Nazaroff WW. Effects of room furnishings and air speed on particle deposition rates indoors. Atmos Environ. 2002; 36: 1811– 1819.
• 57Wells WF. On air-borne infection. Study II. Droplets and droplet nuclei. Am J Epidemiol. 1934; 20: 611– 618.
• 58Liu L, Wei J, Li Y, Ooi A. Evaporation and dispersion of respiratory droplets from coughing. Indoor Air. 2017; 27: 179– 190.
• 59Walker JS, Archer J, Gregson FKA, Michel SES, Bzdek BR, Reid JP. Accurate representations of the microphysical processes occurring during the transport of exhaled aerosols and droplets. ACS Central Sci. 2021; 7: 200– 209.
• 60Marr LC, Tang JW, Van Mullekom J, Lakdawala SS. Mechanistic insights into the effect of humidity on airborne influenza virus survival, transmission and incidence. J R Soc Interface. 2019; 16: 20180298.
• 61Netz RR. Mechanisms of airborne infection via evaporating and sedimenting droplets produced by speaking. J Phys Chem B. 2020; 124: 7093– 7101.
• 62Lieber C, Melekidis S, Koch R, Bauer H-J. Insights into the evaporation characteristics of saliva droplets and aerosols: levitation experiments and numerical modeling. J Aerosol Sci. 2021; 154: 105760.
• 63Hinds WC. Aerosol Technology: Properties, Behavior, and Measurement of Airborne Particles. 2nd ed. Wiley; 1999.
• 64Riley RL. Airborne infection. Am J Med. 1974; 57: 466– 475.
• 65Warfel JM, Beren J, Merkel TJ. Airborne transmission of Bordetella pertussis. J Infect Dis. 2012; 206: 902– 906.
• 66Duguid JP. The size and the duration of air-carriage of respiratory droplets and droplet nuclei. J Hyg. 1946; 44: 471– 479.
• 67Loudon RG, Roberts RM. Droplet expulsion from the respiratory tract. Am Rev Respir Dis. 1967; 95: 435– 442.
• 68Morawska L, Johnson GR, Ristovski ZD, et al. Size distribution and sites of origin of droplets expelled from the human respiratory tract during expiratory activities. J Aerosol Sci. 2009; 40: 256– 269.
• 69Johnson GR, Morawska L. The mechanism of breath aerosol formation. J Aerosol Med Pulm Drug Deliv. 2009; 22: 229– 237.
• 70Han ZY, Weng WG, Huang QY. Characterizations of particle size distribution of the droplets exhaled by sneeze. J R Soc Interface. 2013; 10: 20130560.
• 71He XI, Lau EHY, Wu P, et al. Temporal dynamics in viral shedding and transmissibility of COVID-19. Nat Med. 2020; 26: 672– 675.
• 72Lindsley WG, Blachere FM, Beezhold DH, et al. Viable influenza A virus in airborne particles expelled during coughs versus exhalations. Influenza Other Respir Viruses. 2016; 10: 404– 413.
• 73Buonanno G, Stabile L, Morawska L. Estimation of airborne viral emission: quanta emission rate of SARS-CoV-2 for infection risk assessment. Environ Int. 2020; 141: 105794.
• 74Wölfel R, Corman VM, Guggemos W, et al. Virological assessment of hospitalized patients with COVID-2019. Nature. 2020; 581: 465– 469.
• 75Pan Y, Zhang D, Yang P, Poon LLM, Wang Q. Viral load of SARS-CoV-2 in clinical samples. Lancet Infect Dis. 2020; 20: 411– 412.
• 76To KK-W, Tsang OT-Y, Leung W-S, et al. Temporal profiles of viral load in posterior oropharyngeal saliva samples and serum antibody responses during infection by SARS-CoV-2: an observational cohort study. Lancet Infect Dis. 2020; 20: 565– 574.
• 77Yang Q, Saldi TK, Gonzales PK, et al. Just 2% of SARS-CoV-2—positive individuals carry 90% of the virus circulating in communities. Proc Natl Acad Sci USA. 2021; 118(21):e2104547118.
• 78Baron A, Hachem M, Tran Van Nhieu J, et al. Bronchoalveolar lavage in patients with COVID-19 with invasive mechanical ventilation for acute respiratory distress syndrome. Ann Am Thorac Soc. 2021; 18: 723– 726.
• 79Hou YJ, Okuda K, Edwards CE, et al. SARS-CoV-2 reverse genetics reveals a variable infection gradient in the respiratory tract. Cell. 2020; 182: 429– 446.
• 80Vincent JH, Mark D, Miller BG, Armbruster L, Ogden TL. Aerosol inhalability at higher windspeeds. J Aerosol Sci. 1990; 21: 577– 586.
• 81Volkwein JC, Maynard AD, Harper M. Workplace aerosol measurement. In: P Kulkarni, PA Baron, K Willeke, eds. Aerosol Measurement: Principles, Techniques, and Applications. 3rd ed. Wiley; 2011: 571– 590.
• 82Guo Y, Wei J, Ou C, et al. Deposition of droplets from the trachea or bronchus in the respiratory tract during exhalation: a steady-state numerical investigation. Aerosol Sci Technol. 2020; 54: 869– 879.
• 83 USEPA. Inhalation rates. Chapter 6. In: Exposure Factors Handbook. United States Environmental Protection Agency; 2011.
• 84Baughman AV, Gadgil AJ, Nazaroff WW. Mixing of a point source pollutant by natural convection flow within a room. Indoor Air. 1994; 4: 114– 122.
• 85Drescher AC, Lobascio C, Gadgil AJ, Nazaroff WW. Mixing of a point-source indoor pollutant by forced convection. Indoor Air. 1995; 5: 204– 214.
• 86Cheng K-C, Acevedo-Bolton V, Jiang R-T, et al. Modeling exposure close to air pollution sources in naturally ventilated residences: association of turbulent diffusion coefficient with air change rate. Environ Sci Technol. 2011; 45: 4016– 4022.
• 87Gammaitoni L, Nucci MC. Using a mathematical model to evaluate the efficacy of TB control measures. Emerg Infect Dis. 1997; 3: 335– 342.
• 88Bennett DH, McKone TE, Evans JS. Defining intake fraction. Environ Sci Technol. 2002; 36: 206A– 211A.
• 89Rudnick SN, Milton DK. Risk of indoor airborne infection transmission estimated from carbon dioxide concentration. Indoor Air. 2003; 13: 237– 245.
• 90Melikov AK, Ai ZT, Markov DG. Intermittent occupancy combined with ventilation: an efficient strategy for the reduction of airborne transmission indoors. Sci Total Environ. 2020; 744: 140908.
• 91Nazaroff WW. Inhalation intake fraction of pollutants from episodic indoor emissions. Build Environ. 2008; 43: 269– 277.
• 92Jensen PA, O’Brien D. Industrial Hygiene. In: K Willeke, PA Baron, eds. Aerosol Measurement: Principles, Techniques, and Applications. Van Nostrand Reinhold; 1993: 537– 559.
• 93Chu DK, Akl EA, Duda S, et al. Physical distancing, face masks, and eye protection to prevent person-to-person transmission of SARS-CoV-2 and COVID-19: a systematic review and meta-analysis. Lancet. 2020; 395: 1973– 1987.
• 94Olmedo I, Nielsen PV, Ruiz de Adana M, Jensen RL, Grzelecki P. Distribution of exhaled contaminants and personal exposure in a room using three different air distribution strategies. Indoor Air. 2012; 22: 64– 76.
• 95Chen C, Zhao B, Lai D, Liu W. A simple method for differentiating direct and indirect exposure to exhaled contaminants in mechanically ventilated rooms. Build Simul. 2018; 11: 1039– 1051.
• 96Liu L, Li Y, Nielsen PV, Wei J, Jensen RL. Short-range airborne transmission of expiratory droplets between two people. Indoor Air. 2017; 27: 452– 462.
• 97Qian H, Li Y, Nielsen PV, Hyldgaard CE. Dispersion of exhalation pollutants in a two-bed hospital ward with a downward ventilation system. Build Environ. 2008; 43: 344– 354.
• 98Nielsen PV, Li Y, Buus M, Winther FV. Risk of cross-infection in a hospital ward with downward ventilation. Build Environ. 2010; 45: 2008– 2014.
• 99Bjørn E, Nielsen PV. Dispersal of exhaled air and personal exposure in displacement ventilated rooms. Indoor Air. 2002; 13: 147– 164.
• 100Murakami S. Analysis and design of micro-climate around the human body with respiration by CFD. Indoor Air. 2004; 14(Suppl 7): 144– 156.
• 101Abkarian M, Mendez S, Xue N, Yang F, Stone HA. Speech can produce jet-like transport relevant to asymptomatic spreading of virus. Proc Natl Acad Sci USA. 2020; 117: 25237– 25245.
• 102Tang JW, Nicolle AD, Klettner CA, et al. Airflow dynamics of human jets: sneezing and breathing – potential sources of infectious aerosols. PLoS One. 2013; 8:e59970.
• 103Xu C, Nielsen PV, Liu L, Jensen RL, Gong G. Human exhalation characterization with the aid of schlieren imaging technique. Build Environ. 2017; 112: 190– 199.
• 104Gupta JK, Lin C-H, Chen Q. Flow dynamics and characterization of a cough. Indoor Air. 2009; 19: 517– 525.
• 105Gupta JK, Lin C-H, Chen Q. Characterizing exhaled airflow from breathing and talking. Indoor Air. 2010; 20: 31– 39.
• 106Bourouiba L, Dehandschoewercker E, Bush JWM. Violent expiratory events: on coughing and sneezing. J Fluid Mech. 2014; 745: 537– 563.
• 107Xu C, Nielsen PV, Gong G, Liu L, Jensen RL. Measuring the exhaled breath of a manikin and human subjects. Indoor Air. 2015; 25: 188– 197.
• 108Licina D, Pantelic J, Melikov A, Sekhar C, Tham KW. Experimental investigation of the human convective boundary layer in a quiescent indoor environment. Build Environ. 2014; 75: 79– 91.
• 109Gena AW, Voelker C, Settles GS. Qualitative and quantitative schlieren optical measurement of the human thermal plume. Indoor Air. 2020; 30: 757– 766.
• 110Craven BA, Settles GS. A computational and experimental investigation of the human thermal plume. J Fluids Eng. 2006; 128: 1251– 1258.
• 111Sun S, Li J, Han J. How human thermal plume influences near-human transport of respiratory droplets and airborne particles: a review. Environ Chem Lett. 2021; 19: 1971– 1982.
• 112Licina D, Melikov A, Sekhar C, Tham KW. Human convective boundary layer and its interaction with room ventilation flow. Indoor Air. 2015; 25: 21– 35.
• 113Ma J, Qian H, Nielsen PV, Liu L, Li Y, Zheng X. What dominates personal exposure? Ambient airflow pattern or local human thermal plume. Build Environ. 2021; 196: 107790.
• 114Bhagat RK, Davies Wykes MS, Dalziel SB, Linden PF. Effects of ventilation on the indoor spread of COVID-19. J Fluid Mech. 2020; 903: F1. doi:10.1017/jfm.2020.720
• 115Chin AWH, Chu JTS, Perera MRA, et al. Stability of SARS-CoV-2 in different environmental conditions. Lancet Microbe. 2020; 1:e10.
• 116Marquès M, Domingo JL. Contamination of inert surfaces by SARS-CoV-2: persistence, stability and infectivity. A review. Environ Res. 2021; 193: 110559.
• 117Qian J, Peccia J, Ferro AR. Walking-induced particle resuspension in indoor environments. Atmos Environ. 2014; 89: 464– 481.
• 118Licina D, Tian Y, Nazaroff WW. Emission rates and the personal cloud effect associated with particle release from the perihuman environment. Indoor Air. 2017; 27: 791– 802.
• 119Licina D, Nazaroff WW. Clothing as a transport vector for airborne particles: chamber study. Indoor Air. 2018; 28: 404– 414.
• 120Duguid JP, Wallace AT. Air infection with dust liberated from clothing. Lancet. 1948; 252: 845– 849.
• 121Licina D, Morrison GC, Bekö G, Weschler CJ, Nazaroff WW. Clothing-mediated exposures to chemicals and particles. Environ Sci Technol. 2019; 53: 5559– 5575.
• 122Stephens B, Azimi P, Thoemmes MS, Heidarinejad M, Allen JG, Gilbert JA. Microbial exchange via fomites and implications for human health. Curr Pollut Rep. 2019; 5: 198– 213.
• 123Asadi S, Gaaloul ben Hnia N, Barre RS, Wexler AS, Ristenpart WD, Bouvier NM. Influenza A virus is transmissible via aerosolized fomites. Nat Commun. 2020; 11: 4062.
• 124D’Amato G, Vitale C, D’Amato M, et al. Thunderstorm-related asthma: what happens and why. Clin Exp Allergy. 2016; 46: 390– 396.
• 125Li Y. The secret behind the mask. Indoor Air. 2011; 21: 89– 91.
• 126Verma S, Dhanak M, Frankenfield J. Visualizing the effectiveness of face masks in obstructing respiratory jets. Phys Fluids. 2020; 32:061708.
• 127Grinshpun SA, Haruta H, Eninger RM, Reponen T, McKay RT, Lee S-A. Performance of an N95 filtering facepiece particulate respirator and a surgical mask during human breathing: two pathways for particle penetration. J Occup Environ Hyg. 2009; 6: 593– 603.
• 128Lai ACK, Poon CKM, Cheung ACT. Effectiveness of facemasks to reduce exposure hazards for airborne infections among general populations. J R Soc Interface. 2012; 9: 938– 948.
• 129Cowling BJ, Zhou Y, Ip DKM, Leung GM, Aiello AE. Face masks to prevent transmission of influenza virus: a systematic review. Epidemiol Infect. 2010; 138: 449– 456.
• 130Shakya KM, Noyes A, Kallin R, Peltier RE. Evaluating the efficacy of cloth facemasks in reducing particulate matter exposure. J Exposure Sci Environ Epidemiol. 2017; 27: 352– 357.
• 131Leung NHL, Chu DKW, Shiu EYC, et al. Respiratory virus shedding in exhaled breath and efficacy of face masks. Nat Med. 2020; 26: 676– 680.
• 132Asadi S, Cappa CD, Barreda S, Wexler AS, Bouvier NM, Ristenpart WD. Efficacy of masks and face coverings in controlling outward aerosol particle emission from expiratory activities. Sci Rep. 2020; 10: 15665.
• 133Konda A, Prakash A, Moss GA, Schmoldt M, Grant GD, Guha S. Aerosol filtration efficiency of common fabrics used in respiratory cloth masks. ACS Nano. 2020; 14: 6339– 6347.
• 134Pan J, Harb C, Leng W, Marr LC. Inward and outward effectiveness of cloth masks, a surgical mask, and a face shield. Aerosol Sci Technol. 2021; 55: 718– 733.
• 135Howard J, Huang A, Li Z, et al. An evidence review of face masks against COVID-19. Proc Natl Acad Sci USA. 2021; 118:e2014564118.
• 136Wagner J, Sparks TL, Miller S, Chen W, Macher JM, Waldman JM. Modeling the impacts of physical distancing and other exposure determinants on aerosol transmission. J Occup Environ Hyg. 2021; 18(10-11): 495– 509. doi:
• 137Nazaroff WW, Weschler CJ. Air cleaning effectiveness for improving indoor air quality: Open-path and closed-path configurations. Proceedings of Healthy Buildings 2009, Syracuse, NY. Paper 376.
• 138Nazaroff WW. Indoor particle dynamics. Indoor Air. 2004; 14(Suppl 7): 175– 183.
• 139Nazaroff WW. Indoor bioaerosol dynamics. Indoor Air. 2016; 26: 61– 78.
• 140Loevinger L. Standards of proof in science and law. Jurimetrics. 1992; 32: 323– 344.
• 141Li Y, Leung GM, Tang JW, et al. Role of ventilation in airborne transmission of infectious agents in the built environment — a multidisciplinary systematic review. Indoor Air. 2007; 17: 2– 18.
• 142Azimi P, Zhao D, Stephens B. Estimates of HVAC filtration efficiency for fine and ultrafine particles of outdoor origin. Atmos Environ. 2014; 98: 337– 346.

[ad_2]

Originally Appeared Here