ASHRAE Journal - December 2020 - 14

LETTERS

relatively low. In general, viral airborne contaminant
transport patterns in aircraft cabins are best described
as a near-source region of highly variable exposure
(high and low) that is unmixed and a far-source region
that has lower exposure and is approximately uniformly
mixed. (Bennet reference 10 in my paper).

Modeling Analysis Summary and Conclusion

Achieving a high air change rate in aircraft cabins
requires both relatively high air velocities and a relatively high occupancy density. These velocities generate turbulent fluctuations and disperse airborne contaminant in all directions before their removal. The
USTRANSCOM findings confirm this dispersion and
indicate that aerosols from a single cough can spread
about 11 rows. Some CFD models for aircraft suggest
breath aerosols can spread eight rows from an infector
so the recent USTRANSCOM testing expands the dispersion even more. In my analysis for aircraft and other
settings, I limited the dispersion to three three-person
rows either side of an infector assuming the dispersion
did not cross an aisle.
In this section the letter notes that office building air
is more uniformly mixed than aircraft cabin air and I
agree. However, the implication of this is not a positive
one for aircraft, as the less the dispersion, the higher the
dose received and the more likely an infection will occur.
I note some inconsistency in the letter on this point as in
Section 5, it is said that " localized concentrations of virus
in buildings can be much higher, " implying office building air might not be well-mixed.
As for the HEPA filtration of the recirculated supply air,
it does preclude re-distribution of infectious aerosols.
However, the cabin airflows initially distribute unfiltered infectious aerosols to several rows fore and aft of
an infector.

Discussion

The modelling methodology used in my paper and the
various setting-relative-airborne-viral-infection-risk
predictions from exposure to an ill person's breath have
been peer reviewed several times with the predictions
of relative risk for the eight settings first published a
decade ago.4,5 A more recent publication (Yan et al)3
than the one used in my paper reports an influenza
virus shedding rate of 76,000 virions per hour from one
infected person, rather than the 660 virions per hour
14

ASHRAE JOURNAL

ashrae.org

D ECEM BER 2020

used in my paper or the 4,000 USTRANSCOM used in its
experiments. In doing their influenza shedding measurements, Yan et al3 noted that sneezing was rare and
that sneezing, and coughing was not necessary for the
infectious aerosol generation. For Influenza A Nikitin et
al found the HID50 = 900 virions.6
So, using the Yan influenza shedding rate and the
Nikitin HID50= 900 for Influenza A increases the group
exhaled breath dose in each of the 8 settings analyzed
in my paper by a factor of 115, and that of the coughing
aerosol shedding in the USTRANSCOM coughing tests by
a factor of 19.
Using the Yan and Nikitin values, Influenza A infections due to a coughing aerosol exposure in the aircraft
studied by USTRANSCOM are predicted after 2 hours
and 33 minutes rather than 54 hours. Similarly, the
exposure of 19 healthy persons to one ill person's breath
aerosols on a 14 hour international flight might result in
five of them being infected, while exposing 18 healthy
persons to the breath aerosols of two infected persons,
depending on aerosol dispersion, might result in all 18
being infected. Further, the 10% cabin humidity and/
or smaller group exposures could make the infectioncausing exposure times shorter.

Conclusion

In normal times when mask wearing in public spaces
is no longer common, the risk of an ill person's cough
causing an infection in others can be reduced by coughers covering their mouths. However, no such simple
measure is available to reduce the risk of infectious disease spread from an ill person's normal breathing. We
must rely instead on the design of the HVAC system to
protect us from such infections and that may not be happening as well in some settings as in others.
The USTRANSCOM cough dispersion measurements,
the Boeing CFD modeling of cough aerosols and my
modeling of relative viral infection risk in various settings from the exhaled breath aerosols of infected persons, together with accurate virus shedding rates and
HID50 values, should help medical experts identify the
settings and mechanisms contributing to viral infections. This in turn, along with modelling, should help
engineers design the HVAC/ECS ventilation, filtration,
and humidity control systems of the future for building
and vehicular air, land, and sea settings.

https://www.ashrae.org/

ASHRAE Journal - December 2020

Table of Contents for the Digital Edition of ASHRAE Journal - December 2020