Post #880: Dry indoor air + no masks = outbreak

Posted on October 27, 2020

Source: Underlying data are from Johns Hopkins University, via the NY Times Github COVID-19 data repository.

This is just a linear restatement of my previous post.  Any citations as to sources can be found in the prior post.  This is the argument, in its simplest form.

1:  COVID-19 and flu mainly spread indoors.  Mostly, that’s because we spend almost all our time indoors, estimated at somewhere between 90% and 95% of time, for U.S. adults.  Historically, for adults, outdoor time varied only trivially between winter and summer.  But in addition, outdoor settings are vastly safer than indoor settings, for COVID-19 spread, based on studies of infections in China and Japan.

2:  COVID-19 and flu spread much better in dry air.  That’s been confirmed by (e.g.) laboratory experiments on flu and guinea pigs, by experimental analysis of virus half-life at varying humidities, and by epidemiological analysis of how rapidly these diseases spread relative to prevailing weather.

One study estimated that half of the variation in seasonal flu deaths could be explained by variations in humidity alone. So, not only does humidity matter, it matters a lot.

Greater disease spread in dry air occurs for three reasons.  First, the virus lives modestly longer in dry air.  Second, droplets spread further in dry air because they shrink somewhat, and become lighter, compared to moist air.

But (in my opinion) mainly, third, dry air compromises your body’s first defense against viral pathogens, your nasal membranes.  That has been demonstrated experimentally in mice, showing a very large impact on disease resistance, from the drying of nasal membranes due to reducing relative humidity from 50% to 20%.  That has not been not been directly shown in humans (that I could find), but it’s a good guess that it happens in humans as well.

3:  Wearing a mask means that you breathe humidified air.  (At least, it sure feels that way.)  And so, best guess, wearing a mask keeps your nasal membranes from drying out, or drying out as fast, in dry air.

4: High summer temperatures meant the simultaneous loss of outdoor socializing space, and the drying of indoor air due to use of air conditioning, in southern tier of the US.  Where that combined with low use of facemasks, the result was the second wave of the US COVID-19 pandemic.

5: Low winter temperatures meant the simultaneous loss of outdoor socializing space, and the drying of indoor air due to use of heating, in northern tier of the US.  Where that combines with low use of facemasks, the result is now the third wave of the US COVID-19 pandemic.  In particular, the east slope of the Rockies and high plains are known for their cold, arid winter climate.  (E.g., as I write this on 10/27/2020, it’s 12F in Bismark, ND.)

Below is a picture of self-reported average facemask use, from one national survey, as of mid-summer 2020.  Lighter = less use.

(Source:  NY Times).

6:  Many people have predicted a resurgence of COVID-19 this winter, and for good reasons.  (See last post).  This now appears to be happening in North America, Europe, and Russia.  But that is NOT happening in the three large industrialized Asian nations:  China, Japan, and South Korea. (See pictures, prior post).

One factor that immediately comes to mind is near-universal mask use in those countries during this pandemic.  But a plausible humidity-related factor may be the lack of central heating in much of the housing stock of those countries.

7:  Implications for the rest of the winter.  Right now, the high plains states combine the worst of all worlds:  Frigid temperatures (so no outdoor socializing space), low absolute humidity (so low indoor relative humidity, absent intensive use of humidifiers), and an extreme aversion to wearing masks.

If you agree with my analysis, then you understand why it’s no surprise that the new COVID-19 case rates in those areas are breaking all prior records.

So far, there’s been no sign of an uptick in Virginia (thick black line on first graph in post).  But we really haven’t hit winter weather yet.  (See the note below on the absolute humidity of the air in Bismark, ND versus Washington, DC today.)  And, at least in my neighborhood (Vienna, VA), we have more-or-less universal compliance with the Governor’s executive order #63 mandating use of masks.

By report, there are plenty of places in Virginia where mask use is not the norm.  And by mid-winter, untreated indoor air may be well below the relative humidity threshold at which flu and coronavirus infections rates are boosed by dry air.  The norm for January, in Washington, DC, is 37F with 63% relative humidity (per this source).  If not otherwise treated, that would lead to 17% relative humidity for 74F indoor air (per the calculators used in the note below).  That’s well into the range that the research suggests will boost the infection rates of both flu and COVID-19.

My point is, it’s too early to celebrate yet.  If the theory laid out above is mainly correct, come January (or maybe sooner), we should see an uptick in new cases in the low-mask-use parts of the state.

The other obvious implication is that we ought to focus on keeping indoor air humidified.  Both at home and in public places.  In addition to wearing a mask.  But that’s what the just-prior post was about.

8:  N.B. regarding absolute and relative humidity.  Absolute humidity is simply the amount (mass) of water that is suspended in the air, per unit of air.  Relative humidity is the amount of water relative to the amount the air could hold if saturated.

What matters in this analysis is the relative humidity of room-temperature air. Because we spend almost all of our time indoors.

For a given amount of water in the air, relative humidity drops sharply as air temperature rises.  That’s because warm air can hold a lot more water than cold air.

For example, currently Bismark, ND is 12F with 83% relative humidityIf I were to raise that 12F air to 74F, the relative humidity would fall to under 10% relative humidity (based on this NOAA calculator.)

So, even though the outdoor relative humidity is quite high, at 12F, the absolute humidity isn’t.  There really isn’t much water in the air in Biskark, ND, and that shows up as a low relative humidity when we heat the air to room temperature.  Either way (12F, 83% relative humidity, 74F, 10% relative humidity), that air contains just 1.7 grams of water per kilogram of air (via this calculator).

By contrast, Washington DC is currently 59F with 84% relative humidity.  Using the same calculators, that’s 9 grams of water per kilogram of air.  If I were to raise that air to 74F, the relative humidity would be 50%.  In short, the air in Washington, DC has about five times as much water in it as the air in Bismark, ND.

Studies of humidity and flu use either absolute or relative humidity.  Studies that use weather data, to look at geographic or time-varying differences in infection rates, use absolute humidity of outdoor air.  Studies that look at disease transmission at various indoor conditions use relative humidity of indoor air.

But the point is, during heating season, in norther climates, places with low absolutely humidity outdoors will also be places with low relative humidity indoors.  All other things equal.  They are really looking at the same phenomenon.