Source: Underlying data are from Johns Hopkins University, via the NY Times Github COVID-19 data repository. I have once again had to expand the vertical scale of this graph since the last time I published it. Three days ago.
This post has turned into quite a treatise. Let me cut to the chase. Here’s my hypothesis, regarding the last two waves of the COVID-19 pandemic in the US.
It’s not the heat, it’s the (lack of) humidity.
- Hot weather + no indoor mask use = second wave of US pandemic, centering on southern states.
- Cold weather + no indoor mask use = third wave of US pandemic, centering on northern states.
In short, I think that dry indoor air plus no mask use is going to be a toxic combination this winter. And I’m afraid we’re seeing an acid test of that where the cold, arid winter climate of the US high plains and eastern slope Rockies intersects Republican anti-mask sentiment.
(Source: NY Times). The mask-hating Southern states have avoided masks all along. But they didn’t get into real trouble until air conditioning season hit. The mask-mocking upper Midwest states, ditto — until heating season hit. So it wasn’t the low mask use, by itself, that got them into big trouble. It was widespread lack of mask use when low-humidity indoor settings were common.
Empirically, low humidity is associated with easier spread of most viral respiratory diseases. There’s a deep, broad body of research that backs up that claim. Viruses that have a “lipid coat” are helped by dryer air. And that class of viruses includes COVID-19.
What does mask use have to do with it? You know how uncomfortable it is to wear a mask, outside, in the heat and humidity? The mask gets soaking wet, and you breath in nothing but ultra-humidified air. That’s a drawback in a humid-air environment. But in an environment with room-temperature dry air, that additional humidity may allow your nasal membranes to function normally, and do their job of protecting you. Which they will not do, or do poorly, if they themselves dry out.
This is all by way of preparing for what looks like a coming mid-winter peak of COVID-19. It’s not as if a seasonal peak is unexpected (e.g., Post #714, June 12, 2020). It’s not as if Dr. Fauci hasn’t warned us. So I’m buying an extra humidifier, ahead of time. Because if this is correct, and becomes widely known, we’re (of course) going to have a shortage of humidifiers.
And so, I have purchased a second humidifier for my home. And (see below) 40% relative humidity seems to be one of those “magic numbers” that shows up in a lot of different places, for a lot of different reasons. And so, absent any other information, that’s what I’m shooting for. With both humidifiers running.
A cat is not a hygrometer. Except sometimes. COVID-19 should have a winter peak. Except that’s not happening everywhere.
Years ago, we used used our cat as a hygrometer. When the cat’s fur started attracting bits of dust and lint off the floor, that meant the air was too dry, and we needed to start running a humidifier.
Looking ahead to our first winter of the COVDI-19 era, I’m ditching the cat in favor of a proper hygrometer to track indoor humidity. I’m buying a second humidifier. I’m upping my game on all the other tricks I use to keep indoor air humidified. I’m shooting for a minimum 40% relative humidity. And I’m wondering whether humidifiers are going to be the next COVID-generated shortage, if COVID-19 infection rates increase as room-temperature indoor air dries out.
There is a point to all of that. Low humidity is associated with easier spread of most viral respiratory diseases. Again, there’s a deep, broad body of research that backs up that claim. This is not news. (At some point, I had an extensive literature review here, but does anybody every actually read that when I write it up? For the curious, let me give use one extensive scholarly reference, to show how many different viruses do better in dryer air. See Table 1 in this reference.)
This not the way we normally think of moisture and pathogens. We normally think about mold and bacteria, both of which thrive in moist environments. But many lines of research — from direct experiments on up — show that viral respiratory illnesses spread more easily with dry air.
But there’s more to this than just a lot of hot air. Albeit dry, hot air.
Here’s my casual observation.
There’s a fairly sharp divide between areas where COVID-19 appears to be ramping up broadly, and areas where it is not. As all of those areas move into winter. So, while we expect to see a winter peak in COVID-19, it doesn’t look like that’s happening everywhere. Or, at least, not yet.
Across the northern hemisphere, COVID-19 case loads are ramping up in North America, Europe, and Russia. COVID-19 case load are not ramping up in China, Korea, or Japan. Like so, three graphs for three countries, current as of 10/25/2020.
In the US, with some exceptions, case loads are ramping up sharpy in the northern strongly-anti-mask-use states. Just as the cold weather hits. That is , the Dakotas are setting new records. New England is not. By and large, COVID-19 case loads are not ramping up sharply in northern states with high mask use.
And now I turn away from what I believe to be facts, and move into speculation.
That apparent divide based on mask use may get back to the issue of humidity. Because, as near as I can tell, the direct, physical effect of dry air, on (say) flu virus itself, is relatively modest. Flu virus survives a little longer, and travels a little further, in dry air. But the emphasis is on little. Those effects don’t appear to be nearly large enough to explain the sharp seasonal peak in flu.
My reading of the literature is that the main driver of the winter peak of flu has more to do with what goes on inside the human body during winter than it has to do with the direct effect of dry air on the virus. And I’ll bet masks play a role here. Not just by reducing viral load, but by increasing your resistance as well. They keep your nasal membranes from drying out in warm, dry indoor air. And thus allow your nasal membranes to do what they are supposed to do, which is to protect you from inhaled pathogens.
Take the Dakotas and surrounding areas. The interesting question is, why now? The non-mask-wearing Dakotas have been that way all along. There’s nothing new about their lack of mask use. The Governor of South Dakota has disparaged mask use throughout the pandemic. Certainly, the week-long maskless drunken party that was the Sturgis motorcycle rally didn’t help. But you have to wonder if something else is at play to yield such a broad-based area-wide increase in cases. In short why, only now, as the weather turns cold, are cases going through the roof there. But not, say, in Vermont?
Look at it this way. In the air-conditioned parts of the country, the anti-mask states paid their penalties in mid-summer. That was the second wave of the coronavirus pandemic in the US. But the un-air-conditioned anti-mask areas skated by, at that time. And now, as the weather cools, it’s time for the cold-climate anti-mask states to suffer. That’s the third wave of the pandemic in the US.
It’s almost as if you could take the map of mask use above, cross it with a map of heating degree-days and cooling degree-days, and predict how the third wave will spread.
Flu is seasonal. Non-deadly coronaviruses are seasonal. It’s likely that COVID-19 will be too.
Source: Graph from blog.sas.com, original data from CDC Pneumonia and Influenza Mortality Surveillance.
Seasonality of common human coronaviruses (other than COVID-19) in Stockholm, Sweden:
Source: Potential impact of seasonal forcing on a SARS-CoV-2 pandemic DOI: https://doi.org/10.4414/smw.2020.20224 Publication Date: 16.03.2020 Swiss Med Wkly. 2020;150:w20224 Neher Richard A., Dyrdak Robert, Druelle Valentin, Hodcroft Emma B. Albert J.
Any number of credible sources have suggested that we may see a resurgence of COVID-19 cases this winter. That’s not certain, but it’s a pretty good guess. That’s been in the news enough that it’s not worth giving a lengthy list of citations. Dr. Fauci said it, that’s good enough.
We all know that wintertime is flu season. And, for all the common (non-lethal) human coronaviruses, wintertime is coronavirus season as well. So it’s a pretty good bet — but not a certainty — that wintertime with be COVID-19 season.
But why is flu seasonal?
The exact reasons for the seasonal peak in flu cases are a bit hazy. For something that happens regularly every year, and causes such great economic damage, morbidity, and mortality, there’s not a lot of hard, quantitative data on exactly which factors drive the increase in wintertime flu. And to what extent each factor matters. And, for sure, some of what you commonly read appears to be little more than folklore.
But that’s exactly what you want to know, if you plan to take steps to reduce your risk of contracting COVID-19 this winter. So let me walk through the a) the common explanations for flu season, and b) the available data for factors affecting spread of COVID-19.
In a nutshell, in order of what I think matters:
- Dry nasal membranes provide a less effective barrier against pathogens.
- The size of this effect is unknown in humans, but in mice, experiments show a large impact.
- Masks help keep nasal membranes moist.
- Cold air, and dry air both allow flu viruses to survive longer.
- In addition, dry air allows droplets to travel further because the droplets shrink in dry air.
- Empirically, humidity matters more than temperature.
- But in terms of effect size, neither of these appears to be able to explain the seasonal peak in flu.
And then, things that I think have a much smaller or negligible effect, but that you will see mentioned in the popular press.
- Vitamin D levels drop in wintertime, and that affects the immune system.
- Empirically, that drop is modest in most of America, and does not appear to affect prevalence of other infectious diseases other then flu/pneumonia.
- We spend more time inside, in winter.
- Yeah, a whopping half-hour-a-day more, for the average US adult. That doesn’t seem to be anywhere near large enough to account for seasonal flu peaks.
Humid air means shorter virus lifetimes.
First, viruses simply die faster higher relative humidity, for the range of humidities you would find in a typical living space. For viruses with a lipid envelope (like coronaviruses), survival times in the air are higher at low relative humidity. This has been tested and found true for a wide range of such viruses. (See Table 1 of this study.)
The estimated magnitude of this effect, for COVID-19, is significant but not overwhelming. Per this graph from US DHS, raising relative humidity from 20% to 40%, in a warm (74F) indoor space, cuts the half-life of coronavirus on solid surfaces by just over 20%, from 14 to 11 hours.
Caveat: Other estimates for other airborne viruses show a larger effect. Here’s what I believe to be a typical result. Scientists in this study had one dummy, infected with flu, cough into a closed room. (And by dummy, I mean literally a mannequin.) Other dummies sat around breathing in that room for an hour. At roughly 20% relative humidity, 75% of the collected virus remained infectious. At just over 40% relative humidity, 20% of the collected virus remained infectious. Almost all of the reduction occurred in the first 15 minutes.
Surprisingly, many people have found that 40% relative humidity is roughly the threshold where airborne viruses start to die off. As in this reference.
Humid air means shorter droplet airborne lifetimes.
Second, higher humidity allows droplets to maintain their weight and sink to the ground faster. Lower humidity encourages droplets to shrink and form aerosols. I read though several studies, and was not able to understand them well. As near as I can tell, this also has a relatively small magnitude.
The observed relationship between humidity and disease transmission appears much larger than these simple physical impacts would suggest.
Third, the strongest evidence for a link between humidity and disease transmission comes from epidemiology. In this case, that’s the statistical analysis of the correlation between rates of infection and factors such as temperature and humidity.
A single such study could be discounted. But as far as I can tell, almost every study that has looked at that has concluded that dry air leads to substantially higher rates of disease transmission.
” … approximately half of the average seasonal differences in US influenza mortality can be explained by seasonal differences in absolute humidity alone. ”
Dry air means dry nasal membranes, which do not function as well as moist nasal membranes.
Third, dry air dries out the nasal membranes and makes them less effective at fighting off infectious material. This has been demonstrated empirically mice, with a very large effect size (reference). I found nothing in the literature to provide any estimate for humans, so belief in the large size of that effect is based on mouse models. That said, I can cite experts who point to this issue as being a potentially important one, in this meta-analysis of 10 recent studies of humidity and COVID-19 spread.
One thing that might contribute: Vitamin D levels
Unless you take supplements, you rely on sunshine hitting your skin to produce vitamin D. Vitamin D plays many roles in the body, including in the immune system. And so, the theory goes, maybe reduced levels of vitamin D created a sufficiently weakened immune system to allow flu to thrive. It’s not clear that anyone has quantified that, nor are there seasonal wintertime peaks in other infectious diseases that presumably would be affected by this vitamin D deficiency.
One thing that probably doesn’t matter: Average time spent indoors.
In part, flu season may occur due to behavioral factors. People are indoors more in the winter, or so the theory goes. And so, opportunities for spread are greater than in summer.
But is that true, in the modern world? Or is that a reflection of the way the world was half-a-century ago? And if it’s true, how much more time do we spend indoors in winter, as opposed to summer?
First, in the hotter parts of the US, people spend a lot of mid-summer hours indoors due to extremes of hit and/or humidity. There was some speculation that the COVID-19 outbreaks in Florida, Arizona, and Texas were due to a “reverse summer effect”, as reported in the Washington Post. The theory was that cases rose there in the summer, instead of falling, as people were driven inside by the heat.
So, for sure, the idea that wintertime means higher indoor time likely predates the widespread use of air conditioning. I grew up in the 1960’s in an un-air-conditioned house in Manassas, VA. That was the tail end of the era when restaurants and theaters would advertise that they were air conditioned. Summertime meant open windows and fans running. Constantly moving fresh are was the only way to try to stay cool. (It was also an era when they would shut down the Federal government if it got too hot in Washington, DC).
Source of graph: streets.mn.
Today, literally 100% of new home construction in the South includes central AC. You probably could not sell a new home without it. I can’t recall the last time I walked into a store or restaurant that did not have AC, and I’ll bet that there is not even one commercial establishment in Fairfax County that isn’t air-conditioned.
The fact is, Americans spend almost none of their time outdoors. Typically estimated at less than 10 percent of time, often estimated at around 5 percent of time. And it has been that way for decades.
Source: “It’s about time: A comparison of Canadian and American
time–activity patterns”, JUDITH A. LEECH,WILLIAM C. NELSON, RICHARD T. BURNETT, SHAWN AARON, AND MARK E. RAIZENNE, Journal of Exposure Analysis and Environmental Epidemiology (2002) 12, 427 – 432
In the modern world, there’s almost no winter-summer difference in the amount of time that U.S. adults spend outdoors. The average US adult spends about half-an-hour more outdoors in summer, compared to winter (per the study above). That matches several other surveys that I looked at. The idea that we get a wintertime flu peak due to greater indoor time in the winter does not pass a simple test of face validity. The difference in indoor time, on average, is far to small to account for the vast increase in flu cases.
There is ample evidence that humidity strongly affects flu and COVID-19 disease transmission rates. For wintertime US humidity levels, it looks like 20% relative humidity allows for much easier spread of the virus than does (say) 40% to 50% relative humidity.
In part, that’s through the direct effect of dry air on the virus-containing droplets, and on virus survival times. But it certainly appears to me that the more dominant factor is the effect of low humidity on the human body. And, in particular, in making the nasal membranes less robust to attack by viral pathogens.
For me, the bottom lines are pretty clear. I’m going to maintain 40% relative humidity at home. I would suggest that if you are in charge of an indoor space where people gather — church, office, or store — that you do the same for that space.
And my reading of the data is that dry indoor air, plus a lack of mask use, is going to be a toxic combination this winter. Not just because masks prevent the spread and inhalation of droplets and aerosols. But also because masks prevent dry indoor air from drying out your nasal membranes, and so short-circuiting your primary defense against airborne pathogens.
The mask-hating South had their trial-by-fire this summer. The mask-hating upper Midwest is undergoing it now. Let’s hope we can get a clue. It’s far cheaper to enforce mask use, and to humidify indoor spaces, than to go through another series of shutdowns.