More on Geographic Seasonality

I am trying to get a better understanding of what factors might be responsible for the apparent geographic and seasonal variation in intensity of the coronavirus epidemic.  There has been some work on seasonality of infections by the common coronavirus strains.  There has been more work on influenza, another common respiratory virus, and that work may provide some guidance.  The earth is often divided into climate zones, with the tropics extending to about 23 1/2 degrees latitude north and south of the Equator, the sub-tropics from 23 1/2 to 35 degrees, the temperate zone from 35 to 50 degrees, and then you are in Canada, so well, what can I say.  Because of the Earth’s tilt, the northern and southern hemispheres don’t experience solar insolation and other events exactly the same.  And even in a zone, there is wide variation in topography, climate and weather.

The continental United States basically runs from 25 degrees north to a little under 50 degrees north.  Southern Florida is at around 25 degrees.  The entire Northern border is a couple of degrees under 50 degrees.  It is about 30 degrees latitude along much of the southern border.  40 degrees runs south of the New York metro, along the southern Pennsylvania border, through the middle of Ohio, Indiana and Illinois, out along northern Kansas and Colorado and through the upper part of northern California.  So most of the US is temperate, with some subtropical covering the south.  We have a wide variety of climates and changes in measures like temperature, humidity and precipitation.

One of the most informative studies I have found so far comes from the University of Minnesota in 2013.  (UM Study)  They looked at studies and data sets from 78 global locations and at patterns of influenza.  About 40% of the sites were located in the tropics.  Some regions had two influenza peaks in a year, mostly in East and Southeast Asia and equatorial Central and South American countries.  15 of 17 sites with these dual peaks had a primary peak and secondary, but smaller, peak.  They compared monthly values for temperature, specific humidity, precipitation and solar radiation with influenza activity.  Strikingly, the patterns of influenza in regard to the weather measures were not consistent.  Low temperatures, low solar radiation and low specific humidity were linked to epidemics in higher latitudes (i.e., temperate zones), while high precipitation, high specific humidity and high relative humidity were associated with epidemics in lower latitudes.  The relationship among factors, latitude and influenza was complex.  There appeared to be threshholds of some measures, such as specific humidity and temperature, which were correlated with strength of epidemic.

The pattern is so complex that it is even hard to describe.  There were regions with cold/dry condition peaks and ones with rainy/humid peaks.  In general, areas with at least one-month of low levels of specific humidity and temperature had seasonal influenza peaks during those months.  Sites that had high average levels of specific humidity and temperature tended to have peaks during the most humid and rainy months.  Mid-latitude areas, which includes, most of the US, were the hardest to predict with the factors used, due to high climate variability.  The authors note that other factors, such as changing population-mixing patterns, more time indoors or outdoors, and the effect of climate on both the virus and on host factors like dryness of nasal passages might be the cause of changes in influenza peaks.

If coronavirus is similar to influenza, you are left with an unclear sense of how to understand geographic and seasonal variation.  There definitely appears to be some geographic pattern within the US, and perhaps it is related to measures like specific humidity, temperature and precipitation, but the variation may occur at a micro-level that is harder to detect.  I am left still searching for a clearer explanation.