Thread by @jnordvig

There is still a debate about whether coronavirus will die out naturally due to warmer whether as we enter the summer period, in the US, Europe, etc. But there is remarkably little detailed research on this. A few observations = THREAD

Xu R., Rahmandad H., Gupta M., DiGennaro C., Ghaffarzadegan N., Amini H., Jalali M.S. (2020). Weather
Conditions and COVID-19 Transmission: Estimates and Projections. The finding here is a MODERATE IMPACT

projects.iq.harvard.edu/files/covid19/…

Academics such as @mlipsitch have also made the point that there is an important interaction between herd immunity and seasonality => seasonality will be less important when there is a low level of herd immunity still the message from serology research)

ccdd.hsph.harvard.edu/will-covid-19-…

We have also played with the US statistics @ExanteData, and we are not able to find a strong link to temperature or other weather variables. Population density seems to dominate most other variables (something we can get back to in a different thread).

Here I would just draw on the evidence we have globally currently: There are a lot of VERY HOT places that have problematic outbreaks. Some examples below...

Look at Qatar, new cases are accelerating, and the hit ratio is rising too, pointing to the underlying outbreak accelerating. Today it is 31C/88F there, and and recent averages have been above that.

Look at Brazil. Cases are accelerating (although testing data sporadic, so hard to judge role of test dynamics [no hit ratio]. On the weather, it is cooling a bit now, but we saw acceleration during March-April, which I understand was very hot in Sao Paulo (local color welcome).

Kuwait, UAE, Saudi, Bahrain also look problematic. Hence, that makes you think that those places that are so hot that people stay inside (with AC) are actually high risk.

You can look at charts for 60 countries on your own here:

https://twitter.com/ExanteData/status/1261778674287869952

All told, the weather in the US and Europe is getting warmer. And some academic studies suggest is will help on the margin. But the evidence from the Middle-East (and perhaps Brazil and others too) suggests that warmer weather only helps in certain conditions.

I worry that US states that may get very hot, with a lot of people working indoors in AC'ed environment, could face elevated risk, especially if the re-open process accelerates, and the return to work process leads to crowded offices/factories/shops (this has not happened yet...)

To put it in simple terms: A reasonable hypothesis is that the warmer weather may help, if it leads people to spend more time outdoors. But if the warmer weather causes people to congregate indoors, including in workplaces, then that may invalidate the 'normal statistical link'

I will leave it at that. Warmer weather may help 'on average' getting reproduction numbers (Rs) down. But it will not help in those jurisdictions where very hot weather causes everybody to congregate inside. Policy makers should be pretty careful in those areas. END.

sorry whether=weather in second line...