Spring Weather and COVID-19 Deaths in the U.S.

This study used statistically robust regression models to control for a large set of confounders (including county-level time-invariant factors and time trends, regional-level daily variation, state-level social distancing measures, ultraviolet light, and levels of ozone and fine particulate matter, PM2.5) to estimate a reliable rather than simple regression for the impact of weather on the most accurately measured outcome of COVID-19, death. When the average minimum temperature within a five-day window increased by one degree Fahrenheit in spring 2020, daily death rates in northern U.S. counties increased by an estimated 5.1%. When ozone concentration over a five-day window rose by one part per billion, daily death rates in southern U.S. counties declined by approximately 2.0%. Maximum temperature, precipitation, PM2.5, and ultraviolet light did not significantly associate with COVID-19 mortality. The mechanism that may drive the observed association of minimum temperature on COVID-19 deaths in spring months may be increased mobility and contacts. The effect of ozone may be related to its disinfectant properties, but this requires further confirmation.