COVID-19 spread and Weather in U.S. states: a cross-correlative study on summer-autumn 2020
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (ScreenIT)
Abstract
An effect of weather on sars-cov-2 transmission is regularly proposed as a putative cause of unexplained fluctuations of covid-19 new cases, but clear data supporting this hypothesis remains to be presented. Here I measured longitudinal time-series correlations between outdoor temperature, humidity and covid-19 reproduction number (Rt) in the 50 U.S. states (+DC). In order to mitigate the confounding influence of varying social restriction measures, the analysis spans a 5-month period during summer and autumn 2020 when restrictions were comparatively lower and more stable. I used a cross-covariance approach to account for a variable delay between infection and case report. For a delay near 11 days, most U.S. states exhibited a negative correlation between outdoor temperature and Rt, as well as between absolute humidity and Rt (mean r = −0.35). In 21 states, the correlation was strong (r < −0.5). Individual state data are presented, and associations between cold and/or dry weather episodes and short-term new case surges are proposed. After identifying potential confounding factors, I discuss 3 possible causal mechanisms that could explain a correlation between outdoor weather and indoor disease transmission: behavioral adaptations to cold weather, respiratory tract temperature, and the importing of outdoor absolute humidity to indoor spaces.
Article activity feed
-
SciScore for 10.1101/2021.01.29.21250793: (What is this?)
Please note, not all rigor criteria are appropriate for all manuscripts.
Table 1: Rigor
Institutional Review Board Statement not detected. Randomization not detected. Blinding not detected. Power Analysis not detected. Sex as a biological variable not detected. Table 2: Resources
Software and Algorithms Sentences Resources This approach is similar to normalized cross-covariance (e.g. xcov() Matlab function), but here the delayed variable (Rt) is never clipped at the end of the time series, i.e. n = 154 data value pairs is always verified. Matlabsuggested: (MATLAB, RRID:SCR_001622)I used Matlab® for all analyses and graphics. Matlab®suggested: (MATLAB, RRID:SCR_001622)Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data …
SciScore for 10.1101/2021.01.29.21250793: (What is this?)
Please note, not all rigor criteria are appropriate for all manuscripts.
Table 1: Rigor
Institutional Review Board Statement not detected. Randomization not detected. Blinding not detected. Power Analysis not detected. Sex as a biological variable not detected. Table 2: Resources
Software and Algorithms Sentences Resources This approach is similar to normalized cross-covariance (e.g. xcov() Matlab function), but here the delayed variable (Rt) is never clipped at the end of the time series, i.e. n = 154 data value pairs is always verified. Matlabsuggested: (MATLAB, RRID:SCR_001622)I used Matlab® for all analyses and graphics. Matlab®suggested: (MATLAB, RRID:SCR_001622)Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).
Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:Limitations: The present results are highly dependent on the quality of raw data. New covid-19 case counts are manually collected from states’ public data [3], which can contain variable delays between states, but also even within a state, if the reporting policy changed through months in the pandemic. Large, prolonged surges can amplify issues, with increasing testing delays [6] or reporting backlogs, that add up to variable ICR delays, between and within states. The method used here is quite sensible to “spikes” in the data, as outliers have a heavy contribution to product-moment sums as computed in Pearson’s r. Here I chose to remove only the most disrupting spikes in 3 states (NC, MA, TX), otherwise sticking to the raw covid-19 new cases data (+ 7-day averaging). Other approaches with systematic, stronger smoothing are possible [see e.g. 10] and might allow higher correlation coefficients. In states with small populations and/or very low case counts, Rt becomes more dependent on individual, stochastic cluster emergences, which can hide any potential weather determinism of transmission. For this reason, I do not expect that the present approach translates as well to infra geographic levels, e.g. counties. State meteorological records from the COVID-19 Open-Data [11] were used as-is, and would also benefit from a careful re-examination. Ideally, hourly T and RH should be used to compute hourly AH and WBT before averaging to daily values. Also, the way individual weather sta...
Results from TrialIdentifier: No clinical trial numbers were referenced.
Results from Barzooka: We did not find any issues relating to the usage of bar graphs.
Results from JetFighter: We did not find any issues relating to colormaps.
Results from rtransparent:- Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
- Thank you for including a funding statement. Authors are encouraged to include this statement when submitting to a journal.
- No protocol registration statement was detected.
-
