The COVID-19 pandemic, caused by SARS-CoV-2, has had a significant impact globally, with the U.S. experiencing a disproportionately high number of cases and deaths. Laboratory experiments have shown decreased SARS-CoV-2 stability in warmer temperatures, higher humidity, and simulated sunlight. This, combined with the known seasonality of influenza and other coronaviruses, suggests a potential link between meteorological conditions and SARS-CoV-2 transmission. Previous studies have yielded conflicting results regarding the association between meteorological factors (air temperature, humidity, UV radiation) and COVID-19 cases or reproduction numbers, often due to limitations such as short observation periods, inconsistent data reporting, and lack of sufficient control for confounding factors. This study aims to address these limitations by using a more robust methodology to estimate the association between meteorological factors and SARS-CoV-2 transmissibility, as measured by the effective reproduction number (Re), while accounting for a wide range of potential confounders. Understanding this relationship is crucial for informing public health interventions and communication strategies.

Numerous preliminary studies have explored the relationship between meteorological factors and COVID-19 case numbers, with varying results. Some studies found positive associations, while others found negative or no associations. These discrepancies stem from several factors, including the use of daily new confirmed cases (which are susceptible to underreporting and reporting delays), variations in the lag between infection and symptom onset, and inconsistent control for confounding variables. A few studies have used the reproduction number (R) as an outcome measure, providing a more direct estimate of transmissibility. However, early studies focusing on the basic reproduction number (R0) often found no association, potentially due to the limited range of meteorological conditions observed during the early stages of the pandemic. Furthermore, the fraction of cases or deaths attributable to meteorological conditions remained unclear, with some modeling studies suggesting that the impact of humidity might be overshadowed by population immunity. This study builds upon previous research by utilizing a more comprehensive approach, addressing the limitations of earlier studies and offering a more nuanced understanding of the interplay between meteorological factors and SARS-CoV-2 transmission.

This study analyzed data from 2669 U.S. counties with at least 400 cumulative cases by December 31, 2020. Daily meteorological data (air temperature, specific humidity (SH), and ultraviolet (UV) radiation) were obtained from the North America Land Data Assimilation System and the European Centre for Medium-Range Weather Forecasts ERA5 climate reanalysis. County-level COVID-19 case and death data were obtained from the John Hopkins University Coronavirus Resource Center. Other data included geographic location, population density, demographics, socioeconomic factors, healthcare resources, and air pollution levels. The effective reproduction number (Re), which removes the impact of population immunity on disease transmission, was estimated using a dynamic metapopulation model informed by human mobility data. This model accounts for unreported infections, reporting delays, and county-to-county movement. The association between meteorological factors and Re was analyzed using a generalized additive mixed model, adjusting for spatiotemporal variations and other potential confounders. The attributable fraction (AF) of Re attributable to each meteorological factor was calculated to quantify the contribution of each factor. Sensitivity analyses were conducted to assess the robustness of the results under different modeling choices and with additional adjustments for various factors.

The study found significant associations between meteorological factors and SARS-CoV-2 transmission. Specifically, lower air temperature (within the 20–40 °C range), lower specific humidity, and lower ultraviolet radiation were significantly associated with increased Re. The relationship between specific humidity and Re was non-linear, with higher humidity associated with decreased transmission, except for a stable trend from approximately 7 to 12 g kg−1. The attributable fractions (AF) of Re were 3.73% for temperature, 9.35% for specific humidity, and 4.44% for UV radiation, totaling approximately 17.5% attributed to meteorological factors. The AFs were generally higher in northern counties than in southern counties and highest in winter months. Sensitivity analyses indicated that the estimated relationships were generally robust under different modeling choices and after adjusting for various demographic, socioeconomic, and environmental factors. The exception was the temperature effect, which increased substantially after controlling for long-term PM2.5, suggesting confounding.

The findings indicate that cold and dry weather and low levels of UV radiation are moderately associated with increased SARS-CoV-2 transmissibility in the U.S., with specific humidity playing the most significant role. This is consistent with laboratory evidence showing decreased SARS-CoV-2 stability under conditions of higher temperature and humidity and the impact of cold, dry air on mucociliary clearance. The impact of UV radiation might be mediated through its effects on viral inactivation and immune function. The estimated total contribution of meteorological factors (17.5% of Re) is comparable to a previous modeling study. The higher contribution of specific humidity compared to temperature might reflect the close correspondence between indoor and outdoor humidity levels, especially in developed countries where individuals spend a significant portion of their time indoors. While the study suggests an association, further research is needed to definitively establish causality. Understanding these relationships is crucial for developing effective public health interventions, particularly during winter months.

This study demonstrates a moderate association between meteorological factors and SARS-CoV-2 transmission in the U.S., with specific humidity being the most influential. These findings highlight the importance of considering meteorological conditions when implementing public health interventions, especially during colder months. Future research should investigate the heterogeneity of these associations across different SARS-CoV-2 variants and further explore the causal mechanisms involved. Additional research using individual-level data would strengthen the findings and help eliminate the possibility of ecological fallacy.

This study is an ecological study, which is susceptible to the ecological fallacy. The analysis was limited to the period from March 15 to December 31, 2020, and therefore might not be fully generalizable to other time periods. Data limitations prevented the exploration of potential heterogeneity of associations for different SARS-CoV-2 variants. The model relied on several assumptions, including the constant relative transmissibility of asymptomatic infections and the stability of model parameters over time. The study could be strengthened by analyzing individual-level data and incorporating additional factors that might influence transmission.