tetano
Editor, Senior Moderator
Proc Biol Sci
. 2026 Mar 18;293(2067):20251729.
doi: 10.1098/rspb.2025.1729.
Spatio-temporal spread of COVID-19 over three variant waves in the continental United States
Viviane Callier[SUP] 1 [/SUP], Rob Deardon[SUP] 2 3 [/SUP], Cécile Viboud[SUP] 4 [/SUP]
Affiliations
Understanding the spatio-temporal spread of infectious diseases is important to improve control, yet most studies lack granular epidemiological data and are complicated by heterogeneities in pre-existing immunity. The COVID-19 pandemic provides a unique opportunity to study the invasion of a novel pathogen in a naive population. Here we apply correlation statistics and mathematical modelling to daily city-level COVID-19 case data between 2020 and 2022 to dissect the spatial diffusion of different virus variants across the United States. We find that between-city correlation in case incidences varies with geographical distance, supporting the idea that infections are propagated from neighbour to neighbour up to a radius of 1000-1500 km, a process we call diffusive spread. We fitted mechanistic transmission models to onset times and find that models combining geographical distance with the population size of recipient cities produce the best fit to the winter 2020-2021 and Omicron waves, while indicators for pre-existing immunity and vaccination do not improve model fit. A similar model does not capture transmission dynamics for the Delta wave. In both the winter 2020-2021 and Omicron waves, a few large cities, but also smaller cities, mostly in the Midwest, disproportionately contributed to the onward spread of COVID-19. These could be useful for sentinel surveillance.
Keywords: COVID-19; epidemics; infectious disease modelling.
. 2026 Mar 18;293(2067):20251729.
doi: 10.1098/rspb.2025.1729.
Spatio-temporal spread of COVID-19 over three variant waves in the continental United States
Viviane Callier[SUP] 1 [/SUP], Rob Deardon[SUP] 2 3 [/SUP], Cécile Viboud[SUP] 4 [/SUP]
Affiliations
- PMID: 41844243
- DOI: 10.1098/rspb.2025.1729
Understanding the spatio-temporal spread of infectious diseases is important to improve control, yet most studies lack granular epidemiological data and are complicated by heterogeneities in pre-existing immunity. The COVID-19 pandemic provides a unique opportunity to study the invasion of a novel pathogen in a naive population. Here we apply correlation statistics and mathematical modelling to daily city-level COVID-19 case data between 2020 and 2022 to dissect the spatial diffusion of different virus variants across the United States. We find that between-city correlation in case incidences varies with geographical distance, supporting the idea that infections are propagated from neighbour to neighbour up to a radius of 1000-1500 km, a process we call diffusive spread. We fitted mechanistic transmission models to onset times and find that models combining geographical distance with the population size of recipient cities produce the best fit to the winter 2020-2021 and Omicron waves, while indicators for pre-existing immunity and vaccination do not improve model fit. A similar model does not capture transmission dynamics for the Delta wave. In both the winter 2020-2021 and Omicron waves, a few large cities, but also smaller cities, mostly in the Midwest, disproportionately contributed to the onward spread of COVID-19. These could be useful for sentinel surveillance.
Keywords: COVID-19; epidemics; infectious disease modelling.