Fig 1.
United States population and estimated cumulative SARS-CoV-2 infections per 100,000 distributed across the hexagonal grid.
Panel A, B, and C: United States’, New England’s, and Connecticut’s Meta 30m population estimates on the hexagonal grid. Panel D, E, and F: United States, New England, and Connecticut cumulative infections per 100,000 persons on the hexagonal grid (March 2020–December 2021). Grey hexagons indicate no population (panels A to C) and no infections (panels D to F), void hexagons indicate no infections ever estimated (panels D to F). All maps were generated using United States, state, and county borderlines maps in the public domain from the Census Bureau, which were downloaded through the R package Tigris [30]. The shapefile generated for this analysis with the population estimates and cumulative infection estimates can be found at: https://github.com/covidestim/waves/tree/waves-manuscript/Data/data-products. Note: Numbers are given in a log 10 scale.
Fig 2.
Estimated infections per 100,000 of SARS-CoV-2 in the United States, March 2020–December 2021.
Panel A: Time series of SARS-CoV-2 infection estimates for the United States; the gray shaded areas show the first two large waves of infections. Panels B, C, D, and E: Sequence of the spatially smoothed estimates of SARS-CoV-2 infections per 100,000 associated with Wave 1 at 4 time points. Panels F, G, H, and I: Sequence of the spatially smoothed estimates of SARS-CoV-2 infections per 100,000 associated with Wave 2 at 4 time points. All maps were generated using United States, state, and county borderlines maps in the public domain from the Census Bureau, which were downloaded through the R package Tigris [30]. The shapefile generated for this analysis with the population estimates and cumulative infection estimates can be found at: https://github.com/covidestim/waves/tree/waves-manuscript/Data/data-products. For a more detailed visualization of each wave, see the S1-S2 Movies.
Fig 3.
Daily mean areal expansion rate and wavefront speed for SARS-CoV-2 infection waves for 63 days preceding the infection per capita peak.
(Wave 1: September 8, 2020–November 11, 2020; Wave 2: July 7, 2021–September 4, 2021). Top Panel: Mean daily areal expansion rate. Bottom Panel: Mean daily wavefront speed. Note: Areal expansion rate is the daily change in wave area. Wavefront speed is the distance from a vertex of the hexagonal grid at day d to the nearest point on the line at day d + 1.
Fig 4.
Progression of the wave boundary measured as the distance to the nearest point on the boundary at t + 1 (in km/day).
Each panel shows a specific date; the ones in the right column are related to the first wave and the ones in the left column to the second wave. Vertices in black show points with zero speed (i.e., their location is the same for t and t + 1. All maps were generated using United States, state, and county borderlines maps in the public domain from the Census Bureau, which were downloaded through the R package Tigris [30]. The shapefile generated for this analysis with the population estimates and cumulative infection estimates can be found at: https://github.com/covidestim/waves/tree/waves-manuscript/Data/data-products. For a more detailed visualization of each wavefront, see the S3-S4 Movies.