Fig 1.
High-level overview of the common structure of both models used in this study.
Dark nodes are observed, light nodes are inferred. See S2 Appendix for a detailed model graph for each individual model
Fig 2.
The inferred seasonal R multiplier Γ(t) of the combined models estimate, with 50% and 95% credible intervals.
Gray boxes indicate data range of each dataset, i.e. 22nd January to 30th May 2020 for Brauner et al. and 1st August 2020 to 9th January 2021 for Sharma et al. The zero-width credible intervals around April 1 and October 1 owe to the fact that we model Γ(t) as a seasonal multiplier for R relative to the mid-spring and mid-fall, respectively, which implies that Γ(t) is assumed to be exactly equal to one for these dates.
Fig 3.
Posterior distributions of the R reduction on July 1 relative to January 1 with median, 50% and 95% credible intervals.
Fig 4.
Comparison of the inferred peak-to-trough R reduction effect of seasonality (combined from both models) to the NPI reductions inferred by Brauner et al. [23], with 50% and 95% CIs.
The seasonal effect is lower than the combined NPI effect but higher than or comparable to the individual NPI effects.