Simulations used in this study.

We use the mechanistic dengue model, with different temperature inputs, numbers of locations L, and initial population sizes per location N, to address specific but related questions (column Aim). Each location was run as an independent simulation (no host movement between locations), with a total population size corresponding to the population size of each province (simulation 1), or equal across locations (simulations 2 and 3), using the same host birth and death rates, such that the only differences across locations were temperature (all simulations), and population size (simulation 1 only; see section “Further details on simulation studies” in S1 Appendix). In all cases, simulations were run assuming mean cross-protections between dengue serotypes of 6 months, 1 year, and 2 years. Outputs had a monthly resolution.