Table 1: Parameters for simulation scenarios.

Transmission probability, β is per 5-minute interaction with an infectious individual. All other epidemic transition parameters are per day, giving an average latent period of 3 days and an average infectious period of 7 and 12 days for asymptomatic and symptomatic infections, respectively.

Epidemic Model
β	0.003 – 0.009	probability of transmission per 5-minute interaction*
π ex	0.002	probability of external infection per day
π ai	0.3	probability of asymp. infection15
|I0|	1	number of initial infections
σ a	1/3	transition probability: exposed to asymp. infectious16
σ s	1/3	transition probability: exposed to symp. infectious16
γ a	1/7	transition probability: asymp. infectious to recovered17
γ s	1/12	transition probability: symp. infectious to recovered17
Testing
π se	0 – 0.96	sensitivity, time-varying*18
π sp	0.99	specificity20
π sy	1 day	delay between symptom onset and symp. testing
π sc	0.01	probability non-compliant with scheduled testing
π sy	0.25	probability non-compliant with symp. testing
π fs	0.005	probability non-infectious present for symp. testing
Isolation
τ id	1 day	delay between testing and entering isolation
τ ic	Beta(5, 0.5)	isolation compliance
τ ip	10 days	isolation period21
Transmission Mitigation
η fm	0.15	reduction in transmission probability for mask wearing22
η sd	0.18	reduction in transmission probability for social distancing22
π fm	0–1	proportion of the population wearing face masks*
π sd	0–1	proportion of the population social distancing*
π im	0 – 0.4	proportion of the population immune*

^*Transmission probabilities, proportion of the population wearing face masks, proportion of the population social distancing, and proportion of the population immune varied across simulations. Test sensitivity was time-varying. An overview of the epidemic model is shown in Figure 2. For additional details on parameter values, see the corresponding sections of the Methods.