Table 1. Inferred values of parameters in models for forecasting regional epidemics of coronavirus disease, United States.

Parameter*	Estimate†	Definition
t 0	33 d	Start of transmission
σ	33 d	Start of social distancing
p 0	0.87	Social distancing setpoint
λ 0	0.10/d	Social distancing rate
β	2.0/d	Disease transmission rate
fD	0.12	Fraction of active cases reported
r	12	Dispersal parameter of NB(r,p)‡

*t₀, σ, p₀, λ₀, and β are adjustable parameters of the compartmental model; f_D is a parameter of the auxiliary measurement model; and r is a parameter for the associated statistical model for noise in case detection and reporting.
†All estimates are region-specific and inference-time-dependent. Inferences were conducted daily. These findings reflect the maximum a posteriori estimates inferred for the New York City metropolitan statistical area using all confirmed coronavirus disease case count data available in the GitHub repository maintained by The New York Times newspaper (11) for January 21–June 21, 2020. Time t = 0 corresponds to midnight on January 21, 2020. 
‡The probability parameter of NB(r,p) is constrained (i.e., its reporting-time-dependent value is determined by Appendix 1 Equation 26, https://wwwnc.cdc.gov/EID/article/27/3/20-3364-App1.pdf).