Table 3.

Two-populations model. Optimal model parameters (median and interquartile range). Euclidean-norm minimization with ${\alpha }_1=0,{\alpha }_2=1$. Fixed fraction of hospitalized who died ${\omega }^y=0.0991,{\omega }^o=0.3248$ (see text for justification). Variation range used in the optimization algorithm (initial parameter and initial-condition guesses were uniformly sampled within the ranges shown).

Parameter	Symbol	Median (interquartile range)	Initial value
Population	$N$	127,575,528
Initial populations (young)	($S_0^y;I_0^y;H_0^y;R_0^y;A_{R_0}^y;D_0^y$)	(122,472,507 ($S_0^y∕N=0.96$); 1370; 132; 2; 0; 3)
Initial populations (old)	($S_0^o;I_0^o;H_0^o;R_0^o;A_{R_0}^o;D_0^o$)	(5,103,021 ($S_0^o∕N=0.04$); 207; 39; 2; 0; 2)
Transmission rate ($y$) [per day]	${\beta }_{SI}^y$a	0.3085 (0.2862–0.3295)	$c\in U[0,1]$
Transmission rate ($y$) [per day]	${\beta }_{SA}^y$a	0.2897 (0.2725–0.3126)	$c\in U[0,1]$
Transmission rate ($o$) [per day]	${\beta }_{SI}^o$a	0.5605 (0.5419–0.5796)	$c\in U[0,1]$
Transmission rate ($o$) [per day]	${\beta }_{SA}^o$a	0.6471 (0.6297–0.6625)	$c\in U[0,1]$
Social distancing ($y$) effect	${\eta }_{SI}^y$	0.5283 (0.5110–0.5458)	$c\in U[0,1]$
Social distancing ($y$) effect	${\eta }_{SA}^y$	0.5345 (0.5121–0.5533)	$c\in U[0,1]$
Social distancing ($o$) effect	${\eta }_{SI}^o$	0.5606 (0.5407–0.5790)	$c\in U[0,1]$
Social distancing ($o$) effect	${\eta }_{SA}^o$	0.6068 (0.5876–0.6250)	$c\in U[0,1]$
Latent period [days] ${\tau }_l$	$1∕{\sigma }_1$	2.9699 (2.9673–2.9727)	$1∕k,k\in U[2.75,3.25]$
Preclinical period [days] ${\tau }_p$	$1∕{\sigma }_2$	2.4695 (2.4678–2.4711)	$1∕k,k\in U[2.25,2.75]$
A/P partitioning ($y$)	${\varphi }^y$	0.3710 (0.3405–0.3996)	$c\in U[0,1]$
A/P partitioning ($o$)	${\varphi }^o$	0.2581 (0.2442–0.2740)	$c\in U[0,1]$
Infectivity period (I, $y$) [days] ${\tau }_{\text{inf}}^{I,y}$	$1∕M^y$	4.1815 (4.1167–4.2442)	$1∕k,k\in U[3,14]$
Infectivity period (I, $o$) [days] ${\tau }_{\text{inf}}^{I,o}$	$1∕M^o$	6.6060 (6.4495–6.7548)	$1∕k,k\in U[3,14]$
Infectivity period (A, $y$) [days] ${\tau }_{\text{inf}}^{A,y}$	$1∕M_{\mathrm{AR}}^y$	7.1888 (7.1049–7.2790)	$1∕k,k\in U[5,14]$
Infectivity period (A, $o$) [days] ${\tau }_{\text{inf}}^{A,o}$	$1∕M_{\mathrm{AR}}^o$	7.3453 (7.3183–7.3748)	$1∕k,k\in U[5,14]$
Conversion fraction (I to H, R, $y$)	${\gamma }^y$	0.2919 (0.2874–0.2965)	$c\in U[0.1,0.4]$
Conversion fraction (I to H, R, $o$)	${\gamma }^o$	0.6438 (0.6370–0.6499)	$c\in U[0.45,0.85]$
Recovery period (H to R, $y$) [days]	$1∕{\chi }^y$	7.8047 (7.6836–7.9446)	$1∕k,k\in U[5,15]$
Recovery period (H to R, $o$) [days]	$1∕{\chi }^o$	11.0976 (10.9733–11.2296)	$1∕k,k\in U[7,21]$
H to D period ($y$) [days]	$1∕{\psi }^y$	6.8216 (6.7721–6.8818)	$1∕k,k\in U[5,15]$
H to D period ($o$) [days]	$1∕{\psi }^o$	10.7787 (10.6725–10.8327)	$1∕k,k\in U[5,15]$
Initial population fraction (E, $y$)	$E^y\left(0\right)∕I^y\left(0\right)$	2.0459 (2.0077–2.1059)	$c\in U[1,3]$
Initial population fraction (E, $o$)	$E^o\left(0\right)∕I^o\left(0\right)$	1.9938 (1.9443–2.0433)	$c\in U[1,3]$
Initial population fraction (A, $y$)	$A^y\left(0\right)∕I^y\left(0\right)$	2.0481 (2.0111–2.1125)	$c\in U[1,3]$
Initial population fraction (A, $o$)	$A^o\left(0\right)∕I^o\left(0\right)$	2.0085 (1.9651–2.0669)	$c\in U[1,3]$
Initial population fraction (P, $y$)	$P^y\left(0\right)∕I^y\left(0\right)$	0.5487 (0.5316–0.5677)	$c\in U[0,1]$
Initial population fraction (P, $o$)	$P^o\left(0\right)∕I^o\left(0\right)$	0.4481 (0.4284–0.4669)	$c\in U[0,1]$

^aThe reported transmission rates $\beta$ have been multiplied by $N$. They should be normalized by the initial population when used in the ODE model.