Figure 1

The spatiotemporal distribution of predicted infected population (in natural logarithm scale) across all states under different simulation scenarios: (A) \(\alpha _r = 1\) and \(\alpha _b = 1\), i.e., all parameters took the values of the initial configuration, obtained through data assimilation method using the numbers of confirmed cases during March 1 – March 20, 2020; (B) the travel flow was reduced to \(\alpha _t = 0.05\), while other parameters values remained unchanged; (C) \(\alpha _r = 0.1\) and \(\alpha _b=1\); (D) \(\alpha _r = 1\) and \(\alpha _b=0.1\); (E) \(\alpha _r = 0.1\), \(\alpha _b=0.1\). In the simulations, the transmission rate was set to be \(b = \alpha _b b_0\) and the reporting rate was set to be \(r = 1-\alpha _r(1-r_0)\). Where \(r_0\) and \(b_0\) were the reporting rate and the transmission rate on March 20, 2020, which are inferred from the data assimilation step (Note: The maps are created using Esri’s ArcGIS 10.7 software).