Fig. 5.

Schematic description of our model. We model individual-level transmission of SARS-CoV-2, in continuous time and space and over a heterogeneous landscape with varying population density over 100 m $\times$ 100 m grids. Disease status of an individual is assumed to follow the susceptible–exposed–infectious–recovered framework. Infectiousness of an infectious individual $\beta \left(t\right)$ is time dependent and decreases due to social distancing. Likelihood of transmission from the infectious individual to a susceptible individual, at distance $r$ and angle $\theta$ measured from the infectious source, is determined by 1) a spatial movement kernel (density) function $f$ with mean $\eta$, 2) change of the mean movement distance due to a shelter-in-place order (informed by the aggregate mobility data), and 3) spatial distribution of population denoted by $\mathit{ŝ}$ (i.e., detailed grid-level population density shown in the figure), which, all together, could more realistically account for heterogeneous mixing of individuals in the population (9).