Mobility-informed metapopulation models predict the spatio-temporal spread of respiratory epidemics across scales.

Understanding the spatiotemporal dynamics of infectious disease spread is critical for anticipating epidemic trajectories and guiding public health responses. Accurate forecasts of where and when outbreaks are likely to emerge can support efficient resource allocation, particularly during the early stages of epidemics when surveillance data are limited. In this study, we used empirical human mobility data derived from county-level commuting and air traffic flows, and a theoretical mobility model (the radiation model) to study the relative order of epidemic onset across spatial scales. These mobility models were incorporated into a metapopulation framework to predict the spread of three major respiratory pathogens: COVID-19, seasonal influenza, and respiratory syncytial virus (RSV). We applied this framework to county-level transmission within South Carolina and state-level introductions across the United States. In both empirical and theoretical mobility scenarios, we found that effective distance, a network-based measure of mobility-informed proximity, reliably predicts the relative timing of epidemic onset. These results demonstrate that mobility-informed metapopulation models can capture consistent spatiotemporal patterns across disease systems and spatial scales, even in the absence of detailed epidemiological parameters. This highlights their potential as scalable, data-efficient tools for outbreak forecasting and early public health planning.