We used a one-proton radical-pair model with isotropic hyperfine coupling (a = 0.5 mT), anisotropy (α = 0.3), and a lifetime of 20 μs to simulate the triplet yield in the presence of a 46 μT static magnetic field. We calculated the change in triplet yield caused by an additional 1.3 MHz oscillating magnetic field in resonance with the static field. We also compared this with the change in triplet yield from a 12 μT decrease in the static field. The intensity of the oscillating field required to match the change in triplet yield from the static field decrease was 0.033 μT, which is less than any intensity used in our experiments. The study discusses modeling disease outbreaks in realistic urban social networks using dynamic bipartite graphs generated from individual-based urban traffic simulations. The contact network among people is a strongly connected small-world graph with a well-defined degree distribution, while the locations graph is scale-free, allowing efficient outbreak detection by placing sensors in hubs. The simulation framework analyzed mitigation strategies for smallpox spread, suggesting that targeted vaccination combined with early detection can contain outbreaks without mass vaccination. The dense social-contact networks in urban areas facilitate rapid disease spread. Current urbanization trends increase the risk of disease outbreaks. Recent studies suggest that mass vaccination may not be necessary for smallpox control. A highly resolved agent-based simulation tool, EpiSims, was developed to model disease spread using realistic population mobility data and parameterized disease progression models. The simulation generated a dynamic contact graph replacing differential equations. EpiSims was based on the Transportation Analysis and Simulation System (TRANSIMS), which estimated social networks based on transportation infrastructure constraints. The study presents a model of Portland, Oregon, showing that people move between locations, exposing themselves to infectious agents and transporting them. The social contact network was represented as a bipartite graph, G_PL, with people and locations as vertices. The degree distribution of people vertices showed a sharp peak near the average value, followed by an exponential decay. The degree distribution of location vertices showed a power-law tail with an exponent of about -2.8. The static projections of the bipartite graph, G_P and G_L, were analyzed, showing that G_L is a scale-free network with an exponent of -2.8. The study found that the people-contact graph is more like a small-world graph than a random graph, with a clustering coefficient much higher than that of an Erdős–Rényi random graph. The expansion property of the graph indicated that disease is likely to spread quickly if not controlled early. The results suggested that targeted vaccination combined with fast detection can contain outbreaks. The study also introduced the overlap ratio, a non-local property of the graph crucial for early detection. The overlap ratio was used to determine the minimum dominating set of locations needed to cover the population. The studyWe used a one-proton radical-pair model with isotropic hyperfine coupling (a = 0.5 mT), anisotropy (α = 0.3), and a lifetime of 20 μs to simulate the triplet yield in the presence of a 46 μT static magnetic field. We calculated the change in triplet yield caused by an additional 1.3 MHz oscillating magnetic field in resonance with the static field. We also compared this with the change in triplet yield from a 12 μT decrease in the static field. The intensity of the oscillating field required to match the change in triplet yield from the static field decrease was 0.033 μT, which is less than any intensity used in our experiments. The study discusses modeling disease outbreaks in realistic urban social networks using dynamic bipartite graphs generated from individual-based urban traffic simulations. The contact network among people is a strongly connected small-world graph with a well-defined degree distribution, while the locations graph is scale-free, allowing efficient outbreak detection by placing sensors in hubs. The simulation framework analyzed mitigation strategies for smallpox spread, suggesting that targeted vaccination combined with early detection can contain outbreaks without mass vaccination. The dense social-contact networks in urban areas facilitate rapid disease spread. Current urbanization trends increase the risk of disease outbreaks. Recent studies suggest that mass vaccination may not be necessary for smallpox control. A highly resolved agent-based simulation tool, EpiSims, was developed to model disease spread using realistic population mobility data and parameterized disease progression models. The simulation generated a dynamic contact graph replacing differential equations. EpiSims was based on the Transportation Analysis and Simulation System (TRANSIMS), which estimated social networks based on transportation infrastructure constraints. The study presents a model of Portland, Oregon, showing that people move between locations, exposing themselves to infectious agents and transporting them. The social contact network was represented as a bipartite graph, G_PL, with people and locations as vertices. The degree distribution of people vertices showed a sharp peak near the average value, followed by an exponential decay. The degree distribution of location vertices showed a power-law tail with an exponent of about -2.8. The static projections of the bipartite graph, G_P and G_L, were analyzed, showing that G_L is a scale-free network with an exponent of -2.8. The study found that the people-contact graph is more like a small-world graph than a random graph, with a clustering coefficient much higher than that of an Erdős–Rényi random graph. The expansion property of the graph indicated that disease is likely to spread quickly if not controlled early. The results suggested that targeted vaccination combined with fast detection can contain outbreaks. The study also introduced the overlap ratio, a non-local property of the graph crucial for early detection. The overlap ratio was used to determine the minimum dominating set of locations needed to cover the population. The study