Abstract
Metapopulation models describing cities with different populations coupled by the travel of individuals are of great importance in the understanding of disease spread on a large scale. An important example is the Rvachev–Longini model which is widely used in computational epidemiology. Few analytical results are, however, available and, in particular, little is known about paths followed by epidemics and disease arrival times. We study the arrival time of a disease in a city as a function of the starting seed of the epidemics. We propose an analytical ansatz, test it in the case of a spread on the worldwide air-transportation network, and show that it predicts accurately the arrival order of a disease in worldwide cities.