Exploring factors affecting the reintroduction and amplification of West Nile virus in heterogeneous landscapes in Canada, using a cellular automata approach

West-Nile virus (WNV) is an endemic public health risk in Canada, with outbreaks/reintroduction and amplification that may increase in frequency and size with climate change and urbanization. In this modeling study, we used a compartmentalized and spatialized Susceptible, Exposed, Infected, Recovered (SEIR) WNV transmission model incorporating a cellular automata approach. We tested four scenarios in which we modified the number of infected birds arriving in spring, modified the number of infected mosquitoes emerging from their overwintering/dormancy period, studied the impact of bird abundance on epidemic starting point locations, and examined the progressive shift in mosquito feeding preferences from birds to mammals. First, we observed that WNV amplification may be associated with the arrival of infected migratory birds in the spring, with more severe epidemics as the number of infected birds increases. Secondly, amplification due to the local persistence of WNV virus in surviving infected overwintering female mosquitoes resulted in more severe epidemics in the human population than when amplification was due to the arrival of infected birds. Thirdly, epidemics were more severe when initiated in cells with low bird density than in those with high density. Lastly, the shift in mosquito feeding preference to human blood meals at the end of summer could generate more cases in human populations if reservoir birds delay their migration and stay longer, amplifying the virus locally. A field study is needed to quantify the impact of these mechanisms on WNV reintroduction in southeastern Canada, to better design interventions and early warning systems.