We argue that identifying such bottlenecks will be a fundamental step in future modelling of weather-driven epidemics.