How do you combine simple atmospheric dispersion modelling trajectories like this with environmental data and population biology knowledge to make short- and long-term predictions of desert locust swarm migration?

"Miso-scale" convection