Models of optimal avian migration: state, time and predation

Taking as my starting point the paper by Alerstam and Lindstrom (1990), I review models of optimal foraging and fuel loads during avian migration. As well as summarising previous results, I highlight topics where further work is required. I argue that we now have a good understanding of simple deterministic models based on time and energy. Discrepancies between such models and data have proved instructive. We also have some results about the influence of predation risk. Some work has been done on state-dependent models. This approach needs to be extended to models that can take account of changes in various aspects of internal state, and that put migration into the context of a bird's annual routine. I suggest that a combination of simple analytic models of some aspects of state, together with dynamic programming models of annual migration routines may be a fruitful research plan. We also need to explore the consequences of birds making decisions on the basis of climatic conditions.