Stochastic dynamic programming: An approach for modelling the population consequences of disturbance due to lost foraging opportunities

Since the introduction of the population consequences of disturbance (PCoD) conceptual model, lost energy has become a central component of modelling biologically meaningful disturbance. Long-term datasets on northern elephant seals of Año Nuevo, CA and bottlenose dolphins of Sarasota Bay, FL span several environmental disturbances, allowing us to quantify how behavior, physiology, and vital rates change with natural reductions in prey. While complete PCoD models are possible with large datasets, health metrics in particular can be difficult or impossible to collect for many species. In addition, PCoD models to date have assumed the worst-case scenario, in that animals exposed to a disturbance cannot compensate for lost foraging. Combined with knowledge of energetic requirements of individuals, Stochastic Dynamic Programming (SDP) models examine potential compensatory behavioral mechanisms that individuals can employ in the presence of a disturbance. However, SDP bioenergetic models require data on the energetic landscape, time-activity budgets, movement probabilities between habitats, and metabolic rates on a timescale appropriate for energy acquisition. This approach is currently being implemented for California sea lions and gray whales with the potential to ground-truth SDP models by comparing them with current elephant seal and bottlenose dolphin models.