Agent-based models can predict behavioural and physiological responses of top predators to environmental change: A case study with Weddell seals (Leptonychotes weddellii)

Background/Question/Methods

One of the crucial scientific challenges of this century is characterizing the vulnerability of ecosystems to global climate change. Dynamic bioenergetics models can provide a theoretical construct for addressing specific physiological and ecological questions; however, many models fail to link energy deficiencies with reproductive consequences, and thus cannot be used to predict population-level consequences of extrinsic stressors (e.g. climate change, fishing pressure, sea ice reductions). Further, non-dynamic models cannot be used to address the behavioural responses that may help mammals maximize their fitness under predicted climate scenarios. Here, we present an agent-based, ecophysiological model that simulates the energy balance of adult, female Weddell seals. The input parameters are taken from the literature and include both physiological parameters as well as population-wide ranges for the typical duration and phenology of seasonal life history events. Energy intake depends both on foraging effort and stochastic prey availability at each timestep, whereas energy output is calculated from time- and behaviour-specific energy demands. At each timestep, the simulated animal selects an activity (e.g. forage, nurse pup, molt, rest), based on its body condition and extrinsic conditions (e.g. air temperature). At the end of each timestep, the energy budget is balanced, catabolism or anabolism occurs, and the updated body composition is used to assess the ability of females to continue supporting their fetus/pup. Following model development and validation with empirical data, we ran simulations and studied the responses of individuals to: (1) baseline conditions; (2) reduced prey availability; and (3) variance in ambient air temperature.

Results/Conclusions

As expected, our model replicated the large fluctuations in total energetic requirements that result from seasonal life history events (e.g. reproduction and molt). A 5% reduction in prey consumption rates resulted in seals spending more time foraging and less time resting (from 52.2%±6.2% resting to 40.3±8.4% resting). At the end of the year-long simulations, animals in the baseline simulation were in significantly better condition than animals in the perturbation simulation (T-test, t₂₈=5.6, p<.0001). Our model explores decision-based energy allocation strategies that occur under different energetic stressors and elucidates how extrinsic conditions may impact individual fitness. Identifying the temporal sensitivities of Weddell seals to predicted anthropogenic changes is a valuable contribution to the study of global change biology, and can be used to inform management decisions in polar regions.