Incorporating physiology into mechanistic species’ distribution models for ecological forecasting

Background/Question/Methods

Can including physiology in species’ distribution models improve predictions of responses to climate change? We describe several approaches to include physiology in both correlative and mechanistic range models for lizard and butterfly species.  First, we extend biophysical models to include demography.  This links species’ traits to population dynamics and ultimately species’ ranges and enables examining the implications of geographic trait variation and trait evolution.  Second, we incorporate the output of mechanistic models into correlative models.  This enables using species-specific physiological constraints (e.g., development time) to inform environmental constraints on distributions.

Results/Conclusions

We find that including physiology can improve predictions of species’ responses to climate change.  While correlative and mechanistic distribution models often perform similarly at predicting current distributions, including physiology and other aspects of a species’ biology may be essential to predicting distribution shifts.  Additionally, incorporating geographic trait variation and trait evolution can be important for accurate ecological forecasting.  Incorporating the output of mechanistic models into correlative models may provide a feasible means to incorporate species-specific physiological constraints into correlative models.