Variation in population dynamics across a species range - good bet hedging populations have highest average abundance

Background/Question/Methods

Abundance of a species is known to vary by orders of magnitude across a species range. And, although the patterns are not as simple as the "abundant center" hypothesis, there is clear spatial structure of this variation. Correlational approaches examining how abundance covaries with climate are central to the niche modelling approach. However, developing a more mechanistic theory based on population dynamics is an oft-stated goal that should lead to improved ability to extrapolate and predict species ranges into novel conditions.

Using the North American Breeding Bird survey, I fit a stochastic Ricker population dynamic model to each route and explore how the population dynamic parameters covary with each other, with abundance, and across space. A number of robust altenratives to the Ricker model and some metrics of population time-series are also explored.

Results/Conclusions

Suprirsingly, abundance is negatively correlated with intrinsic rate of population increase (r or λ). Abundance is related to the strength of density dependence in the intuitive way. But by far the strongest signal is the relationship to stochasticity. Using a principle component analysis of parameters in different population dynamic models and descriptors of population dynamics shows that the first component explaining the most variance is related to bet hedging. Sites where a species is most abundant are not those where the population grows the fastest but where the year-to-year variation is minimized. This PCA axis is also a good predictor of abundance across the range. Implications of this result for niche modelling are discussed.