Demographic heterogeneity affects population dynamics through multiple pathways

Background/Question/Methods

Demographic heterogeneity -- variation (other than that due to age, size, or stage) in the propensities of individuals within a population to survive, reproduce, and grow -- has attracted much attention in recent years. Empirical studies have documented substantial heterogeneity in a number of populations. Theoretical studies have shown that it can have substantial impact on demographic variance (the variance in per-capita vital rates due to demographic stochasticity); consequent effects on extinction risk have been suggested, but not analyzed. Furthermore, the correlation structure of the heterogeneity, both within and between individuals, and within and between times; only a few of these correlations (such as "persistent heterogeneity" -- perfect within-individual autocorrelation -- in which an individual carries its demographic trait for life) have been analyzed

Here we focus on extinction risk, using slightly supercritcal branching process models to investigate how the direct effects of heterogeneity on the demographic variance is modified by correlations within and among individuals. Our primary measure is the "extinction effective population size" (N_e), the number of individuals in a heterogeneous population that would have the same extinction risk as a homogeneous population of size 100.

Results/Conclusions

When temporal correlations are absent, heterogeneity's effect on N_e is exactly proportional to its effect on the demographic variance. In contrast, temporal correlations create the opportunity for selection, including natural selection (correlations among individuals) and cohort selection (correlations within individuals). Selection causes survival heterogeneity to increase N_e, above and beyond the direct effects on the demographic variance, by modifying the phenotype structure of the population. Reproductive heterogeneity does not allow cohort selection; natural selection on reproduction has a similar effect as on survival. However, within-individual correlations interact with nonlinear averaging to create transient reverse selection, in which the population becomes dominated by low-fitness individuals. This leads to reductions in N_e, relative to the direct effects on the demographic variance.

We conclude by presenting a conceptual model that integrates heterogeneity's direct effects on demographic variance, indirect effects on the asymptotic demographic mean parameters via selection, and transient effects on population structure. This will help us understand the effects of heterogeneity on population dynamics in a variety of models and ecological contexts.