Associate Professor University of Colorado, Boulder Boulder, Colorado, United States
Abstract: In the current era of climate change and widespread habitat degradation, the theory of evolutionary rescue has emerged to predict which or what kind of populations will be able to avoid extinction by rapid adaptation through natural selection. Despite considerable recent progress, one outstanding unknown is the effects of iteroparity and longevity on rescue, as genetic models of adaptation assume non-overlapping generations. It is often hypothesized that populations of long-lived individuals will adapt more slowly to environmental change due to slower generational turnover or more complex life cycles with stages that “shield” individuals from selection. However, models purporting to show this either make simplifying assumptions, e.g., that selection operates only once in a lifetime, or conflate the effects of life-cycle complexity with longevity per se. We address these issues by combining analytical and stochastic simulation work to model populations with no stage structure and only two vital rates, survival and reproduction, to disentangle effects of longevity and life-cycle complexity. In our model, viability selection acts on all individuals through survival in each time step. Holding maximum lifetime intrinsic fitness constant across groups, we vary survival (a proxy for longevity) to test for response to a single environmental shift on a medium-term (100 time steps) timescale, measuring population size, rates of genetic and phenotypic change, and probability of extinction.
Counter to predictions from models where selection operates only once through the life cycle, we find that with selection acting repeatedly through survival, there is little discernible effect of longevity on the rate of adaptation in a population except under a narrow range of circumstances. With little variation in the rate of phenotypic and genotypic change among populations with varying longevities, differences in maximum intrinsic growth rates and temporal variance in growth rates have a more prominent role in mean population size and extinction rates over time. Differences in the rate of adaptation only emerge at low heritability, in which case longevity is associated with a transient period of rapid phenotypic change due; for perfectly heritable traits, differences in rates of adaptation are minimal. We contrast this with a modified version of our model where selection acts only on viability of newborns and not on adults, providing similar results to those typically hypothesized. We close that several of these results are consistent with demographic and life history theory and note the potential for integration of these fields with evolutionary rescue.
