On natural selection regulation in the population dynamics of birds and mammals

Lars Witting

Abstract: Empirical research is increasingly documenting eco-evolutionary dynamics that shape ecological processes. I examine the population dynamic implications of this, analysing whether natural selection improves our ability to predict population dynamic trajectories. Fitting single-species population dynamic models to 3,368 and 480 timeseries for 900 species of birds and 208 mammals, I find that selection-based population dynamic models are 320 (se:1.3) times more probable on average than models with no selection. Selection is essential in 76% to 90% of AIC-selected models, explaining 80% of the population dynamics variance, with median selection regulation being 1.5 (se:1.1) times stronger than density regulation. The estimated dynamics is cyclic with median damping ratios for birds and mammals of 0.12 (se:0.0068) and 0.083 (se:0.022), and population periods of 8 (se:0.56) and 6.1 (se:1.1) generations, given stable cycles with damping ratios around zero. These results highlight the necessity of integrating natural selection into population dynamic theory, and they are discussed in relation to the literature resolving several enigmas of population dynamic cycle.