When habitat is lost matters: patterns of population decline and time to extinction in a seasonal, density-dependent model

Nearly all wild populations live in seasonal environments in which they experience regular fluctuations in environmental conditions that drive population dynamics. Recent empirical evidence from experimental populations of fruit flies suggests that demographic signals inherent in the counts of seasonal populations, including reproduction and survival, can indicate when in the annual cycle habitat loss occurred. However, it remains unclear whether these signatures of season-specific decline are detectable under a wider range conditions. Here, we use a bi-seasonal Ricker model previously developed and applied to the same experimental system to examine season-specific signals of population decline induced by different rates of habitat loss in the breeding or non-breeding season and different strengths of density dependence. Consistent with the findings in Drosophila, breeding habitat loss was accompanied by reduced reproductive output and a density-dependent increase in survival during the subsequent non-breeding period. Non-breeding habitat loss resulted in reduced non-breeding survival and a density-dependent increase in reproduction in the following breeding season. These season-specific demographic signals of decline were present under a wide range of habitat loss rates (2-25% per generation) and different density-dependent regimes (weak, moderate, and strong). We show that stronger density dependence can negatively influence time to extinction when non-breeding habitat is lost, whereas the strength of density dependence does not influence time to extinction with breeding habitat loss (although, in all cases, density dependence itself was an important modulator of population dynamics). Our results illustrate the need to incorporate seasonality in theoretical models to better understand when populations are being driven to decline.