The limits of demographic buffering in coping with environmental variation

Animal populations have developed multiple strategies to deal with environmental change. Among them, the demographic buffering strategy consists in constraining the temporal variation of the vital rate(s) that most affect(s) the overall performance of the population. Tortoises are known to buffer their temporal variation in adult survival, which typically has the highest contribution to the population growth rate λ, at the expense of a high variability on reproductive rates, which contribute far less to λ. To identify the effects of projected increases in droughts in its natural habitat, we use field data collected across 15 locations of Testudo graeca in southeast Spain over a decade. We analyse the effects of environmental variables on reproduction rates. In addition, we couple the demographic and environmental data to parameterise an integral projection model to simulate the effects of different scenarios of drought recurrence on λ under different degrees of intensity in the survival–reproduction tradeoff. We find that droughts negatively affect the probability of laying eggs; however, the overall effects on λ under the current drought recurrence (one/decade) are negligible when survival is constant (independent of the reduction of reproduction by drought events) and when survival increased as a tradeoff with the reduction of reproduction rates, with a threshold to population viability at three or more droughts/decade. Additionally, we show that, although some species may buffer current environmental regimes by carefully orchestrating how their vital rates vary through time, a demographic buffering strategy is insufficient to ensure population viability in extreme regimes. Our findings support the hypothesis that the demographic buffering strategy has a limit of effectiveness when adverse conditions occur frequently. Our methodological approach provides a framework for ecologists to determine how effective the management of environmental drivers can be for demographically buffering populations, and which scenarios may not provide long‐term population persistence.