Global effects of climate change on animal population dynamics

Abstract

Climate change affects many statistical descriptions of the environment. The impact of changes in mean environmental conditions on animal populations has been widely reported. The variability and autocorrelation of environmental variables are also changing over time, yet less research has focussed on what impact these changes may have on populations. Is the research focus on changes in mean conditions justified? How do changes in different statistical descriptions of climate change affect populations, and how do the impacts compare? To answer these questions, we developed a simple stochastic population model, explicitly linked to the environment, and compared the impacts of changes in environmental mean and variability. We found, using both the long-term stochastic growth rate and extinction risk as proxies for population fitness, that changes in variability have a significant impact on population dynamics. The main gradient along which the relative importance of changes in environmental mean and variability varied was the population's distance from its ideal environment. We also re-analysed existing population models to yield the sensitivity of the population to changes in environmental mean and variability. Results support the findings from our model, and confirm the importance of changes in variability for population dynamics. Previous theoretical and laboratory studies concluded that the autocorrelation in the environment in part affects the autocorrelation in population time series. So far, this hypothesis has not been tested using empirical data. We used a database of population time series to find that the autocorrelation in mean summer temperature is significantly correlated with the autocorrelation in population time series. Results also show that environmental variables have become less autocorrelated in most geographical regions, suggesting that populations' autocorrelation may also be changing. Autocorrelation in population time series has been linked to extinction; these results may therefore have important implications for animal populations.