‘Supersynchrony’ and the Moran effect in spatio-temporally variable environments

Background/Question/Methods

Spatial synchrony in population dynamics is a major determinant of important ecological processes, including regional extinction risk. It is well known that spatially synchronized environmental fluctuations can drive population synchrony (the Moran effect), which increases the probability of extinction by causing multiple populations to experience low abundance simultaneously. However, recent analyses indicate that stochastic environmental factors such as temperature and precipitation exhibit significant spatial gradients in temporal autocorrelation structure (‘noise-color’), which intrinsically limits spatial environmental synchrony. We used theoretical two-patch metapopulations to examine how such spatial environmental variation affects population synchrony, both with and without dispersal. In our model, environmental fluctuations in each patch ranged from uncorrelated (‘white-noise’) to highly autocorrelated (‘reddened-noise’) and directly affected the demographic rates of inhabiting populations. We tested a wide range of spatial gradients in environmental fluctuations and dispersal pressures using simulation analyses and calculated population and environmental synchrony between paired patches.

Results/Conclusions

Increasing the difference between patches in the structure of environmental fluctuations systematically reduces spatial population synchrony. This desynchronizing effect becomes more pronounced as dispersal pressure approaches zero (because dispersal itself is a strong synchronizing factor). However, the rate at which synchrony deteriorates with increasing environmental gradients depends on how populations track environmental fluctuations. Surprisingly, undercompensatory dynamics can cause population synchrony to exceed environmental synchrony, even in the absence of dispersal (a phenomenon we call ‘supersynchrony’). This happens because some of the high frequency components of environmental variation are inaccessible by undercompensatory populations, causing populations in different environments to be intrinsically ‘reddened,’ and consequently more dynamically similar than their respective environments. The Moran Effect may therefore be stronger than expected when environmental fluctuations vary across space. This consideration has the potential to significantly alter estimates of regional extinction risk via environmental forcing, which is especially important given the ubiquity of temporally autocorrelated environmental factors and predicted changes in autocorrelation structures under future changes in global climates.