Andrew Gonzalez and David Matthews. McGill University
Environmental change is creating heterogeneous landscapes with increasing proportions of habitat of low quality and low conservation value. A consequence of dispersal in such landscapes is that individuals may move into habitats of low quality, creating sink populations. By definition, the long-term intrinsic growth rate of a population sink is negative and in the absence of dispersal from a source population extinction is assured. However, there is growing empirical evidence that temporal variation in habitat quality may cause transient periods of positive growth that may lead to long-term sink persistence. We test recent theory that suggests, surprisingly, that stochastic environmental variability may have an inflationary effect that enhances the long-term abundance and persistence of metapopulations made up entirely of sink populations. Using experimental populations of Paramecium aurelia we show that it is possible for a sink metapopulation to persist for many generations in a stochastic environment. In accordance with the theory, we show that positive temporal autocorrelation, and low spatial correlation in the environment, can ensure the long-term persistence, and enhance the mean and maximum abundance of sink metapopulations. We show that metapopulation persistence was due to positive spatial growth-density covariance that resulted in arithmetic mean growth rates across space greater than zero. High levels of spatial correlation in the environment created strong population synchrony (a Moran effect) and greatly reduced the persistence time of the sink metapopulations. These results have important implications for the development of a theory underlying the synergistic effects of habitat fragmentation and environmental change on population persistence.