Dispersal between two populations can synchronize their temporal dynamics, resulting in a phenomenon known as spatial synchrony. For populations that undergo cycling, such as those exhibiting predator-prey fluctuations, the amount of dispersal required to synchronize their dynamics is thought to be quite low, on the order of 1% per generation. If this untested view is correct, it implies that low levels of local dispersal could induce synchrony across entire metapopulations of cycling systems. Given the importance of synchrony on such factors as metapopulation stability, we were interested in answering the following question: How much dispersal is necessary to induce and maintain synchrony between cycling populations?
We addressed the question empirically, in a model system consisting of the protist predator Euplotes patella and its prey Tetrahymena pyriformis. For each of 11 different dispersal rates, ranging from 0% to 12.5% per day, we established 5 pairs of initially asynchronous predator-prey populations. Over the course of 90 days, we monitored how quickly the population pairs from each dispersal treatment became synchronous and measured the degree to which synchrony was maintained for the remaining period.
Results/Conclusions
Dispersal as low as 0.125% per day, approximately 0.04% per prey generation, resulted in some degree of synchronization between prey population pairs. Time to synchrony varied with dispersal rate, with greater dispersal corresponding to faster synchronization. Mean synchrony over the duration of the experiment increased with dispersal rate in a nonlinear asymptotic fashion, broadly as predicted by theory. However, even under relatively high dispersal rates, populations did not necessarily remain in phase for the duration of the experiment. Based on a parameterized model of our system, we attribute such phase drifting to demographic stochasticity. Thus, while low levels of dispersal can apparently trigger synchrony between populations, dispersal alone cannot maintain synchrony due to the desynchronizing effects of demographic stochasticity. We suggest, therefore, that the amount of demographic stochasticity within populations may impose a hard limit on the degree to which they can be synchronized.