Dispersal in Patchy Vole Populations: Role of Patch Configuration, Density Dependence, and Demography

We studied dispersal movements in 12 enclosed, patchy populations of root voles (Microtus oeconomus) during the breeding season. The habitat was manipulated experimentally so that there were four replicates of three types of habitat patch configuration: two large patches, six small isolated patches, and six small patches of which two patch triplets were connected by corridors. The total habitat area (1350 m2) and the distance to nearest neighboring patch (15 m) were the same in the three different configurations. The matrix area between the patches was kept open and uninhabitable for voles by weekly mowing. Dispersal was defined as shifts between patches. Movements across the open matrix habitat to traps located along the edge of the enclosures were also recorded. The frequency of shifts between patches was higher among small than among large patches. Corridors channeled dispersal between corridor-connected patches but did not enhance the frequency of shifts between patches at the population level. Dispersal was strongly density dependent, and most so for subadult animals. High-density patches had low emigration rates. Root voles immigrated onto patches with a smaller number of individuals, especially of their own sex and reproductive state, than were present in the patch they left. Thus most shifts between patches took place from patches with relatively low density to patches with even lower density. Small patches had higher spatiotemporal variability in density and demographic composition than large patches, and this probably caused the higher dispersal rate among small patches. Emigration and immigration contributed most to the demographic turnover in small patches. In particular, emigration was the main demographic parameter behind declining numbers and patch extinction in small patches with few individuals. Our study highlights the importance of taking patch-specific conditions such as patch size, demographic composition, and spatiotemporal demographic variability into account when studying dispersal rates. Moreover, the kind of density-dependent emigration–immigration dynamics found in our study does not match the common perception that dispersal works primarily to reduce extinction probabilities through rescue effects. In particular, the impact of emigration as a factor that may increase the extinction probability of small, isolated patches with few individuals needs to be considered in future studies.