Description:
1. Habitat fragmentation has major implications for demography and genetic structure of natural plant and animal populations as small and isolated populations are more prone to extinction. Therefore, many recent studies focus on spatial fragmentation. 2. However, the temporal configuration of suitable habitat may also influence dispersal and gene flow in fragmented landscapes. We hypothesize that short-term switching of inhospitable matrix areas into suitable habitat can mitigate effects of spatial fragmentation in natural and seminatural ecosystems. 3. To test our hypothesis, we investigated the hairy-footed gerbil (Gerbillurus paeba, Smith 1836), a ground-dwelling rodent, in fragmented Kalahari savannah areas. Here, rare events of high above mean annual rainfall suggest short-term matrix suitability. 4. During the field survey in 'matrix' areas in the Kalahari (shrub encroachment by heavy grazing) we never observed the hairy-footed gerbil in years of average rainfall, but observed mass occurrences of this species during rare events of exceptionally high rainfall. 5. In a second step, we developed an agent-based model simulating subpopulations in two neighbouring habitats and the separating matrix. Our mechanistic model reproduces the mass occurrences as observed in the field and thus suggests the possibly underlying processes. In particular, the temporary improvement in matrix quality allows reproduction in the matrix, thereby causing a substantial increase in population size. 6. The model demonstrates further how the environmental trigger (rainfall) impacts genetic connectivity of two separated subpopulations. We identified seasonality as a driver of fragmentation but stochasticity leading to higher connectivity. 7. We found that our concept of temporal fragmentation can be applied to numerous other fragmented populations in various ecological systems and provide examples from recent literature. We conclude that temporal aspects of fragmentation must be considered in both ecological research and conservation management.