Detectability of landscape effects on recolonization increases with regional population density.

Variation in population size over time can influence our ability to identify landscape-moderated differences in community assembly. To date, however, most studies at the landscape scale only cover snapshots in time, thereby overlooking the temporal dynamics of populations and communities. In this paper, we present data that illustrate how temporal variation in population density at a regional scale can influence landscape-moderated variation in recolonization and population buildup in disturbed habitat patches. Four common insect species, two omnivores and two herbivores, were monitored over 8 years in 10 willow short-rotation coppice bio-energy stands with a four-year disturbance regime (coppice cycle). The population densities in these regularly disturbed stands were compared to densities in 17 undisturbed natural Salix cinerea (grey willow) stands in the same region. A time series approach was used, utilizing the natural variation between years to statistically model recolonization as a function of landscape composition under two different levels of regional density. Landscape composition, i.e. relative amount of forest vs. open agricultural habitats, largely determined the density of re-colonizing populations following willow coppicing in three of the four species. However, the impact of landscape composition was not detectable in years with low regional density. Our results illustrate that landscape-moderated recolonization can change over time and that considering the temporal dynamics of populations may be crucial when designing and evaluating studies at landscape level.