Animals use and select habitat at multiple hierarchical levels and at
different spatial scales within each level. Still, there is little
knowledge on the scale effects at different spatial levels of species
occupancy patterns. The objective of this study was to examine nonlinear
effects and optimal-scale landscape characteristics that affect
occupancy of the Siberian flying squirrel, Pteromys volans, in South- and Mid-Finland. We used presence–absence data (n = 10,032
plots of 9 ha) and novel approach to separate the effects on site-,
landscape-, and regional-level occupancy patterns. Our main results
were: landscape variables predicted the placement of population patches
at least twice as well as they predicted the occupancy of particular
sites; the clear optimal value of preferred habitat cover for species
landscape-level abundance is a surprisingly low value (10% within a 4 km
buffer); landscape metrics exert different effects on species occupancy
and abundance in high versus low population density regions of our
study area. We conclude that knowledge of regional variation in
landscape utilization will be essential for successful conservation of
the species. The results also support the view that large-scale
landscape variables have high predictive power in explaining species
abundance. Our study demonstrates the complex response of species
occurrence at different levels of population configuration on landscape
structure. The study also highlights the need for data in large spatial
scale to increase the precision of biodiversity mapping and prediction
of future trends.