Fig 1.
Conceptual figure illustrating the rationale behind the approach described in this study.
The example refers to the identification of a wildlife corridor connecting two core areas, i.e., a hypothetical elk winter range and a new one reached after a dispersal event. Typical connectivity models (a) would depict the least cost corridor connecting the two areas, which is the most likely path given a friction map on the background. Our approach highlights the importance to implement behavioural ecology (in this case, dispersal ecology) into connectivity modelling science. A young male elk usually migrates during late spring—early summer with the mother’s group, moving from the natal winter range to the early summer range (b). During summer, the young elk may disperse to a new suitable summer home range (c), and, in autumn, eventually move to the new winter range (d). If the animal will adopt the migratory strategy, then it will periodically migrate between the new winter range and the new summer range (d). Implementing dispersal ecology into connectivity modelling means combining a sequence of migratory and dispersal movements (e, resulting from b + c + d), which suggests a very different potential wildlife corridor (e) compared to the one predicted by simply connecting the two winter ranges (a).
Fig 2.
Relative probability of selection estimated by Resource Selection Functions (RSFs) in winter and summer for a) elevation, b) terrain ruggedness, c) normalized difference vegetation index NDVI and d) canopy cover.
Fig 3.
Relative probability of selection estimated by Step Selection Functions (SSFs) during spring and autumn movements for a) terrain ruggedness, b) canopy cover, c) distance to highways and d) distance to minor roads.
Fig 4.
Corridor map of spring movements connecting winter ranges with summer ones.
a) Normalized least-cost-corridors, with low values as optimal, connecting winter with summer core areas; b) close up section along highway 3; c) close up friction map produced from spring SSFs. Note that elk are predicted to move parallel or perpendicular to highway 3 (map b), in contrast to what is depicted in S5 Fig, where elk are predicted to move along highway 3 at the bottom of the valley, which would be the movement if there were no highway 3.
Fig 5.
Corridor map of autumn movements connecting summer ranges with winter ones.
a) Normalized least-cost-corridors, with low values as optimal, connecting summer with winter core areas; b) close up section along highway 3; c) close up friction map produced from autumn SSFs. Note that elk are predicted to move parallel or perpendicular to highway 3 (map b), in contrast to what depicted in S6 Fig, where elk are predicted to move along highway 3 at the bottom of the valley, which would be the movement if there were no highway 3.
Fig 6.
Corridor-road intersections of a) 168.8 km in spring considering SSFs predictions computed with actual distance to roads, b) 355.5 km in spring assuming as there were no roads (i.e., distance to roads set to maximum value when predicting SSFs), c) 172.9 km in autumn considering roads, and d) 379.5 km in autumn assuming as there were no roads. Maps b) and d) depict road segments that would be crossed by elk if there were no roads. About half of these segments are predicted not to be crossed by elk (maps a, c) as a result of road avoidance by elk.