Co-occupancy
We also found differences in predictors of co-occupancy between seasons for both lynx-wolf and wildcat-wolf co-occupancies. In winter, lynx-wolf and wildcat-wolf co-occupancy were predicted by forest cover (Fig 3 B, C), but in autumn, co-occupancy for both pairs were predicted by terrain ruggedness (Fig 3 E, F). Lynx-wildcat co-occupancy was predicted by terrain ruggedness for both winter and autumn seasons and was positively associated with terrain ruggedness in both winter and autumn (Fig 3 A, D), but in autumn the relationship was less linear (Fig 3D). In contrast, both lynx-wolf and wildcat-wolf co-occupancy were negatively associated with terrain ruggedness in autumn (Fig 3 E, F). In winter, wildcat-wolf co-occupancy was negatively associated with forest cover while lynx-wolf co-occupancy was positively associated with forest cover, but only at >75% forest cover (Fig 3 E, F).