Résumé
The emission of artificial light at night (ALAN) generates a light pollution. The impacts on fauna, flora and ecosystems have been increasingly studied in recent decades. However, mammals - except bats or rodents - remain under-studied, particularly in terms of space use. Here, we implemented a three-year in-situ before–during–after exposure protocol to assess the effect of artificial light at night at five underpasses of a motorway in a French regional natural park. Using camera traps, we recorded movements of medium-sized wild mammals and collected data on 12 species, especially European badger Meles meles, red fox Vulpes vulpes and martens (Martes martes and M. foina). Our results showed that lighting significantly decreased the probability to cross the underpasses for European badger in spring and autumn and for red fox in spring, while there was no significant effect of lighting for martens. Lighting also reduced crossing speed for badgers. We can conclude that, for some medium-sized wild mammals, ALAN triggers an avoidance behaviour that prevents them from crossing lit underpasses during certain seasons. This suggests that ALAN can act as a nightscape fragmentation, which is in line with previous studies on other taxa - as bats, insects, amphibians or eels. This additional barrier effect confirms the value of dark infrastructure; i.e. ecological network policies to preserve dark habitat patches and dark corridors.