Résumé
Some protected species have benefited from human activities to a point where they sometimes raise concerns. However, gaps in knowledge about their human-related behaviour hamper effective management decisions. We studied non-breeding common ravens Corvus corax that aggregated and predated livestock in the surroundings of a landfill. Combining several approaches, we first assessed the effectiveness of relocations at distances ranging from 20 to 240km, and of one-time disturbance consisting in non-lethal shots performed at night roosts and at surrounding foraging areas during two consecutive evenings, in reducing ravens' presence at the landfill. Then, we documented the spatial ecology of both relocated ravens and control ones (i.e. released in situ). Although the return probability widely varied with relocation distance and time after release, homing behaviour (87/102 relocated ravens marked with wing tags [85.3%] returned to the landfill in 3years) prevented relocations from sustainably reducing the presence of ravens. Likewise, one-time disturbance only decreased ravens' abundance during a few hours. These results could be related to the extensive movements of ravens equipped with GPS trackers. The total area occupied by control ravens reached 40,492km(2), i.e. 7.4% of the area of France and 21.7% of the French species distribution range. Individuals used smaller home ranges (min=84; max=1814km(2)), consisting in a network of foraging areas and roosts that they visited. The daily and weekly turnover rates in the observed area of the landfill were high (on average 0.68 +/- 0.2 and 0.36 +/- 0.17, respectively) and the actual presence of ravens displayed strong seasonal variation. Hence, one-time/local management actions affected only a limited proportion of the population, partly explaining their small impact. Our study provides information needed to improve future management plans in a context of increasing ravens populations and conflicts with human activities.