Sudden changes in habitat quality during the breeding season may mislead individuals when selecting their nesting site and result in population declines. In such cases, even semi-natural and extensive agricultural lands may become ecological traps. We examined how the availability of six open habitat types (i.e. agricultural lands, open forests, alder swamps, young forests, regeneration, and wetlands) could be affecting the habitat selection process, as well as the hatching and fledging successes of American kestrels (Falco sparverius). We hypothesized that natural open habitats are less disturbed by anthropogenic activities than extensive agricultural lands and thus represent higher quality habitats for kestrels. We also considered weather conditions during the breeding season as possible factors affecting hatching and fledging successes. We monitored 200 pairs of American kestrels during 11 years (2006–2016) within a network of 155 nest box stations and we characterized landscape composition metrics within 800?m radii from each nest box. We used generalized linear mixed models and multimodel inference to quantify the effects of landscape composition metrics on the probabilities of using nesting site, hatching success, and fledging success of American kestrels. We also tested the effects of weather conditions and clutch initiation date on hatching and fledging successes of kestrels. We found that the probability of nesting site use increased with the amount of agricultural lands. Hatching success decreased with the amount of agricultural lands, whereas the fledging success of kestrels did not vary with the amount of agricultural lands. Both the probabilities of hatching and of fledging increased with the area of young forests. There was no evidence of a weather effect on hatching success. However, the probability of fledging success increased with mean temperature during the raising period of nestlings. Although fledging success alone does not determine fitness or population dynamics, our results suggest that kestrels nesting in this region at the northern limit of their range may be caught in an ecological trap by extensive agricultural lands. Indeed, kestrels were attracted by meadows when selecting nesting habitat, but the hatching and early nestling periods coincided with hay harvesting which could reduce the hunting success of breeding adults and suddenly alter food availability. Although the causes of recent kestrel declines remain unclear, our results suggest that harvesting practices, even those related to extensive perennial agriculture, may have a negative effect on the breeding success of the species compared to areas dominated by young forests.

The North American beaver is considered a keystone species because its behaviour leads to profound changes in the wetland systems within forested landscapes. Such changes influence species composition and their interactions. However, in some cases, beavers are considered as an important source of disturbance and conflict with anthropogenic activities. In this paper, we reviewed regional studies using geomorphology, food availability and anthropogenic infrastructure on spatial modelling of beaver habitat. Even though all studies used different sets of variables and methodologies, important factors affecting beaver occurrence or abundance are mainly stream gradient, watershed size and hardwood cover that is adjacent to the streams. However, the identification of key habitat indicators often varies between studies depending upon the object being modelled (colonies vs. dams), the geomorphological characteristics of the region, and the scale of the study area. Recent developments in geomatics and improved data quality now allow spatial modelling of beaver habitat across larger areas, and make models using at least stream gradient and forest cover types more accessible to managers. Such large-scale predictive beaver habitat models could have valuable applications for the prevention of infrastructure damage and related costs, and for managing wildlife species that rely upon beaver ponds.

Le castor du Canada est considéré comme une espèce clé car son comportement cause d’importantes modifications aux milieux humides au sein des paysages forestiers. De tels changements influencent la composition des espèces ainsi que leurs interactions. Cependant, dans certains cas, les castors sont considérés comme une source importante de perturbation et de conflit avec les activités anthropiques. Dans cet article, nous faisons la recension des études régionales qui ont utilisé la géomorphologie, la disponibilité de ressources alimentaires et la présence d’infrastructures anthropiques pour réaliser une modélisation spatiale de l’habitat du castor. Même si les études ont utilisé des variables et des méthodologies différentes, les principaux facteurs affectant l’occurrence ou l’abondance du castor sont le gradient des ruisseaux, la superficie du bassin hydrographique et le couvert de peuplements feuillus adjacents aux cours d’eau. Cependant, l’identification des principaux indicateurs d’habitat varie selon les études en fonction de la variable modélisée (huttes ou barrages), des caractéristiques géomorphologiques de la région et de la taille de la zone d’étude. Les développements récents de la géomatique et l’amélioration de la qualité des données disponibles permettent maintenant la modélisation spatiale de l’habitat du castor pour de grandes régions et rendent les modèles utilisant au moins le gradient des ruisseaux et les types de couverts forestiers plus accessibles aux gestionnaires. De tels modèles prédictifs de l’habitat du castor à grande échelle pourraient avoir des applications importantes pour la prévention des dommages aux infrastructures et des coûts qu’ils engendrent, ainsi que pour la gestion des espèces fauniques qui dépendent de la présence d’étangs de castors.