Den selection criteria

Published: 14 February 2023| Version 1 | DOI: 10.17632/53kgyy3rth.1
Audrey Moizan,


Climate warming is favoring the expansion of non-native species onto the Arctic tundra, where they may compete over resources with native species. In the harsh tundra conditions, sympatric red foxes (Vulpes vulpes) and Arctic foxes (Vulpes lagopus) may compete over denning sites, which are important for their reproduction and survival. We studied den selection by red and Arctic foxes in spring and summer, and their possible competition over this resource in an ecotone near Churchill, Manitoba, on the west coast of Hudson Bay, by examining patterns of den occupancy related to den characteristics and spacing patterns between neighbors. Based on 11 years of occupancy data for 42 tundra dens, we determined that red and Arctic foxes favored dens based on shelter quality in both spring and summer, rather than proximity of specific habitats (and thus specific prey). Mechanisms of den selection differed between species, which may promote co-existence, and areas of high den density were avoided by red foxes and preferred by Arctic foxes. We did not find evidence of exclusion of Arctic foxes by red foxes: spacing patterns showed that foxes spaced themselves based on their need for space, territoriality and food availability but not interference. In the current abiotic Arctic conditions, taiga species settling on the tundra could coexist with tundra endemics, at given density thresholds of both competitors. As Arctic conditions may become milder, an increase in newcomer abundance could disrupt the current balance that favors species coexistence. This study was carried out in and around Wapusk National Park, on the western edge of Hudson Bay near Churchill, Manitoba, Canada (58°45’ N; 94°10’ W), where red foxes have settled in tundra dens to reproduce exhibiting a continuous presence since 2010.


Steps to reproduce

From 2011 to 2021, we monitored fox dens geolocated on the tundra in and around Wapusk National Park twice a year, to assess the occupancy of both red and Arctic foxes. We first visited dens in March-April by snowmobile, when foxes have paired up at den sites to start breeding, and again in late July-August by foot, all-terrain vehicle and/or helicopter, near the end of the breeding season shortly before pups disperse. The species occupying a den cannot always be assessed accurately from the air, so we used 42 of the 120 natal dens of our study area that were visited from the ground both in spring and summer. We determined if a den was occupied in spring based on the presence of tunnels through the snow and assessed the species based on shed fur. In summer, we assessed occupancy based on recently excavated burrows, tracks, fresh scats, urine odor, or direct sighting of foxes visually or from trail cameras (since 2015, 21 dens were equipped with cameras). When the species using a den could not be verified (e.g., no shed fur or sightings), we recorded the species as “unidentified” and excluded the den from subsequent analyses. To characterize the dens, we measured 16 variables. We recorded 4 of those variables in the field from March to June 2019 (8 dens still covered in snow in June were revisited in August) and used habitat characteristics as proxies for food availability, based on published habitat preferences of prey from our study area. We extracted these 12 remaining habitat variables using ArcGIS 10.6 (Esri 2018) and ArcGis Pro 2.4 (Esri 2019). Five environmental variables were extracted from Canada’s land cover map (version 2015) . Hydrographic data (length of stream and length of pond bank) were extracted from Natural Resources Canada (2017). We extracted intertidal flats from a 5-m resolution microhabitat map of Wapusk National Park and merged it with the other variables in one raster. Coastline and treeline were drawn as a vector file based on this same microhabitat map. We defined coastline using the western edge of the intertidal flats, and defined treeline as the limit between tundra and forested habitats (spruce larch forests and woodland) using the eastern and northern edges of the woodland and forest-type microhabitats. We estimated lemming densities using mark-recapture methods each June, after snow melted, from 2011 to 2021. We captured small mammals using Longworth and Sherman live-traps on 2 grids (8x8, with 15-m spacing and 2 traps per station) and 2 transects (1x20 with 15-m spacing and 3 traps per station) on hummocky lichen-heath tundra, the habitat most commonly inhabited by collared lemmings, Dicrostonyx richardsonii. Traps were baited with peanut butter and oats and left open for 72 hours, and captured animals marked by hair-clipping.


University of Manitoba, Sorbonne Universite


Animal Science, Arctic Tundra Biome, Polar Ecosystem, Animal Ecology


Natural Sciences and Engineering Research Council of Canada

University of Manitoba

Churchill Northern Studies Centre

Polar Continental Shelf Program