Time vs. distance: Alternate metrics of animal resource selection provide opposing inference

• Published in: Ecosphere
• Main Authors: Serrouya, Robert, Kellner, Amanda, Pavan, Gary, Lewis, Doug W., DeMars, Craig A., McLellan, Bruce N.
• Other Authors: British Columbia Forest Science Program, British Columbia Forest Investment Account, Simpcw First Nations Band, World Wildlife Fund Canada, Parks Canada
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 2017
• Subjects: caribou; Rangifer tarandus
• Online Access: http://dx.doi.org/10.1002/ecs2.1730; https://api.wiley.com/onlinelibrary/tdm/v1/articles/10.1002%2Fecs2.1730; https://esajournals.onlinelibrary.wiley.com/doi/pdf/10.1002/ecs2.1730

Abstract Quantifying resource use or selection by valued species on a human‐altered landscape is important for wise conservation action. Here, we contrast metrics of resource selection based on Global Positioning System ( GPS ) telemetry, which indexes time spent in various habitats, with tracking in snow, which measures distance travelled. When animals move at different speeds within different habitats, the two methods can produce different results. The study was conducted in winter on endangered mountain caribou ( Rangifer tarandus caribou ) in a landscape fragmented by forestry roads and cutblocks. From 2001 to 2009, 25 caribou were monitored with GPS collars obtaining locations every two to six hours, and compared to 63 caribou trails followed in snow in the same study area and time period. Logistic regression was used to estimate selection for both metrics, with the same definition of use and availability employed. Forest age, tree species, edge distances, and road densities were the covariates of interest. For most covariates, the two metrics produced similar results that agreed with the literature: selection for primary forest stands that represent forage patches. Both metrics indicated selection for habitat edges that potentially enhance foraging. The distance metric (snow trailing) indicated strong selection for forestry roads (42% more than available), whereas GPS locations suggested an avoidance of roads (33% less than available). The GPS analysis was in agreement with the vast majority of work published on woodland caribou, whereas the distance metric suggests some of the first evidence of selection for anthropogenic linear features. Our results highlight a potential bias against detecting selection for habitat features used for movement when using long fix intervals (>2 h) that typify most GPS studies. Avoidance of linear features should be carefully examined and designs considered to asses this bias. Finally, the selection of roads by caribou exacerbates an already desperate situation by ...