Mothers’ Movements: Shifts in Calving Area Selection by Partially Migratory Elk

• Published in: The Journal of wildlife management 2021-09, Vol.85 (7), p.1476-1489
• Main Authors: BERG, JODI E., REIMER, JODY, SMOLKO, PETER, BOHM, HOLGER, HEBBLEWHITE, MARK, MERRILL, EVELYN H.
• Format: Article
• Language: English
• Subjects: Algorithms; Animal behavior; Animal migration; Bears; Breeding success; calving; Canis lupus; Cervus canadensis; Costs; Elk; Forage; Foraging habitats; Global positioning systems; GPS; Habitat Relations; habitat selection; Implants; Learning algorithms; Machine learning; Migrants; Migratory species; National parks; Offspring; partial migration; Parturition; Predation; Predators; random forest; Reproduction; Risk; Summer; Tactics; Ungulates; Ursus; Vagina; Wildlife habitats; Wildlife management
• ISSN: 0022-541X; 1937-2817
• DOI: 10.1002/jwmg.22099

Loss of migratory behavior or shifts in migratory ranges are growing concerns to wildlife managers. How ungulates prioritize safety from predators at the expense of high-quality foraging opportunities during calving may be key to understanding these shifts and long-term reproductive success. We compared trade-offs in selection for forage and predation risk by elk (Cervus canadensis) following 3 migratory tactics (western and eastern migration and resident) during 2 time periods in a declining (by almost 70% from 2002–2016), partially migratory elk population adjacent to Banff National Park in Alberta, Canada. We hypothesized that regardless of migratory tactic, maternal elk would show stronger trade-offs between high-quality foraging associated with higher predation risk and forage resources of lower-quality yet lower risk on calving ranges than on ranges used during summer because of vulnerability of their offspring. Additionally, we hypothesized these trade-offs would occur at high (2002–2006) and low (2013–2016) elk population sizes. We used a machine-learning algorithm to predict dates of parturition based on global positioning system (GPS) movements of elk equipped with vaginal implants (n = 60) and predictions were within 1.43±0.85 (SE) days of the known date. We applied the model to an additional 58 GPS-collared elk without vaginal implants. Based on changes in localized movements, we defined calving areas as the 26 days post-parturition and compared habitat characteristics of calving areas to 10 similar-sized areas centered on random locations during summer for the same individual in a latent selection framework. Across the 2 time periods, parturition occurred from 8 May–11 July with median parturition dates differing among migratory tactics and residents shifting towards an earlier parturition date in the later period. All elk, regardless of migratory tactic and time period, selected calving areas with greater forage resources than were available on areas used during summer, with no evidence for greater selection of areas that reduced predation risk at the expense of higher-quality foraging. Calving season selection for areas with abundant forage exposed western migrants to high risk of bear (Ursus spp.) predation, residents to high risk of wolf (Canis lupus) predation, and eastern migrants to low risk of bear and wolf predation. Patterns in exposure to predation risk during calving between migratory tactics were consistent with the recent decline