Craig DeMars (MS, PhD candidate), Stan Boutin (PhD), University of Alberta
December 2014
Executive Summary
The boreal ecotype of woodland caribou (Rangifer tarandus caribou) is federally listed as Threatened and provincially designated as Red-listed in British Columbia due to population
declines throughout much of its distribution. High rates of calf mortality due to predation are a key demographic factor contributing to population declines and increasing predation of caribou
has been linked to landscape disturbance within and adjacent to caribou range. Developing effective management strategies for sustaining caribou populations in multi-use landscapes
therefore requires an understanding of the spatial dynamics of caribou and their predators during the calving season.
In 2011, we initiated a four-year project to evaluate spatial factors influencing predation risk to boreal caribou calves in northeast British Columbia. The project was a collaborative effort
among government, industry, non-governmental organizations, First Nations and academia. The two primary objectives of the project were to: i) identify key attributes of calving habitat
and determine whether calving habitat constituted a discrete, identifiable habitat within caribou range; and ii) evaluate spatial factors influencing survival of neonate calves (< 4 weeks
old). The latter objective required an assessment of space use by wolves (Canis lupus) and black bears (Ursus americanus), the two main predators of caribou calves.
Over the project’s three years of data collection, boreal caribou continued to sustain high rates of neonate mortality. We also documented relatively low rates of parturition. Collectively,
these results translated to calf-to-cow ratios that dropped below 30 calves: 100 cows by mid- July. While our findings of high rates of neonate mortality are consistent with the predation-mediated
hypothesis for caribou population declines, the low rates of caribou productivity (fecundity and calf survival) may also suggest declining winter and/or summer range conditions.
Using GPS data from 56 radio-collared female caribou, we identified calving habitat in a multiscale framework that also assessed whether females were selecting calving habitat to reduce predation risk or to access higher forage quantity and/or quality to meet maternal nutritional demands. Across all scales, reducing predation risk was a dominant factor driving calving habitat selection by females. At the finest scale, calving sites were predominantly situated in treed bogs and nutrient-poor fens – land covers considered to be predator refugia – and forage attributes of calving sites did not differ from winter locations used by the same animals. Females continued to select for treed bogs and nutrient-poor fens when moving within calving areas, defined as those areas used by females with neonate calves. Females generally avoided locations within high densities of linear features and showed weak selection for locations with higher forage productivity.
Our largest scale of analysis focused on female selection of calving areas within caribou range. We used an individual-based, comparative approach that assessed for selection differences based on season and maternal status (e.g. with calf versus barren). In general, females moved from winter ranges dominated by treed bogs to calving areas situated in landscapes mosaics with a high proportion of nutrient-poor fen. This shift may indicate a forage-risk trade off because fens are more productive than bogs but provide less of a predator refuge. Within these mosaics, females situated calving areas away from rivers, lakes and anthropogenic disturbance. Comparisons based on maternal status suggested that the presence of a neonate calf intensified selection behaviours associated with reducing predation risk.
We conducted similar multi-scale analyses of predator habitat selection. During the calving season, wolves were not confined to specific areas within caribou range; rather, pack territories were tightly spaced and overlapped significantly with caribou range and core areas. At a finer scale, wolves were closely associated with aquatic areas, showing selection for nutrient-rich fens and being closer to rivers and lakes than expected. This association is consistent with the hypothesis that wolves switch to beaver (Castor canadensis) as primary prey during the spring and supports previous studies highlighting the importance of water to wolves during the denning period. Wolf response to disturbance was counter to expectations as early seral vegetation and areas of high linear feature density were generally avoided. We further assessed wolf response to linear features by determining whether wolves preferentially select certain linear features over others. Of the two factors assessed, our results suggest that wolves select lines that increase movement efficiency and, secondarily, sightability.
In contrast to wolves, black bears were more predictable at larger spatial scales, favouring landscapes dominated by upland deciduous forest. Areas used by bears were also closer to early seral vegetation and had higher densities of linear features. Across all scales, bear were closer to aquatic features than expected and showed strong selection for rich fens when in caribou range. In general, selection patterns by bears suggested a preference for habitats associated with higher grass and forb abundance, which are important food sources for bears in the early spring.
We assessed the influence of spatial factors on the probability of calf survival by evaluating four hypotheses that described impacts from disturbance, refuge effects from lakes and peatlands, and predation risk from bears and wolves. We assessed each hypothesis at multiple scales and related calf survival to spatial factors using two metrics: exposure and maternal selection of habitat. The probability of calf survival was best predicted by a model representing predation risk from bears. Specifically, the model suggested that calf survival depended on the density of high quality bear habitat surrounding locations selected by females within the calving area. All other hypotheses were generally unsupported and we found no evidence to suggest that any one specific landscape feature contributed disproportionately to the probability of calf survival. This lack of support may suggest that: i) the degree of disturbance within caribou range has exceeded thresholds where differences in neonate survival may be detected; and/or, ii) neonate survival may be driven by predator densities more than variation in spatial factors.
Results from this project highlight the challenges of managing calving habitat for caribou in multi-use landscapes. Management actions will need to be conducted at large spatial scales because caribou are at their most dispersed at calving and small-scale actions will likely be ineffective at improving rates of neonate survival. Targeting potential management actions toward large fen complexes may be most effective because of their importance to calving caribou. For rapidly declining populations residing in ranges highly impacted by disturbance, habitat conservation and restoration initiatives may need to be augmented by more intensive actions (e.g. maternal penning, predator management) to improve calf survival rates in the short-term.
# 09-9196-50 / 13-AU-ERPC-01