2.2. Bear capture and monitoring

We captured, collared, and monitored bears from 2005 to 2018. Bears were immobilized by darting from a helicopter (see McLellan & McLellan, 2015). Capture was carried out in spring, shortly after den emergence when bears were feeding in avalanche chutes and open alpine meadows, or early autumn when they fed on huckleberries (Vaccinium membranaceum). Once immobilized, we weighed, measured, and fitted bears over 2 years old with either GPS or VHF collars (Lotek Inc.). We also obtained a tissue sample for genetic identification and a vestigial premolar for determining age via cementum annuli. We classified 2 to 5‐year‐old female bears as subadults and those 6 years and older as adults (Garshelis et al., 2005). Insufficient data were collected from males to include in this analysis. All collars had canvas spacers to ensure that the collar would drop off, and we weakened the canvas on the collars of subadult bears to ensure they dropped in about 1 year. The Animal Care Committee of the British Columbia Fish and Wildlife Management Branch approved and permitted all capture and handling protocols.

Capture effort, defined as the time we spent searching for bears, was evenly distributed between the study populations until 2014 when it had become apparent that the North Stein‐Nahatlatch population was not only small, but had an unusually high incidence of adult female mortality. Although no bears had been injured or killed in our research, the risk of additional female mortality due to capture was deemed too high, so captures from 2015–2018 were limited to the McGillvary Mountains population. We continued to monitor the vital rates of all collared bears in both populations until 2018.

Throughout the first part of the study (2005–2008), we located collared bears by fixed‐wing aircraft every 2 weeks from May to November. On each flight, we downloaded GPS location data and attempted to visually locate each bear and count the dependent offspring of females. If we did not find a bear for more than 8 consecutive weeks, we censored them from the analyses estimating survival and reproductive parameters at the time of their last known status. The populations were also monitored using genetic capture–recapture, and all females were genetically detected after their collars had dropped or stopped working; therefore, all had known fates while collared. In the second part of the study (2010–2018), some females were fit with VHF collars to reduce the number of recaptures needed to maintain continuity of monitoring. Collared bears were subsequently located and observed from a helicopter at least once each spring and then again in summer and fall. We intermittently located bears from the ground between aerial attempts. Offspring age was determined by size for cubs (bears <1‐year‐old) and yearlings. We grouped attendant offspring that were 2 years of age and older because we could not visually distinguish between 2 and 3‐year‐old bears with certainty if the previous year's status was unknown.

All collars were programmed to signal if the collar had not moved in 24 hr, and these mortality signals were investigated as soon as possible after detection. Whenever we found a dropped collar with rotted canvas, we assumed the bear was still alive. If a bear was found dead, we performed an investigation and necropsy in the field to determine the cause of death.