Assessment of key reproductive traits in the Apennine brown bear population

Although knowledge of reproductive parameters is critical to project the chances of persistence of small populations, no data on basic reproductive traits have ever been estimated for the relict Apennine brown bear population (central Italy). From 2005-2014, as part of an ongoing ecological investigation, we compiled re-sight data on marked adult female bears (3 ≤ n ≤ 10 per year, 78 bear-years) and unmarked, distinct family groups (n = 18) to estimate litter size (1.9±0.7 SD cubs), weaning time (range: 7 May-6 June), interbirth interval (3.3-3.7 years), and reproductive rate (0.19±0.13 SD female cubs/adult female/year). We also applied multi-event models in a capturerecapture, robust design framework to our re-sight dataset of marked adult female bears to account for imperfect detectability of female bears with non-functioning radiocollars. Due to a high annual probability of re-sighting (0.77-0.82 for reproductive and non-reproductive females, respectively), and a negligible classification error (p = 0.003), multi-event models produced results similar to naïve estimates. Female bears had highest probability to reproduce 3-4 years after their last reproduction, and their average reproductive rate was 0.24 (95% CIs = 0.07-0.59). In addition, we produced a naïve estimate of apparent cub survival (0.49) as the proportion of cubs that was seen again the following spring before weaning, and a bias-corrected estimate of adult female bears survival (0.93, 95% CIs = 0.83-0.97), as a derived parameter of multievent models. Our findings tentatively place the Apennine bear population at the lower bound along the spectrum reported for other non-hunted brown bear populations. Coupled with high levels of human-caused mortality, a relatively limited reproductive performance could explain why Apennine bears failed to expand their range during the last decades. More in-depth demographic investigations are required to support our results and eventually assess whether they reflect density-dependent mechanisms or inbreeding depression.