Detection of pregnancy in a hibernator based on activity data

Abstract

Listen

Translate article

Pregnancy and parturition are often difficult to detect in wild-living animals, especially in species that give birth to altricial young in burrows or caves. Current methods to detect pregnancy and birth in hibernating animals often entail disturbances that can affect the animals' reproductive success. We developed a new method to confirm pregnancy and parturition in hibernating brown bears (Ursus arctos) using activity data recorded in dual-axis motion sensors mounted on GPS–GSM neck collars on 30 brown bears during 52 hibernation seasons in Sweden. We adjusted recorded activity to the individual basal activity level for each bear. Pregnant females showed characteristic activity patterns during the period of pregnancy, with significantly higher daily activity levels and frequency of active periods during pregnancy than nonpregnant bears. Pregnant females were active on average 2.20 times more often during the pregnancy period than during the postpartum period, compared with 0.97 times for nonpregnant females. A pregnancy index was defined as the average of the proportion of mean daily activity levels and the proportion of activity events during the pregnancy period compared with the postpartum period. It averaged 2.61 ± 1.73 (SD) for pregnant females and 0.94 ± 0.24 for nonpregnant females. Using this method, we evaluated a group of adult females that had an unverified reproductive status but were classified as not pregnant because no cubs had been observed after den emergence. The results suggested that 35 % of those females had, in fact, been pregnant.