Modeling the demography of species providing extended parental care: A capture-recapture approach with a case study on Polar Bears (Ursus maritimus)

1. In species providing extended parental care, one or both parents care
for altricial young over a period including more than one breeding season.
We expect large parental investment and long-term dependency within family
units to cause high variability in life trajectories among individuals
with complex consequences at the population level. So far, models for
estimating demographic parameters in free-ranging animal populations
mostly ignore extended parental care, thereby limiting our understanding
of its consequences on parents and offspring life histories. 2. We
designed a capture-recapture multi-event model for studying the demography
of species providing extended parental care. It handles statistical
multiple-year dependency among individual demographic parameters grouped
within family units, variable litter size, and uncertainty on the timing
at offspring independence. It allows for the evaluation of trade-offs
among demographic parameters, the influence of past reproductive history
on the caring parent’s survival status, breeding probability and litter
size probability, while accounting for imperfect detection of family
units. We assess the model performance using simulated data, and
illustrate its use with a long-term dataset collected on the Svalbard
polar bears (Ursus maritimus). 3. Our model performed well in terms of
bias and mean square error and in estimating demographic parameters in all
simulated scenarios, both when offspring departure probability from the
family unit occurred at a constant rate or varied during the field season
depending on the date of capture. For the polar bear case study,
we provide estimates of adult and dependent offspring survival rates,
breeding probability and litter size probability. Results showed that the
outcome of the previous reproduction influenced breeding probability. 4.
Overall, our results show the importance of accounting for i) the
multiple-year statistical dependency within family units, ii) uncertainty
on the timing at offspring independence, and iii) past reproductive
history of the caring parent. If ignored, estimates obtained for breeding
probability, litter size, and survival can be biased. This is of interest
in terms of conservation because species providing extended parental care
are often long-living mammals vulnerable or threatened with extinction.