The polar bear (Ursus maritimus) serves as a flagship species in zoological institutions, contributing
to both indirect and potential future direct conservation efforts. However, concerns over the welfare
of these bears in captivity have been raised, and contemporary initiatives soon require institutions to
evaluate and monitor welfare through evidence-based welfare assessment tools. The overall
objective of this PhD study was to therefore to commence the development of a welfare assessment
protocol for zoo-housed polar bears, focusing on the appropriate behaviour welfare principle of the
Welfare Quality® framework.
In a critical review of scientific publications, several potential animal-based welfare indicators were
found. Based on content, construct and criterion validity, only limited evidence of validity was
established for the indicators, with abnormal repetitive behaviour being the only thoroughly
validated behavioural indicator. The identified gaps in knowledge encouraged the remaining parts
of this PhD study, concerning developing and validating indicators within the welfare criteria of
‘positive emotional state’, ‘expression of species-specific normal behaviour’ and ‘appropriate social
environment’.
To enable inference of the emotional state, a fixed-term Qualitative Behaviour Assessment (QBA)
list was developed, of which validity and short-term consistency was investigated. QBA was carried
out on 22 polar bears housed in nine zoos in Denmark, Germany, The Netherlands and France,
concurrent with collection of several behavioural, postural and health-related indicators. Two
components were extracted through Principal Component Analysis (PCA) and coined as Valence
and Arousal, which displayed evidence of construct validity (both convergent and divergent)
through meaningful significant associations of component scores to the other animal-based welfare
indicators. Positive association of valence scores was found to behavioural diversity, environmental
interaction, rest and negative significant associations was found to abnormal repetitive behaviour
and activity level. Arousal scores showed significant positive associations to abnormal repetitive
behaviour, environmental interaction and activity level as well as a negative association with awake
inactivity. The valence scores of the QBA was moreover found to be consistent within and between
days (short term), useful for feasibility of future assessments.
In the same study population, behavioural diversity, based on the Shannon Index (H), was assessed
for its potential use in polar bear welfare assessment concerning the criterion relating to species-
appropriate behaviour. Behavioural diversity showed both construct (convergent) validity through association with QBA valence scores, and construct (divergent) validity with time spent engaged in
abnormal repetitive behaviour. However, several caveats of this index persist, and the behavioural
diversity index should not be treated as an independent or fully validated indicator.
Lastly, the effect of the social environment on polar bear welfare is of current focus, yet no
immediate measure to assess this has been developed. Potential indices for monitoring social
dynamics were therefore investigated, along with the effects of these relationships on welfare,
assessed through associations to other welfare indicators. Potential dyadic factors explaining
variation in social qualities were moreover investigated. The social environment of the same study
population of polar bears (excluding solitary housed bears and a mother-cub dyad), was investigated
through PCA of multiple diverse social parameters, as well as through existing social indices used
for other species. Three components were found explaining dyadic relationships labelled Value,
Security and Incompatibility, and useful social indices capturing information on the social
environment were identified, along with a proxy measure (inter-individual proximity) for
feasibility. Relatedness was found to be significantly associated with positive dyadic relationships,
and bears in positive relationships was significantly less engaged in awake inactivity compared to
other bears, assessed through the adapted social indices. Although no clear-cut relationship between
the social environment and welfare was found, the proposed indices may prove useful for
monitoring social dynamics, which is imperative in socially housed polar bears.
The generated novel knowledge and indicators were compiled into an adapted welfare framework
based on Welfare Quality® and the 24/7 approach, serving as a prototype welfare assessment
protocol for zoo-housed polar bears. The protocol highlights current indicators, scoring schemes,
gaps in knowledge and future perspectives, and may serve as a starting point for protocol
development, important for the care, management and conservation efforts of this vulnerable
species.