Socio-behavioural and spatial ecology of African elephants (Loxodonta africana)

Abstract

African savannah elephants (Loxodonta africana) are among the most dominant terrestrial species in Sub-Saharan ecosystems. Their complex spatial and social ecology make them a valuable model-species for behavioural field study. This thesis investigates selected aspects of their spatial and social behaviour in three populations in southern Africa. Satellite tracking data of 19 elephants were collected in two different ecosystems, the woody bushveld neighbouring Kruger National Park, South Africa (six individuals) and arid savannah of Etosha National Park, Namibia (13 individuals). The spatio-behavioural pattern of elephants (frequency and duration of visits in habitat patches) was quantified using Time-scaled Local Convex Hull (T-LoCoH) models and tested against that calculated from random walk models. Seasonal resource selection functions (RSF) were constructed for each individual and population to examine the effects of different environmental features on elephant habitat use pattern. In both study areas, elephants return to selected habitat patches more frequently than expected by random, likely to access water. The duration of visits, however, differs between ecosystems. In the bushveld, elephants stay longer at habitat patches in dry seasons and the opposite is true in the arid savannah. They are generally attracted to surface water and productive areas. However, as resource availability and quality decrease, particularly in arid conditions, individuals demonstrate different seasonal habitat use pattern (e.g. natural vs. artificial water sources, habitat productivity vs. stability). These findings indicate that elephant spatio-behavioural resource use patterns are subject to local ecological regime. Furthermore, substantial intra-population heterogeneity among individuals suggests that socio-sexual factors, such as dominance hierarchy and reproductive state, play important roles in elephant spatial behaviour and habitat selection. Concurrent with the above, a photo-identification study was conducted on a socio-demographically disrupted elephant population in Pilanesberg National Park, South Africa. Social complexity and structure of the population were assessed with social network modelling and clustering techniques. Grouping pattern and temporal stability of associations were quantified and compared against socio-behavioural models. Elephant sociality was measured as different network metrics. Sociality types were investigated with a two-step principal component analysis (PCA)-clustering analytical framework, depicting similar sociality traits. The social structure of female elephants indicates a stable, family-centred hierarchical society. However, a noticeably more fluid bond-group structure is likely an effect of past socio-behavioural disturbance that happened four decades ago. Male associations, generally fluid and weak, evolve with males’ age. Their temporal association pattern follows a ~300-day cycle, suggesting the influence of annual musth cycle, which may drive natal dispersal of young males. Correspondingly, the population demonstrates sociality types that are interrelated with social units for females and individual age for males. These types may function as a platform for information exchange and repositories of socio-ecological knowledge, which further underscores the importance of functional socio-demographic structure in the wellbeing of elephant population. With several diverse research topics, this study contributes to a more holistic understanding of elephant social and spatial ecology. I discuss relevant conservation implications, and the links between socio-demography and behaviour of elephants, which may help to better understand the development of their complex society.