Abstract:
Understanding elephant habitat use and movement is essential for conservation in dynamic, resource-limited
semi-arid
ecosystems.
Remote sensing and GPS telemetry provide powerful tools for quantifying elephant ecological patterns across heterogeneous
landscapes. This study investigates the habitat preferences, movement patterns and spatial clustering of African savannah
elephants (Loxodonta africana) in Mana Pools National Park (MPNP), Zimbabwe, using GPS telemetry data and remote sensing–
derived environmental variables. Habitat suitability was modelled using the maximum entropy (MaxEnt) modelling technique.
Seasonal home ranges and movement dynamics were analysed using minimum convex polygons (MCP), Voronoi polygons and
the Time local convex hull (T LoCoH) techniques. Clusters were identified using the K-means
method. The MaxEnt results revealed
that proximity to permanent water sources particularly, the Zambezi River was the most significant predictor of elephant
distribution, contributing over 96% to the model's performance. Home range size estimates varied across methods and seasons.
During the dry season, MCP, Voronoi polygons and time local convex hull (T-LoCoH)
estimated ranges of 406 km2, 521 km2 and
325 km2, respectively. In the wet season, home ranges expanded markedly to 975 km2 (MCP), 713 km2 (Voronoi) and 527 km2 (T-LoCoH).
The transition season recorded the largest ranges, with 1065 km2 (MCP), 1032 km2 (Voronoi) and 714 km2 (T-LoCoH).
MCP consistently produced the largest estimates, while T-LoCoH
yielded the smallest, highlighting methodological sensitivity in
quantifying elephant ranging behaviour across seasonal landscapes. K-means
clustering identified spatially distinct movement
clusters across all seasons, showing non-random
habitat use driven by environmental constraints such as water availability,
vegetation productivity and terrain. Findings support data-driven
conservation strategies for sustainable elephant management
in protected areas.
