Contrasting African elephant movescapes across xeric and mesic savannas

Expanding human activities increasingly influence animal movement. Yet, conservation interventions often overlook intraspecific variation in how movement is structured across populations inhabiting distinct ecosystems. For wide-ranging generalists such as African savanna elephants (Loxodonta africana), understanding how responses to resource availability and anthropogenic pressure vary across populations is critical for effective conservation planning. Here, we analyze long-term GPS tracking data from 175 African elephants to contrast how movement is structured across mesic (Mara-Serengeti) and xeric (Samburu-Laikipia) savannas in East Africa. We delineate the landscape of elephant movement (a "movescape"), using a network-based approach that integrates movement path, use intensity, and connectivity metrics to describe how animals use discrete locations for different functions. We focused on identifying highly used areas and movement corridors, then used mixed-effects spatial regressions to evaluate how environmental context influenced the occurrence of these areas within and across ecosystems. All analyses in Mara-Serengeti represent a new application of this framework and are compared with previously published results from Samburu-Laikipia. Elephant movement structure differed between ecosystems. In the Mara-Serengeti, movement networks were dense and clustered, with more locations used repeatedly and connected by redundant pathways. Movement in Samburu-Laikipia was more spatially dispersed. Most locations were used by elephants with low intensity, and connectivity was concentrated in fewer corridors. Across both ecosystems, core-use areas comprised a small proportion of the landscape, highlighting locations where conservation efforts could be targeted in unprotected areas. In the Mara-Serengeti, elephants concentrated use in areas with low boma density and away from villages. In Samburu-Laikipia, core areas and corridors were strongly associated with permanent water and consistent vegetation productivity, likely reflecting more ecological constraints on movement. Conservation planning in the Mara-Serengeti could prioritize riparian woodlands and bush thickets on hillsides to protect connectivity between core elephant areas. In Samburu-Laikipia, protecting existing limited corridors that provide access to seasonally productive forage and access to permanent natural water sources could be effective. Our combined delineation of core-use areas and the corridors that link them provides a framework for prioritizing both core habitat patches and the movement pathways that sustain landscape-scale connectivity.