Fig 1.
Composite of suitable areas for elephants for individual spatial layers.
Areas suitable for elephants are calculated for each geospatial data separately. These include vegetation (A) [29], slope (B) [28], distance to rivers (C) [30], human population (D) [31], cattle density (E) [32], crop probability (F) [33], and protected areas (IUCN I-VI) (G) [35]. The intersection of these seven layers provides our projection of suitable landscapes for elephants (H).
Fig 2.
Amount of land and elephant presences within distances of varying river sizes.
Accumulation curve of elephant telemetry points as the distance increases away from rivers of varying flow orders. The red dashed line indicates the 10 km preference threshold.
Fig 3.
Fences (red) and elephant paths (lines between consecutive telemetry points) (black) across southern Africa. Protected areas (grey in inset maps) with elephants (dark brown) or without (light brown) appear on the large map.
Fig 4.
Map of elephant paths in northern Namibia.
Fenced boundaries (red) [4] around Etosha National Park prevent elephant movement (black) towards the north, east, and south. The arid landscape restricts movement in the west to riverbeds (symbolised by average annual flow rates from blue to purple) [30]. Green areas in the subset map mark protected areas from the WDPA [35]. Basemap Source: ESRI, MAXAR.
Fig 5.
Map of elephant paths in Luangwa and Zambezi.
Telemetry paths of elephants (black lines) in part of the Luangwa and Zambezi clusters overlaid on a map of slopes [28]. Green areas in the subset map mark protected areas from the WDPA [35].
Fig 6.
Distribution of elephant presences by time inside and outside Kasungu National Park.
Histograms show elephant telemetry records’ hourly distribution for elephants living in Kasungu National Park in Malawi. The data are split between those within the national park boundaries (A) and those outside (B). The shaded regions of the graphs approximately represent night (17:00–5:00). The difference in distributions suggests elephants more often appear outside the park at night, a behaviour typically associated with crop raiding.
Fig 7.
Map of elephant paths and crop probability in the Limpopo cluster.
Telemetry paths of elephants in Kruger National Park (South Africa), the adjacent Limpopo National Park (Mozambique), and Gonarezhou National Park (Zimbabwe) overlaid on a map of crop probability (blue) [33]. Fences are marked in red [4], and the WDPA I-VI protected area boundaries are green [35]. The proposed Greater Limpopo Transfrontier Conservation Area is in orange.
Fig 8.
Map of elephant paths and human population in the Chobe cluster.
Telemetry paths of elephants (black) in the Zambezi and northern Botswana regions overlaid on a human population density map (orange) [31]. Rivers (blue-purple lines) [30] and fences (red lines) [4] also act as barriers to elephant movement. Green areas in the subset map mark protected areas from the WDPA [35].
Fig 9.
Map of notable routes of connectivity or lack thereof.
Predicted suitable habitat (green), protected areas with elephants (orange), and protected areas without elephants (grey). The numbers refer to interesting connections and are described further in the main text.