Landscape Genetics of Fragmented Elephant Habitats

Abstract

Habitat fragmentation is the foremost driver of genetic erosion in large terrestrial mammals, severing dispersal corridors that historically maintained gene flow between populations and accelerating inbreeding depression in isolated remnant groups. This study characterised landscape genetic structure in African savanna elephants (Loxodonta africana) across 18 fragmented reserves and wildlife management areas in East and Southern Africa, using 22 microsatellite loci genotyped from non-invasive faecal DNA samples of 842 individuals collected between 2018 and 2023. Bayesian clustering (STRUCTURE, K = 5) identified five genetically distinct population clusters corresponding broadly to reserve boundaries, with mean pairwise FST of 0.148 +/- 0.038 across cluster pairs. Landscape resistance modelling using circuit theory (Circuitscape) identified human settlement density (beta = +0.84, p < 0.001), road network intensity (beta = +0.62, p < 0.001), and cultivated land cover (beta = +0.71, p < 0.001) as the strongest barriers to elephant gene flow, collectively explaining 74% of variance in pairwise genetic distance (Mantel test r = 0.74, p < 0.001). Effective population sizes (Ne) estimated by the linkage disequilibrium method ranged from 28 (most isolated reserve) to 214 (largest connected cluster), with five of 18 reserve populations showing Ne below the minimum viable population threshold of 50. A Connectivity Priority Score (CPS) integrating genetic isolation, Ne, and landscape resistance ranked three transboundary corridor zones as highest priority for legal protection and human-wildlife conflict mitigation to restore functional genetic connectivity across the study network