Habitat Fragmentation Effects on Large Carnivore Metapopulation Dynamics

Abstract

Habitat fragmentation is widely recognised as the foremost driver of large carnivore population decline in human-dominated landscapes, yet the quantitative relationships between landscape configuration metrics and metapopulation viability remain incompletely characterised across taxa and biogeographic regions. This study integrated GPS telemetry, landscape genetic analysis, and spatially explicit metapopulation modelling to evaluate fragmentation effects on three large carnivore species — Panthera pardus, Canis lupus, and Lynx lynx — across 14 study landscapes in Central and Eastern Europe spanning a gradient from highly fragmented agricultural plains to continuous montane forest. A total of 312 GPS-collared individuals (P. pardus: n = 84; C. lupus: n = 128; L. lynx: n = 100) generated 2.4 million relocations over three years (2019–2021). Landscape resistance surfaces were calibrated against empirical movement data using circuit theory (Circuitscape 4.0). Metapopulation viability analysis (VORTEX 10.5) projected 100-year population trajectories under four habitat scenarios. Patch area (r = 0.79, p < 0.001) and inter-patch connectivity (measured as least-cost corridor width) were the strongest predictors of local occupancy probability across all three species. Minimum viable patch size was estimated at 380 km2 for P. pardus, 1,240 km2 for C. lupus, and 620
km2 for L. lynx. Corridor widths below 2.8 km reduced functional connectivity by > 60% for all species. Metapopulation persistence probability over 100 years dropped below 50% when landscape-scale connectivity fell below a threshold effective mesh size of 420 km2. These findings provide species-specific, empirically grounded thresholds for habitat network design applicable to EU Habitats Directive conservation planning.