Snow leopard connectivity through trans-boundary areas surrounding the Dzungarian Basin in Kazakhstan

The assessment of landscape connectivity is central to the development of robust conservation strategies. Such knowledge is particularly important to understand isolation and dispersal patterns of wildlife species. Using snow leopard (Panthera uncia) occurrence data from Kazakhstan, this study sought to explore connectivity patterns in trans-boundary areas, with a particular emphasis on the Dzungarian basin, a geographic feature regarded as a barrier to snow leopard movement and range connectivity. We employed a multiple-scales species distribution model to reveal drivers of snow leopard occurrence, generating also a reference model reflecting the historical distribution of snow leopards in the country. We then adopted resistant kernels and factorial least-cost paths to infer core areas to snow leopard persistence and connectivity corridors throughout the study area. Snow leopard occurrence was driven by fine-scale topographic attributes and the configuration of open areas and forest patches. Kazakhstan possessed a low proportion of core patches but hosted critical stepping-stone areas to range connectivity along borders. Factorial least-cost paths suggested limited connectivity through trans-boundary areas surrounding the Dzungarian Basin in Kazakhstan, which only appeared as stepping-stone patches when snow leopard dispersal was high, suggesting the existence of source-sink population dynamics. Conservation efforts must be directed to ensuring and restoring functional connectivity across the main trans-boundary mountain ranges and from these to the stepping-stone core patches in the Dzungarian basin. Increased monitoring efforts in trans-boundary areas and collaboration among range countries are advocated to improve knowledge on snow leopard dynamics in this part of its range.