Tage der Biodiversität 2027

The global biodiversity crisis is particularly dramatic in freshwater ecosystems which have suffered exceptionally high species loss. High local endemism of fish in Swiss perialpine lakes and the number of imperiled species in the surrounding rivers makes them sensitive to multiple anthropogenic threats. This creates an urgent need to allocate limited resources strategically across conservation and restoration amid competing demands for land and resources. Spatial prioritization models can support such decisions, but conservation and restoration are rarely integrated, and restoration planning often fails to quantify how specific threats shape species distributions, and therefore which actions would most effectively support recovery.

We present an interdisciplinary project involving scientists, government agencies, and NGOs that develops data-driven models integrating conservation and restoration in a heavily regulated watershed in Switzerland. We combined national observations of fishes with high-resolution environmental data to model habitat suitability and used a causal forest and explainable AI framework to estimate how individual threats (e.g., flow alteration, habitat fragmentation, pollution) constrain the distributions of 45 riverine fish species. We mapped these local causal effects as stressor-specific recovery potential and used them as features in a spatial restoration optimization explicitly linked to candidate actions (e.g., barrier removal, riparian revegetation, pollution abatement). In doing so, our approach moves beyond area-based habitat gains to identify spatial “leverage points” where feasible actions remove limiting threats and yield the greatest ecological benefits, directly connecting threat–species relationships, restoration actions, and expected recovery outcomes within a single quantitative planning framework.