Habitat loss and fragmentation drive biodiversity decline, yet debate persists on their relative impacts and how to design landscapes for biodiversity conservation. While some prioritize habitat connectivity, others emphasize habitat quantity. However, the role of the landscape matrix and its resistance in population persistence remains poorly understood. To address this, we conducted a multigenerational, landscape-scale experiment using the model arthropod Folsomia candida, assessing the effects of matrix resistance and inter-patch distance on colonization, population size, and extinction while keeping a fixed habitat amount. We found that the amount of reachable habitat, integrating habitat amount and matrix resistance, is a strong predictor of population size and extinction risk. Survival across matrix types was the key mechanism, influencing both colonization and demography. Our study enhances understanding and predictive ability of population fate at the landscape scale, offering new insights for landscape ecology theory and valuable perspectives for applied conservation.

Problem statement: To halt global biodiversity decline, many conservation measures are set up by citizens, companies, or stakeholders. However, even if originally well-intentioned, some of these actions could have direct or indirect negative effects on biodiversity when ecology is not accounted for. The management of bees is a good example of such misplaced conservation practices. We identified three successive errors in the management of bees which can disrupt the focus on real conservation issues: the multiplication of honey bee hives, the installation of insect hotels, the trade of solitary bee cocoons for release into the wild. To help the bees, as well as biodiversity in general, we must consider prioritizing efficient conservation measures which consider more broadly the complexity of ecosystems.