ABSTRACT
Knowledge about the cumulative impacts of anthropogenic activities and environmental conditions on marine ecosystems is incomplete and details are lacking. Compositional community changes can occur along gradients, and community data can be used to assess the state of community resilience against combined impacts of variables representing human pressures and environmental conditions. Here we use a machine learning approach, i.e., Gradient Forest, to identify explanatory variable thresholds and select relevant epibenthic fauna and demersal fish species, which can be used to inform an integrated management of multiple human pressures and conservation planning in the southern North Sea. We show that a broad selection of anthropogenic and environmental variables, such as natural disturbance of the seafloor and euphotic depth, determined community composition thresholds of 67 epibenthic fauna and 39 demersal fish species along environmental conditions and human pressure gradients in the southern North Sea between 2010 and 2020. This has the potential to inform resilience assessments under the Marine Strategy Framework Directive to promote and retain a good environmental status of marine ecosystems.
