Marine spatial planning provides a framework to minimize conflicts of interest between conservation and economic use in European shelf sea regions. One outcome of marine spatial planning is the designation of marine protected areas, which are an important tool to safeguard marine biodiversity and improve the marine environmental status. Site selection for marine protected areas requires detailed habitat maps, which display at the ultimate, most detailed level of resolution the composition of the inhabiting species communities. Decades of environmental research and monitoring have produced a wealth of data on the seafloor sediment composition and the occurrence and distribution of marine benthic species. We used an extensive data set based on 14 years of sampling to develop an updated map of the infauna associations of the southeastern North Sea. Fuzzy clustering was applied to identify and differentiate specific infauna associations. The distribution of the associations was modelled using Random Forests. Four different infauna associations were identified in the study region based on their inventories of characteristic and accompanying species. A high structural complexity of the benthic environment became evident where numerous scattered occurrences of course sediment occurred, particularly inside the marine protected areas. The predictive uncertainty of the spatial modelling was displayed allowing for a realistic representation of the dynamic transition zones between neighboring habitats. These transition zones may deserve attention in future investigations as habitats of elevated benthic biodiversity. The updated map of infauna associations of the southeastern North Sea facilitates the identification and localization of vulnerable biotopes to support administrative management processes. • Statistically modelled infauna associations of the southeastern North Sea • High-resolution predictive mapping to support marine spatial planning and management • Detailed spatial patterns inside Marine Protected Areas revealed by modelling
Gutow et al. (Fri,) studied this question.