Identifying key-conservation areas for Posidonia oceanica seagrass beds

The configuration of a habitat often reflects its conservation status and, to be effective, the conservation actions to be implemented must be adapted to local environmental conditions and human activities. Here, we take advantage of a fine resolution map (1:10000) of marine habitats to study the spatial configuration of Posidonia oceanica seagrass beds, a marine habitat of great ecological and economic importance. Six different composition and landscape descriptors were calculated at different resolutions (grid cells of 400 m x 400 m, 200 m x 200 m and 100 m x 100 m) between 0 and 40 m deep along 1700 km of French Mediterranean coastline (mainland and Corsica). A Random Forest approach was applied to relate these landscape descriptors to anthropogenic and environmental factors and to assess their relative importance. The best predictive power of the Random Forests models was obtained for 100 m x 100 m grid cells with models explaining 87% of the variance of the decline index and 70% of the variance of the cohesion index. The identification of threshold points for all environmental variables allowed to localize seagrass beds in either good or bad environmental conditions. We also identified sites whose spatial configuration is degraded despite good environmental conditions. These are sites with greater influence from human activities that could benefit from proactive conservation measures.