Understanding environmental predictor variables of species occurrence may contribute to conservation management. In this paper we test the use of a spatial binomial model, estimated with the combined INLA-SPDE method, to relate the probability of occurrence of the mussel Mytilus edulis in the subtidal part of the western Dutch Wadden Sea in the period 1992–2022, to a range of environmental and human-impact related variables. Salinity and orbital velocity appeared to be the most important driving variables, and maximum probability of occurrence was predicted at intermediate salinity levels around 22 PSU and low orbital velocity. Mussel occurrence was also lower in the shipping lanes that are regularly dredged. One of the hypotheses is that at lower salinity physiological stress occurs, but that at higher salinity levels predation limits the occurrence. The spatial structure of the unexplained variation is described by a Gaussian field, but it remained unclear what the type of underlying explanatory mechanisms could have been that some areas had much lower probability of occurrence than expected on the basis of environmental conditions. Further understanding of these observed patterns, for example by including temporal dynamics or experimentally testing settlement limitations, could benefit future decision making for conservation management.

Sea level rise, increased storminess, and changes in sediment supply due to nourishments are all expected to drive coarsening (i.e., ‘sandification’) of muddy coastal sediments in the decades to come. Since the composition of soft-bottom benthic communities is associated with the sediment grain-size and mud content, this may result in habitats becoming less suitable for some species, leading to species shifts. Species-sediment relations can help to predict how this foreseen sandification may affect benthic fauna. We explore and quantify the sandification-sensitivity of benthic communities, with a tidal basin in the Dutch Wadden Sea as a model system. We identify the species' sediment optima and tolerance ranges using non-linear quantile regression models, summarise preference and sensitivity at the community level, and determine the difference between optimal and realised sediment habitat. We find that sediment optima are taxon-specific and that most species in this area are sediment generalists. On community level, there is a difference between the preferred and realised sediment habitat. In many areas, the actual inhabited sediment is coarser and sandier than expected based on the preferences of the resident species. Future sandification of the area would further decrease sediment habitat suitability for benthic communities in these places. This detailed knowledge of area-specific sensitivity of benthos can be used to inform coastal management decisions.