Identifying soft sediments at sea using GIS-modelled predictor variables and Sediment Profile Image (SPI) measured response variables

Macrofauna composition and diversity in soft sediments are commonly used as ‘‘health indicators’’ in various pollution monitoring programmes worldwide. Hence, finding a modelling method for predicting the presence of soft sediments and enable production of digital maps of where soft sediments are likely to be found would be valuable for developing a cost-effective sampling design. This study presents a first-generation model that can predict where to find soft sediments in coastal areas with a complex topography and a mosaic of seabed habitat types. We used geophysical data that were quantitative, objectively defined (through GIS modelling) and integrated over time. We analysed, using a Generalised Additive Model (GAM) and the model-selection approach Akaike Information Criterion (AIC), the influence of in-situ measured depth and GIS-modelled terrain structures, wave exposure and current speed on the distribution of soft sediment measured using a Sediment Profile Image (SPI) camera. Our analyses showed that the probability of finding soft sediment was best determined by depth, terrain curvature and median current speed at the seafloor. These predictors were used to develop a spatial predictive GIS-model/-map (for parts of Skagerrak, Norway, with a spatial resolution of 25 m 25 m) of the probability of soft seabed occurrence.