Laminaria hyperborea kelp is modelled for the Sandoy area at the More and Romsdal coast on the West coast of Norway (Bekkby et al. 2009). The model was carried out in the projection UTM zone 32N. Kelp forest was defined as the dense kelp forest (see Bekkby et al. 2009), not the scattered occurrences. The model was developed based on 384 data points and GAM analyses of presence and absence points (presence being kelp regardless of density). Data were collected 2008 and the model was run by NIVA in 2009. The distribution model is based on depth, slope, terrain curvature, wave exposure and a light exposure index (Bekkby et al. 2009), depth being the most important variable, followed by terrain curvature and wave exposure (see Bekkby et al. 2009 for explanation of the GIS environmental layers). The input depth, slope, terrain curvature, wave exposure and light exposure index models had a spatial (horizontal) distribution of 10 m, the current speed model was resampled from 25 m resolution. The output model has a spatial (horizontal) distribution of 10 m. Analyses showed that coverage (density of kelp defined as classes) increased with predicted probability. The work was funded by the Research Council of Norway.