Kelp distribution model across Brittany, France

A habitat distribution modelling approach was used to model the spatial distribution of kelp forests along BritannyÔÇÖs coast (France). Lineage: Data represents percentage cover of kelp forest. The Habitat model was generated from kelp forest presence/absence determined from acoustic surveys of laminarial algae. KdPar, sea surface temperature, depth and current speed were used as predictive variables. Habitat model represents only the areas where kelp forests may occur, not the areas confirmed in field observations. Type of occurrence data used: Acoustic surveys of kelp forest were carried out in spring 2005-2006-2007 in 10 study areas along the coast of Brittany. A specific algorithm was developed for automatic detection of presence/absence of kelp forest from the acoustic data. Environmental covariates/explanatory variables: Covariates used were: - KPAR (proxy of light availability): spatial resolution 1km, averaged values (temporal coverage 7 years -1998-2004-, temporal resolution 1 week) - Sea surface temperature: spatial resolution 1km, averaged values (temporal coverage 20 years, temporal resolution 1 week) - Tidal current velocity: spatial resolution 300m, maximum values (temporal coverage one mean spring tide) Algorithm/modelling approach: A stepwise multiple regression with a backward selection of variables was used to predict values of kelp frequency (%)For EMODNet Seabed habitats portal, the dataset was scaled between 0-1. Original data is 0-100.