Statistical multi-model climate projections of surface ocean waves in Europe

•A statistical methodology for regional projections of wave climate was developed.•The methodology is based on a weather type classification.•The low computational requirements of the method allow a multi-model approach.•Changes in several wave parameters under different climate scenarios are estimated and analyzed.•Results indicate mean wave height and mean period reductions in the European coasts throughout the twenty-first century. In recent years, the impact of climate change on sea surface waves has received increasingly more attention by the climate community. Indeed, ocean waves reaching the coast play an important role in several processes concerning coastal communities, such as inundation and erosion. However, regional downscaling at the high spatial resolution necessary for coastal studies has received less attention. Here, we present a novel framework for regional wave climate projections and its application in the European region. Changes in the wave dynamics under different scenarios in the Northeast Atlantic Ocean and the Mediterranean are analyzed. The multi-model projection methodology is based on a statistical downscaling approach. The statistical relation between the predictor (atmospheric conditions) and the predictand (multivariate wave climate) is based on a weather type (WT) classification. This atmospheric classification is developed by applying the k-means clustering technique over historical offshore sea level pressure (SLP) fields. Each WT is linked to sea wave conditions from a wave hindcast. This link is developed by associating atmospheric conditions from reanalysis with multivariate local waves. This predictor–predictand relationship is applied to the daily SLP fields from global climate models (GCMs) in order to project future changes in regional wave conditions. The GCMs used in the multi-model projection are selected according to skill criteria. The application of this framework uses CMIP5-based wave climate projections in Europe. The low computational requirements of the statistical approach allow a large number of GCMs and climate change scenarios to be studied. Consistent with previous works on global wave climate projections, the estimated changes from the regional wave climate projections show a general decrease in wave heights and periods in the Atlantic Europe for the late twenty-first century. The regional projections, however, allow a more detailed spatial characterization of the projected changes under different c