Super-ensemble techniques applied to wave forecast: performance and limitations

Abstract

Abstract. Nowadays, several operational forecasts of surface gravity waves are available for a same region through different forecasting systems. However, their results may considerably diverge, and choosing one single forecasting system among them is not an easy task. A recently developed approach consists in merging different forecasts and past observations into a single multi-model prediction system, called the super-ensemble. First implemented in meteorology, the method has also already been tested with success in oceanography for determining temperature, acoustic properties or surface drift. During the DART06 campaigns organized by the NURC, four wave forecasting systems were simultaneously run in the Adriatic Sea, while significant wave height was measured at six stations and along the tracks of two remote sensors, hence offering an opportunity to evaluate the skills of the super-ensemble techniques. The improvement shown during both the learning and testing periods was essentially due to a bias reduction, though the correlation was also increased. The possibility of extrapolating locally obtained results in the whole domain of interest was assessed against satellite observations. Though definitive conclusions can not be drawn from these experiments, the results open the door for further investigations.