Since the early 1990s, ensemble methods have been increasingly used to address predictability issues, and provide estimates of forecast uncertainties, e.g., in the form of a range of forecast scenarii or of probabilities of occurrence of weather events. Although there is an overall agreement on the main objectives of ensemble-based, probabilistic prediction, different methods have been followed to develop ensemble systems. In this chapter, we will review the methods followed at the major weather prediction centres to develop global ensembles, and we will highlight the links between the method followed and the ensemble forecast performance. The material presented in this chapter is based on the operational, global, medium-range ensembles operational at the time of writing (2014).
Authors who have authored or contributed to this publication.