Clustering
In order to compress the amount of information being produced by the ENS, and to highlight the most predictable parts, individual ensemble members that are "similar" according to some measure (or norm) can be grouped together. This process is known as Clustering. The norm for measuring which members are “similar” can be defined in different ways. Clustering can be performed over different geographical areas or for different parameters, and can be for each forecast time or over different forecast intervals. Clustering is a compromise between the advantage of condensing information against the disadvantage of losing possibly important information. A convenient overview of the ENS forecast information can be gained from using Cluster Means (the average over all members in a cluster) or by a Most Representative Member which typifies the members within the cluster.
In the current medium-range ENS forecasts, the clustering summarises the range of synoptic flow patterns (weather scenario clustering) in a restricted area covering Europe and the northwest Atlantic. Each cluster is represented by the ensemble member closest to its centre, referred to as the Most Representative Member (MRM or cluster scenario) for that cluster. Each MRM is then attributed to one of the four climatological regimes (weather regime clustering) which have been derived from reanalysed data. This shows the differences between scenarios in terms of the large-scale flow and provides information about the possible transitions between regimes during the forecast.
In the current extended-range and seasonal forecasts, the "weather regime" clustering is used from the outset, although the weather regimes that are used differ from those used with ENS medium range weather regime clustering.
Overview
Community Forums
Content Tools