Cluster Parameters

The ECMWF offers additional information on their ensemble forecasts describing the behaviour of the individual ensemble forecast members with respect to each other. This comparison happens by grouping similar members into clusters.
The information offered can be grouped into two categories:

Information about the whole set of clusters

Number of Clusters

The number of clusters depends on the variability of the model and thus can vary over the course of a model run. Hence the number of clusters at a specific time step is of interest:
number_of_clusters:x 
Example:

Control Cluster

Number of the cluster containing the control run.
control_cluster:idx

Deterministic Clusters

Number of the cluster containing the deterministic run.
deterministic_cluster:idx 
Example:

Information about specific clusters

Please not that cluster-specific parameters need a cluster_select parameter to specify the cluster for which to display the information. The following information can be queried:

Ensemble Members

Number of ensemble members grouped into the specific cluster.
number_of_forecasts_in_cluster:x

Representative Ensemble Members

Returns the id of the ensemble member of the selected cluster that is considered representative for the cluster
representative_member:idx 

Climatological Regime

Climatological weather regimes are computed for cold (October to April) and warm (May to September) seasons. Four patterns, consistent with those documented in the literature (see for example Michelangeli et al. 1995) are considered:
  1. a strong positive anomaly over the Greenland area, together with a trough elongating from the southwest to the northeast Atlantic Ocean
  2. an east-west dipole, [...] associated with a pronounced wave structure
  3. a trough centered over Iceland together with a small-amplitude positive anomaly in southern part of the domain
  4. a ridge in mid-Atlantic
climatological_regime:idx

Start Time

First time step of the clusters existence (in hours since model start)
start_time_step:h 

End Time

Last time step of the clusters existence (in hours since model start)
end_time_step:h

Geopotential Height Cluster

Since the grouping of the ensemble members into clusters is based on the geopotential height at 500hPa, that parameter can also be queried for each cluster
geopotential_height_500hPa:m 
Example: