Ensemble model = a model that is derived from a group of input models (called the ‘ensemble’).
Ensemble models create the average, mean, mode, median, spread, standard deviation and percentiles of parameters such as temperature, pressure, precipitation, etc.
Ensemble models attempt to statistically compensate for unknowns of input data, chaotic aspects of weather, and other conditions.
The ensemble group of input models on which these averages and other stats are based can be different models OR it can be based on a group of the same model’s input variations (called “perturbations”). The perturbations are used as the input models.
Creating the proper “perturbations” is a science in itself.
The GEFS is the ensemble model derived from the GFS and consists of 30 perturbations of the GFS plus itself!
The SREF (short range ensemble) is a model where the group are different short range models like the NAM, NAM-NEST.
The model blend (NBM) can also be thought of as an ensemble model comprised of many different models for its ensemble.
Since these models are composites of many models or many perturbations, they can take 6-7 + hours to be computed by the supercomputers used for numerical weather prediction.
« Back to Glossary Index