Ensemble forecasts are a computationally feasible method to predict

forecast uncertainties depending on the meteorological situation of
the day. The individual forecasts in an ensemble differ due to the
representations of initial uncertainties and model uncertainties.
At the European Centre for Medium-Range Weather Forecasts
(ECMWF), model uncertainties are represented stochastically. The talk
will begin with an overview of the current methodology of stochastic
model uncertainty representations used at ECMWF.  Recent progress and
challenges associated with the representations will be described.

Then, future directions for the representation of model uncertainties
in ECMWF ensembles will be summarised.  The coming years are likely to
see a further increase in the use of ensemble methods in forecasts and
assimilation. This will put increasing demands on the methods used to
perturb the forecast model. An area that is receiving greater
attention than 5 to 10 years ago is the physical consistency of the
perturbations.  Other areas where future efforts will be directed are
the expansion of uncertainty representations to the dynamical core and
to other components of the Earth system as well as the overall
computational efficiency of representing model uncertainty.