Seasonal prediction

The group develops forecasting capabilities on seasonal timescales.

The main prediction system currently used for seasonal timescales is GloSea5. It is an ensemble prediction system using a coupled ocean-atmosphere model (variant of HadGEM3), to generate probabilistic forecasts up to six months ahead. GloSea5 includes uncertainties due to initial conditions and model formulation.

In a major change from the previous seasonal forecasting system, GloSea5 is fully integrated in the model-development cycle. This is expected to lead to faster improvements in forecasting skill for regional, seasonal scales.

The output from the system is used to generate products.

Key aims

  • To develop an operational forecasting system.

  • To contribute to model development.

  • To generate products required by World Meteorological Organization, from its global producing centres of long-range forecasts (of which the Met Office is one).

Current research

  • Model initialisation, including ocean data assimilation

  • Representation of model uncertainties

  • Calibration of model output and verification

Last updated: 16 August 2013