4.4 Impact of the Representation of the Stratosphere on Tropospheric Weather Forecasts in the Met Office Unified Model

Tuesday, 18 June 2013: 4:00 PM
Viking Salons DE (The Hotel Viking)
David Jackson, MET OFFICE, Exeter, United Kingdom; and S. Mahmood, J. C. H. Cheung, J. D. Haigh, and A. Charlton-Perez

This study investigates the importance of the representation of the stratosphere for tropospheric weather forecasts using two approaches. The first is to look at three different vertical model resolutions available in the Unified Model (UM) forecast system at the Met Office. One vertical resolution is a “low top” (with an upper boundary near 40 km) and the other two are “high top”, with an upper boundary near 85 km but with different resolutions in the middle atmosphere. The second is to examine the impact of changing the UM ozone field from a climatology to one based on observations (valid at the forecast start time) from the Earth Observing System Microwave Limb Sounder (EOS MLS).

Ensemble forecasts at the three vertical resolutions have been run around the February 2010 Stratospheric Sudden Warming. Innovative methods of diagnosing Rossby wave breaking in the troposphere have been developed, with the aim of furthering our understanding of the stratosphere-troposphere link and its impacts. Results of the low top and high top ensemble experiments will be discussed and the mechanisms by which the stratosphere may be influencing the troposphere investigated through these experiments. The information gained will inform future decisions regarding the optimum representation of the stratosphere in Global Circulation Models, and will contribute towards the selection of future numerical weather prediction forecast setups at the Met Office.

Experiments with different UM ozone representations have been run for a number of periods. We focus on March 2011, where the cold Arctic vortex persisted for unusually long and where there was a particularly large difference between EOS MLS ozone and climatological values. Our results show that tropospheric forecast errors in the medium-extended range are dominated by the spread of ensemble members and hence no significant reduction of root mean square errors due to improved representation of ozone is seen in the troposphere.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner