18th Conference on Weather and Forecasting, 14th Conference on Numerical Weather Prediction, and Ninth Conference on Mesoscale Processes

Monday, 30 July 2001: 1:20 PM
Recent developments of the Met Office 3D-Var system
N. Bruce Ingleby, Met Office, Bracknell, Berks., United Kingdom; and S. P. Ballard, R. S. Bell, A. Berney, B. Candy, S. English, D. Li, A. C. Lorenc, R. J. Renshaw, and J. K. Ridley
Poster PDF (46.8 kB)
The Met Office 3D-Var system was implemented operationally in the main global forecast suite in March 1999. It was implemented in the mesoscale (local area) suite in October 1999 and in the troposphere-stratosphere suite in November 2000. This presentation will give an overview of the global system.

Use of NOAA-15 ATOVS data (via a 1D-Var retrieval system) was also implemented in March 1999, since then we have moved to direct use of (A)TOVS radiances, use of some tropospheric channels over Siberia and shortly we will begin assimilating ATOVS moisture (AMSU-B) channels and data from the second ATOVS satellite NOAA-16.

Various other changes to observation usage have taken or will take place: use of retrieved SSM/I surface wind speeds, reintroduction of scatterometer surface winds with dealiasing within 3D-Var, improvements to use of surface pressure reports, introduction of wind profiler data and use of 'first-guess at appropriate time'.

There have been changes to the representation of background error covariances - to reduce vertical correlation scales slightly and to increase horizontal scales, particularly in the stratosphere. Minor changes to humidity error covariances will be implemented shortly.

Since the introduction of 3D-Var there have been significant improvements to forecast verification, particularly for winds, the southern hemisphere and at medium range. The (A)TOVS and background error covariance changes have played a major role in this - supported by the other observational and forecast model changes. Future developments will be briefly outlined.

Supplementary URL: