Coupled Ocean-Atmosphere Modelling

The UGAMP coupled ocean-atmosphere modelling effort has made tremendous progress since the Spring 1996 newsletter. The spin-up run of the tropical domain (29S-29N) MOMA ocean model has been completed and is documented in UGAMP Internal Report 45. And our first full coupled GCM run is well under way.

The coupled run uses the tropical domain MOMA ocean GCM coupled to the ECMWF IFS version 13r4 atmosphere GCM. The ocean GCM has roughly 1^ox 1^o resolution, increasing to 0.3^o in the north-south direction in the deep tropics, and 30 levels. The atmosphere GCM resolution is T42, 31 levels. The ocean state after 5 years of spin-up was used as the initial state for the coupled run. The first 33 months of the coupled run were carried out on the Cray C90 computer at ECMWF. When the C90 service was terminated in September 1996 the run was restarted on the Cray J90 at RAL. At the time of writing about 5 years of integration have been completed. One month of coupled integration requires roughly 1 CPU day and produces about 300 Mb of output!

Fig. 25 shows that two periods of dramatic cooling occur in the eastern equatorial Pacific during the first year of the run. Preliminary diagnostics suggest that these occur initially through a wind-driven upwelling signal propagating at the Kelvin wave speed. The zonal wind stress in the coupled model is initially significantly stronger than the climatological values used to drive the ocean spin-up, so that the initial coupled model state is "out of balance"; the dramatic cooling is part of the adjustment towards a more balanced state.

The figure also shows a systematic cooling of the equatorial SST at nearly all longitudes by about 4K over 4 years. Preliminary diagnostics suggest that, except in the eastern Pacific, this drift is due to excessive latent heat flux from the ocean to the atmosphere.

Despite the adjustment and drift, an annual cycle that exhibits many realistic features appears to be emerging during the third and fourth years of the run. In line with observations, the SST in the east Pacific is warmest around March, and the warm pattern propagates westwards accompanied by a decrease in the central Pacific westward wind stress. The model also captures the phase of the equatorial Atlantic SST annual cycle and the September maximum in wind stress in the western Atlantic, as well as the January maximum in wind stress off the east coast of Africa. However, some features of the model's annual cycle are unrealistic. The observed semiannual cycle in SST in the west Pacific is not captured, the September maximum in wind stress in the west Pacific is greatly overestimated, and there is unrealistically strong eastward wind stress in the Indian ocean during November-February.

In the coming year we will carry out detailed analyses of two specific aspects of the coupled integration. First is the process of adjustment and drift. This problem is of more than merely technical interest. Its investigation gives insight into the numerous interactions that maintain the mean climate in the coupled system. Furthermore better understanding of adjustment in the coupled system is essential for the development of seasonal forecasting. Initially we are examining the comparative importance of local thermodynamic processes and nonlocal dynamical processes for drift in SSTs. The second topic we will investigate is the role of coupled processes in the seasonal cycle in the east Pacific (see last FebruaryÕs newsletter). The fact that the coupled GCM reproduces the westward propagation in both SSTs and zonal windstress is a most encouraging stimulus for this work.

A considerable loss this year was the departure of Ian Udall. Ian made a tremendous contribution to the project through his work on the development of the ocean model, MOMA. Fortunately, he left this work essentially complete and Ñ what is more Ñ WELL DOCUMENTED! To replace Ian, we look forward to being joined shortly by Ben Edgington who for the last 18 months has worked on MPP projects at CGAM with Lois Steenman-Clark.

John Thuburn and Rowan Sutton
(Reading)

Adaptation of the Unified Model for Coupling via OASIS

The UKMO Unified Model (UM) can at present be run either in atmosphere only mode or linked to an ocean model as part of a coupled system. However, the design of the UM system is such that the ocean model is essentially hard wired. It cannot easily be replaced with a different ocean model. The facility to easily couple a variety of ocean models to the UM atmosphere would be very useful. For example, the UKMO are keen to compare, within the coupled system, the performance of an isopycnic ocean model with the performance of their current level model. In UGAMP, we wish to use a variety of simpler and limited domain ocean models to help elucidate the physical mechanisms by which ocean-atmosphere interactions influence the climate.

Recognizing the mutual benefit to be gained, we have set up a collaboration with the UKMO to adapt the Unified Model for coupling to arbitrary ocean models via OASIS. At the Hadley Centre Jean-Christophe Thil is working on the Fortran level changes to the UM, while at UGAMP Roger Brugge is working on the script level changes. A substantial part of the work has already been done, and we aim to start the first test integrations in February/March of this year. After thorough testing, the option of using OASIS will be included in the UM User Interface, and any user with a new ocean model should find coupling it to the UM atmosphere will be a straightforward task. Because the UM is currently undergoing major changes associated with porting onto the new Cray T3E, the option of using OASIS is unlikely to be available as standard until the end of the year, but test versions should be available sooner.

Rowan Sutton
(Reading)

Announcement of Town Meeting on COAPEC thematic programme proposal

Coupled Ocean Atmosphere Processes and their Effect on Climate (COAPEC) is the title of a proposal for new thematic programme to be submitted to NERC later this year. The proposal is being developed as a collaboration between members of the UK atmospheric and oceanic science communities in discussion with users and other interested parties.

A Town Meeting has been organised to discuss the content of the proposal. This meeting will be held at the University of Reading on 24 and 25 March 1997. Anyone wishing to attend should contact Anne Pinnock at the Centre for Global Atmospheric Modelling, Department of Meteorology, University of Reading (anne@met.reading.ac.uk). For further information about the proposal please contact Alan O'Neill, Julia Slingo or Rowan Sutton at CGAM. See also the WWW site
http://www.met.reading.ac.uk/coapec/coapec.html

Rowan Sutton
(Reading)

© 1997 Centre for Atmospheric Science/UGAMP. All scientific articles are unpublished. No text or graphics may be copied or used without permisson. Newsletter Editor: Glenn Carver, Cambridge University.