EXECUTIVE SUMMARY

 A major scientific thrust for the remaining years of WOCE is to assess which ocean processes are important for climate prediction. To have predictive value, models must accurately describe the present ocean state. In particular, which processes below the surface layer must be observed and modelled and at what minimum resolution in space/time to increase predictive skill?

 Three overlapping activities are recognised in WOCE AIMS: analysis (which includes the synthesis of data with simple ocean models), model development (which includes the comparison of models and data) and assimilation (which combines data with models directly). These activities all presuppose the existence of a comprehensive WOCE data set, which results from the assembly of quality-controlled data from disparate sources, its documentation, the ready and widespread access of this data set and ultimately its secure archival.

 DATA

 The smooth and continuous functioning of the data system is necessary through the end of WOCE. Although the observational phase of WOCE ends in 1997, many data sets will still be processed after that time. Some sources will be producing WOCE data through 2002 (e.g., drifters and floats), and some data streams will be continuing and in the process of transition to other climate programs (e.g. XBT's, surface fluxes, etc.). Assembly, quality control, documentation and distribution of all of these data sets will continue to be required, necessitating full support for the Data Assembly Centres, Data Information Unit and archives.

 Two critical elements exist in the WOCE data stream, the supplier and the data centre: at the latter the data is subject to review and evaluation, before being made publicly available. Most WOCE Principal Investigators have been willing to submit their data to the centres, especially once it was made clear that the data was made public only after their permission was given. This report identifies some areas where data flow to the data centres is unsatisfactory and must be improved; namely, ADCP measurements and tracer data and bathymetry.

 In the matter of data distribution, the onus is put on the centres to stay at the forefront of changes in computing, distribution and storage, to create links to useful data products which exist in the community, and to distribute their holdings periodically under the guidance of the Data Products Committee. The data centres are charged to acquire older, high-quality data alongside WOCE data, thereby allowing the estimation of both the nature and extent of oceanic climate change in the case of hydrography, and more complete description of the ocean in the case of direct velocity data. All centres are also requested to provide assessment of data errors and their moments. The WOCE Data Information Unit is commended for its continuously updated summary of the status of all WOCE data and is tasked to continue during WOCE AIMS.

 The role of the data centres in bringing together the 13 data streams is emphasised in the report and the benefits of a single interface to all types of quality controlled data, their error fields, their quality flags and associated metadata (creating the WOCE data resource) are described. Such an integrated data set is especially important within WOCE for model testing and the developing data assimilation effort. Its success relies on the ongoing efforts of the data centres; without dedicated agency support for the data centres, the free access of data to the oceanographic community will wither and the ultimate benefit of the investment in the programme will remain unrealised.

 ANALYSIS

 Analysis activities will include the construction of data climatologies and atlases, an assessment of the representative nature of these climatologies; and a determination of the mean circulation of the global ocean, of its inherent variability and variability resulting from changing forcing. These, along with the quantification of ventilation and the renewal rates of water masses, in both of which tracers have a key role, and the determination of the oceanic transport of heat, freshwater and tracers, their divergences, and a description of ocean variability during the WOCE period - all have the highest priority, as was recognised in the enunciation of the goals of the programme. An important task will be to establish the mutual consistency of air-sea exchange and ocean flux divergences. The Meteorology Data Centre has been tasked to evaluate various flux products for such researches and also to advise ocean modellers as to the most appropriate products for their work.

 Most of these researches will be carried out initially using data and simple models, such as inversion techniques, but as the AIMS phase matures, increasing use will be made of prognostic ocean models and models with data assimilation. The more data-based results will serve as benchmarks against which the more sophisticated techniques will be evaluated. WOCE goals also target the understanding of the dynamical balances in different areas of the ocean. The report argues that this will result in part from the understanding of processes revealed through the description of the circulation as well as through process experiments such as the subduction, tracer release, and deep basin experiments. Tests of dynamical balances will be made from models too, once they are deemed to be sufficiently realistic and are accurately forced.

 MODELS

 Model development has benefited during the WOCE programme from the breathtaking growth in the speed and power of (super)computers, and from continued development of parameterizations essential to making modelled processes more realistic. In WOCE, ocean-only models will continue to be developed on two fronts, coarse and high resolution. A working definition of the first is one that can be integrated for (model) centuries or can be combined with an atmospheric model in coupled mode; high resolution models resolve the smallest scale consistent with the computer power and storage available. In broad terms, the task for WOCE AIMS described in this report is the continued testing of established and new ocean models, the further development and evaluation of high resolution models, and new parameterisations developed for coarse models and evaluated from the high resolution results. Close co-operation between WOCE and CLIVAR modellers is encouraged. While CLIVAR emphasises coupled models, WOCE will continue to provide the expertise and development of ocean models, including thorough testing of the coarse resolution ocean models used for climate prediction.

 Evaluation of models must take place both through thorough statistical tests of model runs and comparison of models with each other, and also through thorough comparison with data sets.

 The report highlights the need for alternative approximations to the fluid mechanics and to small scale parameterisations, perhaps resulting from interaction with other areas of continuum dynamics. Also desirable are improvements to domain geometry and numerical algorithms and better forcing, and boundary and initial conditions. The connection between model resolution and unresolved processes, and the methods for evaluating models from WOCE data are significant issues too. For coarse resolution models new parameterisations must be sought for isopycnal and diapycnal mixing, and for the effects of topography including narrow straits. The report urges the encouragement of young scientists to tackle these problems through post-doc and junior faculty positions, and warns of the need for heavy computing commitment by national agencies. Some tasks require a collective effort which exceeds the capability of small groups. Here the US has taken a step forward in setting up a cross-project Community Consortium for Ocean Modelling and other National Committees are invited to examine it and consider if similar action is appropriate within national or neighbouring boundaries.

 ASSIMILATION

 It is widely agreed that the most complete analysis of the WOCE data, revisiting all of the tasks described in the Analysis summary, will result from the combination of the data with advanced numerical models, or assimilation. This technique brings together data and models to yield a dynamically consistent estimation of ocean fields, of the errors of these fields, and even of certain model parameters (e.g., diapycnal diffusion). A set of steps of increasing complexity is outlined in the report which match the expected and necessary growth in computer power along with the development of tools and experience in dealing with the practical and scientific problems.

 Assimilation is already underway in a number of laboratories and universities, currently making use of the satellite altimeter and upper ocean thermal data streams. Early techniques are relatively crude but the problems of more rigorous methods coupled with the full range of data types are formidable. The computing resources projected for such a task along with major data handling needs has led to the concept of Assimilation Centres (see below).

 Successful assimilation of WOCE data will eventually be used for assessing the realism of ocean models and will also assist in the ongoing design of a climate observing system. Considerable experience is needed in this oceanographically-underdeveloped area to meet these expectations. As in other areas of ocean modelling, resources of manpower are clearly lacking, a deficiency which must be met by training programmes, as well as by recruitment. The report emphasises the need for increased skilled manpower in the field of assimilation, and hence its implementation would offer clear career opportunities.

 COLLECTIVE

 To a large extent the activities listed above will be carried out by individuals and small groups of researchers in academic centres and government laboratories in the 25 or so countries participating in the WOCE programme. Some tasks involve co-ordination of effort and resources; these include:

 * The formation of Assimilation Centres, consisting of 10-30 persons and massive super computer support, required for the most ambitious assimilation procedures: steps have been taken along this path within France and the US.

 * The creation of consortia for ocean modelling to take best advantage of large computing resources and the intellectual contributions of numerous small groups.

 * The development of a single interface(s) to all of the WOCE data sets, thus building on the efforts of all of the WOCE DACs, SACs and DIU.

 * International meetings of wide-ranging nature, all to promote WOCE AIMS activities, stimulate collaboration and disseminate results:-

 Regional Science Workshops: those for the Pacific (August, 1996), South Atlantic (June, 1997), Southern Ocean (July, 1997) have already been held and that for the Indian Ocean (September, 1998) is planned, and for the North Atlantic (1999) is proposed;

 Global Workshop (2000 or 2001) to discuss global models, global analyses and processes common to all oceans or benefiting from regional and basin comparisons, is proposed;

 WOCE Conference, to celebrate the completion of the field phase and announce preliminary results (May, 1998) is planned;

 WOCE/CLIVAR Workshop on Ocean Modelling for Climate Studies (August, 1998) is planned;

 WOCE Tracer Workshop is planned for Fall 1998 or Spring 1999;

 WOCE/GODAE Workshop on Data assimilation is planned for Spring 199.

* Collective publications of WOCE results and achievements in the scientific literature: each of the regional science workshops is expected to yield a collection of articles which will be published in JGR (Oceans). Atlases of WOCE data and climatologies, both printed and on-line, static and interactive, will be produced. A series of monographs and review articles synthesising WOCE results will be solicited.

 WOCE publications, the advances in ocean modelling and assimilation, and the existence of the comprehensive data set, which will serve as a baseline for observations of climate change within the ocean, will be the legacy of WOCE. In addition the results of WOCE will be exploited by follow-on programmes such as CLIVAR, GCOS and GOOS.

 Return to Table of Contents