Presentation of the 2002-2004 research plan

The chairman thanked Dominique Giard for her paper concerning the Third Medium Term Research Plan (2002-2004) that had been received by each partner one week before the meeting. He thanked also all those who had taken part in the working groups and spent their precious efforts to help in providing this interesting document to the Assembly. The presentation covered the following individual topics : Training, Maintenance, Operations, Applications, Verification, Coupling, Dynamics, Physics, Data assimilation (Methods, Observations, Surface), Predictability and Organisation of research. The discussion aimed especially the verification point. All the partners found it abnormal that verification, important component for both research and operations, is still at the same level than in 1999 : no improvement had been done since this date. The several reasons behind this fact are :

1. Lack of manpower dedicated to this subject.
   
2. Portability problems between partners : absence of unified procedure.
   
3. No local databases implemented.
   
4. No data exchange,
   
5. No parallel suites for the model (this could push people to take care of observations).
   
6. Considering this subject as local and specific to each country.
   

The possible solutions suggested by the different participants could be :

1. To consider verification as a great part of the project.
   
2. To encourage people to build local observation data bases.
   
3. To exchange results (and if possible data and procedures): use of "verifala" mailing list.
   
4. Verification must be done by an independant team at each level.
   
5. New approach of community organisation based on a reasonable coordination scheme.
   

The chairman decided to suspend this discussion and resume it during the AOB item of the agenda.

About the other topics of this futur research plan, M. Mersich mentioned the (too ?) high number of topics with a high "priority". M. Beysson was surprised to see that a high priority was not assigned to satellite observations. Jean-François Geleyn answered that variational data assimilation as implemented for ARPEGE/IFS could not be adapted to ALADIN LAM model without taking into account some very important LAM specificities related to coupling, high resolution, frequency of data, ...etc. He added that for the time being, making an assimilation in ALADIN with the context and observations as in ARPEGE is equivalent to redoing an ARPEGE analysis on a limited area. Dominique Giard underlined that the definition of "priorities" included the time-range necessary to reach them. For example, IASI data, to be really useful within ALADIN, were to be used over land, not only over sea as is done in ARPEGE. This requires significant developments and thus cannot be considered as a "high priority/short term" action.

The chairman suggested to distinguish between the so-called "transversal priorities" and the other less general priorities ; this had been taken into account later by the creation of a new working group (cf. 7.2) having as a charge to take care of this kind of management tasks.