During the last Assembly of ALADIN Partners, some amendments to the ALADIN MoU were proposed. All minor amendments were approved by the Assembly. Others were discussed and the directors chose between a few possible options. Main accepted modifications are :
- regarding the access to ALADIN code for institutions which were neither ALADIN participating Partners, nor from an ECMWF Member, nor from an ECMWF Co-operating country, it was decided : no rule of exception but regulation of new problems by the amended mechanism for full adhesion;
- the case of code benchmarks was specifically added to the non-dissemination and restricted-use rules;
- regarding the "Commitment and monitoring of manpower associated to the ALADIN Project" : a third category of work for reporting ("Prague-LACE" in addition to "Toulouse" and "at Home") and applying of strict deadline and using single contact points were agreed upon.
The modified
version of the MoU is currently visiting all ALADIN countries to be
signed. Countries are visited in an alphabetic order which is certainly not
the order I would have chosen in case I would have the opportunity to travel
with it.
This Agreement has been approved by ECMWF Council and is expecting to be signed any day now.
The work of the ALADIN community is officially recognized by ECMWF who considers "the advantage as IFS/ARPEGE partners to profit from the ALADIN community's contributions on outstanding NWP issues that may be beneficial both for synoptic- and meso-scales".
This Agreement is linked to the ALADIN MoU since it regulates an access to those countries which are neither full nor associated Members of ECMWF to the ECMWF-born parts of the ALADIN software.
A draft of this agreement was presented during the last Assembly of Partners who decided to declare the agreement of ALADIN Partners with the presented document.
This document has been approved (for its purely scientific part and accordingly edited) at the third Assembly of Partners.
It was the second exercise of this type in the history of ALADIN, both plans coinciding with important events in the life of the project : the quasi-operational start of ALADIN-PECO in 1994 (first plan), and now the operational or pre-operational status of ALADIN by most of the NMS involved in the project.
This new plan has been constructed starting from a few basic assumptions :
- sufficient means will be allocated to cover strong burden of code maintenance resulting, for the IFS/ARPEGE/ALADIN software, from a more and more distributed research effort;
- the different operational applications of ALADIN will not generate any unnecessary diverging tendencies inside the overall ALADIN community for what concerns the links between research and operations.
Three main topics are discussed in the plan :
- operations : the code maintenance is a key for all operational versions but it is equally important for the common research and development work. Co-ordinators have been proposed for dynamics, physics and observation handling topics and orientations are defined.
- applications : this part of the document deals with the application of the model itself and its products for further use and for other special applications. Two topics and their co-ordinators have been proposed : modelling problems and production related problems.
- investment : in this part, orientations for the research for tomorrow's numerical weather prediction are given once divided in two main topics : data assimilation and high resolution dynamics.
Please consult the complete document for details.
The next IFS/ARPEGE cycle 21 is an automatic processing of cycle 20 introducing a strong typing (through the IMPLICIT NONE usage) and the FORTRAN 90 free format. This automatic processing will introduce the definition of KIND parameters, dedicated macros for declaration of integers and reals, and macros to replace superstar real constants.
Free format, which means that there is no more specific positions on a FORTRAN line, should drive us to more flexibility and legibility of the source code.
See more details in annex in this Newsletter.
On November the 19th, the second coordinate change of the clim files used by all ARPEGE and ALADIN operational suites was successfully completed. It allowed a better description of vegetation, especially over Eastern and Northern Europe.
A major bug in ALADIN 09 has been discovered by the Prague/LACE team. It concerned AL09/CY19T1 : the bug was active in the forecast mode in the case you had more than 7 coupling files. The 8th file was the fatal one : it was not read and the previous one was kept. For example, for ALADIN/FRANCE that is coupled every 3 hours, the 8th file was valid for +21h. Coupling was realized with +15h instead of +21h. Results were false at +21h. This error had still an influence at +24h even though the coupling was right at that range. The coupling sequence was : ... +15h, +18h, +15h, +24h ... . For ALADIN/LACE that is coupled every 6 hours, the fatal file was for +42h.
The regular examination of ALADIN-France scores (realized only every 6 hour !...) have not shown any spectacular break corresponding to the bug.
The problem has been corrected. The ALADIN09 export version number 5 is OK. In case you use ALADIN09, please make sure that every coupling files has been used (check in the job listings).
Last December, ALADIN source code has been phased with a preliminary ARPEGE library preCY20. Phasers are now validating CY20/AL10 code... and they are doing their best to avoid such a big bug as the one introduced in cycle 9. This will be completed by the end of February. Validations on CRAY and on workstation (SUN) will be performed afterwards in Belgium and in France respectively.
Then, a cleaning cycle (CY21/AL11) will be performed with free-source format and IMPLICIT NONE statement. Please contact us in case of bad experience with local compiler and free-source format.
Marta Janiskova successfully defended her PhD on the 9th of November 1998 in Brastislava.
