LINK TO NEW ALADIN WEB
Lampe ALADIN
 ALADIN Consortium
 ALADIN Documents
 ALADIN Model

The ALADIN Verification Project


Verification within the ALADIN Partners

The Aladin Verification Project (AVP) was started in the Aladin community in order to develop a common objective verification system at synoptic scale. The goal is to make a simple and reliable system to monitor time evolution and compare classical verification scores over different domains and versions. The development of the system has started in 2002 following the method proposed and accepted during the 12th ALADIN workshop: presentations and the conclusions of the working group on verification

Link to the prototype


Background of Verification
The verification issue is handled in Europe at several levels:
  • EWGLAM (with a co-ordination of DWD no longer active)
  • SRNWP (with a focal point by the Netherlands mainly operational for HIRLAM models)
  • and ALADIN with not real effect yet !
The 4th Assembly of ALADIN Partners (Lisbon, 1999) decided to endorse the following actions:
  • (1) the need for each NMS to nominate a contact point dedicated to verification : the e-mail list (verifala2++at++meteo.fr) contains the addresses of the nominated contact points;
  • (2) start a WMO like, i.e. decentralised, exchange of scores (but using the same definitions and same parameters as in the EWGLAM project);
  • (3) computation by everyone of scores for its geographical domain using the stations recommended by EWGLAM (with special attention of Marocco, not included in EWGLAM);
  • (4) scores sent back to all other contact points.
Up to end of 2001, only two ALADIN countries produce monthly scores according to the EWGLAM rules.
The ALADIN informal meeting (Cracow, 2001) decided to have a second attempt to realise an ALADIN verification : the e-mail list for verification (verifala2++at++meteo.fr) should be updated and ... used for the exchange of the scores ! It has been completed at the end of October 2001.

During the 6th Assembly of ALADIN Partners, vivid discussions were taken about verification matters of the ALADIN model. It was decided to have a stronger emphasis on verification.


Verification within SRNWP

As discussed during the 2001 EWGLAM/SRNWP meetings, ALADIN will participate to the exchange of scores within the SRNWP groups. Per Unden proposed to collect these scores. For ALADIN group, only ARPEGE, ALADIN-FRANCE and ALADIN-LACE are concerned by these SRNWP scores (not the other individual ALADINs).
Guidelines for the computation of the SRNWP scores and examples of results (for HIRLAM only) can be found on the HIRLAM website : http://www.hirlam.org/open/publications/NewsLetters/, on "HIRLAM Verification scores, 1st Quarter 2001" (pdf document).

The method is more or less the same as previously defined for EWGLAM scores by DWD, with a selection of parameters, ranges and periods to be compared. Corresponding scores are (ARPEGE, ALADIN-FRANCE) or will be (ALADIN-LACE) produced and sent to Per Unden.