Newsletter No 30

¡La Puerta al Futuro ya está Abierta!

Version 4.1 Adopted

Peter Lynch

In defining HIRLAM Version 4.0 we bid a Grand Slam. A large number of new components, each of which had been separately and comprehensively tested, were combined in the new version. The results of the parallel tests were presented at the meeting in DMI in November. They were less than spectacular. The results with the new version (H40) were much worse than the reference, H27. A serious bug in H40 was (and is) suspected. Suspicion pointed to the new surface scheme, to the semi-Lagrangian scheme, or to some interaction between these. The mood at the Copenhagen meeting was subdued: Mark Twain's synposis of progress (see below) seems to catch the spirit.

To continue the Bridge metaphor, we failed to make the contract; but, as it was the first bid, we were non-vulnerable. A revised new model configuration was agreed for testing. This was designated Version 4.1 (H41), and the new components comprised a subset of those which had been included in Version 4.0. A suitable correction set for the new version was prepared at DMI and distributed in December. Parallel tests comparing H41 to the reference H27 were carried out and the results were presented and discussed at a meeting in FMI in late January.

The results of these tests, carried out in Madrid and Helsinki, were very much more encouraging. They are reported in two articles in this Newsletter, one by Simo Järvenoja (page 20), the other by José Antonio García-Moya and Isabel Martínez (page 29). The anomalous results produced with H40 were no longer evident. The overall skill of the new configuration is comparable to that of the reference, and it is clearly superior in a number of respects. In the light of these more encouraging results, it was agreed at the Helsinki meeting that the new version should be accepted as the reference. Following a discussion there, the following details of the new reference version were agreed:

  1. Holtslag vertical diffusion
  2. Mass flux convection scheme
  3. Eulerian advection (Delta t=240s)
  4. Implicit 4th-order horizontal diffusion
  5. Resolution: 0.5 degrees, 31 levels
The surface parameter analysis has been temporarily disabled; its reinstatement will be given top priority. It was also agreed that the alternative cloud scheme STRACO (see HIRLAM Newsletter 29, page 37) should be included as an option in the new version. At the time of writing (16 February) the installation of the new reference version on the Fujitsu is nearing completion. Release notes on Version 4.1.0 have been prepared by Gerard Cats and are available on the HIRLAM Extranet.

The new Reference Model should provide the basis for testing of future refinements and enhancements. Intercomparison of results from different centres is greatly facilitated if tests are carried out using the same control model. There are also some clearly identified weaknesses in the new system, but we can be optimistic about early progress in removing them. The articles following this give more details of problems and possible solutions. The enhancement of Version 4 is now the focus of our work. We are in a good position to make progress. The gateway to the future is now open --- ¡La puerta al futuro ya está abierta! 

             Researchers have already cast much darkness on the subject,
             and if they continue their investigations we shall soon 
             know nothing at all about it.

                                                           ---Mark Twain