ALADIN developments in Toulouse during the forth quarter of 1999
Toulouse ALADIN visitors... the return !
As usual, the number of Toulouse visitors increased during
the last quarter of 1999. The growth was all the more obvious since there were
very few visitors during the last quarters. However, this Toulouse visitors
participation during the forth quarter of 1999 is the smaller one within the
participations of the forth quarters since 1993. But, do not worry, things
will change in 2000. The first quarter of 2000 is reasonably crowded in
Toulouse and the second quarter is already overbooked !... and ALATNET has not
begun yet ...
1. Operational applications ...
- No major problem was encountered in operational ALADIN suites when
entering year 2000, only some GRIB coding details in Prague.
2. Main events in Toulouse this quarter
- The ALADIN effort in Toulouse this quarter was mainly devoted to the
preparation of a new cycle, AL12, phased with the cycle 22T1 of ARPEGE. The
Toulouse team was helped by Radi Ajjaji, Martin Bellus, Lora Gaytandjeva,
Ilian Gospodinov, Hassan Haddouch, Jure Jerman and Gabor Radnoti. Apart from
the developments issued from AL11T2, this new library will allow an improved
management of spectral transforms, the use of simplified physics, a new
formulation of envelope orography, the use of CANARI in diagnostic mode at
high resolution, the identification of lakes, the computation of the TL and AD
parts of non-hydrostatic dynamics, ... and some cleaning, with some obsolete
options suppressed.
- As you can see, ALADIN life went off quite well in Toulouse but, outside,
the last quarter of 1999 was not so quiet : strong storms over North Sea and
Baltic Sea, snow events over Central Europe, dramatic floods over Southern
France (12-13/11/99), and the two french Christmas storms (25-28/12/99) were
as many challenges for NWP. ARPEGE behave rather well, but unexpectedly ALADIN
was not so good over some situations. For example, ALADIN/LACE missed the
first Christmas storm (26/12/99) while ALADIN/FRANCE was not better than
ARPEGE. Some experiments were performed afterwards by Ryad El Khatib and
Radmila Bubnova, to understand the reasons of this failure. The main impact
came from coupling : results improve as the frequency of coupling, from a 6h
span (operational ALADIN/LACE) to a 3h span (operational ALADIN/FRANCE) and
further with a 1h span. Using a quadratic rather a linear time-interpolation
also led to better results. A quadratic 3h-coupling proved better than a
linear 1h-coupling. On the other hand, changing the time-step or using
non-hydrostatic dynamics brought only minor changes.
3. Other visitors research or development studies that ended during this quarter
- Jan Masek has thoroughly analysed the discretization of equations in the
non-hydrostatic case, and justified some initial choices.
- André Simon studied the impact of the new envelope orography and of
the lift parameterization in ALADIN / France and ALADIN / LACE on a wide set
of situations.
4. Mixed Toulouse-deported studies
- July 97's floods
- In the framework of the Barrande exchanges between Toulouse and Prague,
Elisabeth Gérard and Dagmar Dufkova studied the behaviour of the two
last operational versions of ARPEGE/ALADIN (i.e. cycle AL11-CY21T1 and the
CYCORA modset) on the dramatic floods of July 1997 over Central Europe. For
each configuration, they launched a 3-weeks global assimilation suite starting
on the 3rd of July and 2-days forecasts with increasing resolution, using
ARPEGE (21 km), ALADIN/LACE (12 km) and ALADIN/CZ (8 km) successively. The
detailed comparison of fields to observations will be pursued in Prague.
However the first maps show a clear improvement from the former ALADIN/LACE
forecast to ALADIN/CZ using CYCORA, with a significant increase in precipitations.
- Using high resolution data for clim files
- The Balaton (Toulouse-Budapest) exchanges benefited to configuration 923.
A new module (the 7th) was written by Sandor Kertesz. It allows a modification
of the characteristics of water points (type, temperature, depth) and an
identification of lakes using local high resolution data. It is the
counterpart of module 5 for land points. In the meantime both modules were
modified by Dominique Giard in order to better preserve sharp gradients from
initial data. The management of sea domains was also corrected.
5. Other research or development studies by the Toulouse permanent staff
- MAKDO, a graphical tool for the creation of ALADIN domains, is born. In
the meantime, Jean-Daniel Gril wrote a new version of EGGX, simpler,
preserving namelist directives, and avoiding some of the bugs discovered in
the previous one.
