ALADIN developments in Prague during the second half of 2001

Note: all ALATNET related R & D, representing the majority of the effort, is reported in ALATNET Newsletter. Here we sum-up topics which are not referred as ALATNET ones for Prague centre.

Developments in physics

- Convective onset study

A behaviour of the deep-moist convection scheme in ALADIN was studied in the Alpine area. As a first step the summer testing situations were chosen in such a way that the large-scale synoptic forcing was weak, hence the orographic forcing and local heating were dominant. ALADIN was able to simulate the convection in quite a realistic way, except that it was starting systematically too early in the model with respect to the reality. One plausible reason for this early start of convection is that the model PBL is too moist. However it was very difficult to evaluate the realism of humidity profiles: there are too few soundings available around the studied area.

More details can be asked to : Stephan Greilberger and Thomas Haiden.

- "Adriatic storms" study

We could observe a few cases of a very strong cyclogenesis simulated by ALADIN over the Adriatic Sea last summer. The model produced quite a deep, small-scale cyclone over the sea, looking unrealistic (too deep). It was figured out that this strong cyclogenesis was present in all ALADIN applications covering the area. There were 3 periods with this strong cyclogenesis : a) 17-18th  June  2001, b) 20th  July  2001, c) 3-4-5th  September  2001. For all these situations the model had a very similar behaviour, regarding the sensitivity to CYCORA-x packages. The weakest cyclogenesis was observed for the pre-CYCORA settings. When getting convinced on the similarity of the model simulations, the case of 20th  July was studied more in depth. What was possible to conclude from the available observations (unfortunately absent on the sea surface) was that the forecast was very realistic despite the exaggerated deepening above the sea. There was a cyclogenesis in reality and the diagnosed model fields were corresponding to the cyclogenesis process known in literature for the Adriatic Sea area. A few tests changing the tunings of the physics were made but without having any impact on the forecast. The only found modification to weaken the cyclone was to cut the evaporation from the sea surface. We did not succeed to discover which mechanism(s) were responsible in the model to get so active deepening. Nevertheless these cases are very interesting and shall be used as benchmark for new modifications made in the ALADIN physics.

More details can be asked to : Zoran Vakula, Radmila Brozkova and Jean-François Geleyn.

- Sensitivity study of the radiation scheme

The aim of this work was to help in the effort made on development and tuning of the radiation scheme. The radiation scheme is a part of CYCORA-x packages, however, the very last modifications meant to enter the CYCORA-bis package had to be abandoned due to their unstable behaviour when going to 41 levels. Since the problem is quite complex, the work made in Prague aimed at the detection of the most sensitive namelist parameters for the computation of gas absorption coefficients. This was made by comparing to reference results obtained from a more complex but expensive radiation scheme (with 1d column model), all this for 50 vertical profiles chosen randomly around the globe, with special emphasis on one of them. As outcome some better understanding of the scientific problem was reached.

More details can be asked to : Helga Toth, Roger Randriamampianina, Jean-François Geleyn.

Developments in the verification

The verification script based on VERAL was created in order to produce monthly scores for the EWGLAM list of stations. This procedure will take into account all the developments made up to now.

More details can be asked to: Dijana Klaric, Zuzana Huthova, Martin Janousek.

Technical developments

- TUC: Transparent Use of Clearcase

When moving the R & D environment to new work-group server voodoo, the source-code-utility tool, previously based on MaK software was rebuilt on the shareware software CVS. In order to make the life of developers easy, a set of scripts was prepared in the similar spirit like for ClearCase use at GMAP. Hence, "happy tax payers" may find the familiar commands like "cc_get", "cc_edit", "cc_add", etc..., and also the ". cc_quit" one. There is a documentation available on voodoo and the first cycle to be tried is AL15/CY24T1_op1.

More details can be asked to : Filip Vana, Tomas Kalibera.

- porting the cycle AL15

Within the fall of 2001 the cycle AL15/CY24T1 was ported to the NEC SX4 platform. In fact, two branches were ported: the branch "op1" for validation, since it is scientifically the same like the current operational version of ALADIN/LACE (CYCORA-bis), and branch "op2", containing the modifications of CYCORA-ter. Within the porting two general bugs were already found and reported: missing interface in one routine in the ALADIN transform package and a bug in computation and write-out of the fluxes. Despite the fact, that these bugs were real ones, they were not detected on the VPP platform. The cycle was validated on 1 processor and gave bit-identical results in the adiabatic mode as the previous cycle. However, another bug manifests itself when the run is switched to more processors: straight after reading the initial conditions and filling the SPA3 & SPA2 arrays the spectral norms of vorticity are affected in a systematic way. The investigation continues.

More details can be asked to: Martin Janousek

- porting ODB

The ODB software to create and treat the observation files was ported to the SX4 platform last summer. However, it was not tested yet due to the problems detected in the model cycle.

More details: Metod Kozelj

Home