"Réseau Formation Recherche" : PhD Studies
- Doina BANCIU : "Specific small scale diabatic forcing in ALADIN at
the limit of the hydrostatic assumption" : report in the next
Newsletter.
- Marta JANISKOVA : "Realization of a simplified physical
parametrization for incremental four-dimensional data assimilation" : see "
PhD".
- Ilian GOSPODINOV : "Conservation Properties of 2 Time Level
semi-Lagrangian" : report in the next Newsletter
.
- Filip VANA : "The dynamical and physical control of kinetic energy
spectra in a NWP spectral semi-Lagrangian model"
- In October 1998 was finished the phasing of all work previously done with
the new cycles to be possible to continue on new supercomputers in Toulouse
and in Prague. The initial configuration slightly modified was phased with
AL08 and because it is modified configuration e926 which doesn't exist any
more in Toulouse, it is possible to create the initial files only in Prague
now. The integration itself (e001) was moved to the cycle AL08T3.
- After this more technical than theoretical work another tool was prepared
allowing creation of nested subdomain with the higher resolution isotropic in
both directions (horizontal). Due the higher resolution the smaller
diffusivity of numerical scheme was achieved which allowed to go further to
the frontal collapse. It was supposed that spectra obtained from this higher
resolution frontal collapse would be sensitive enough to study diffusivity of
the semi-Lagrangian interpolators. To create this tool was not so easy,
several configurations had been created to transform all atmospheric fields
from ideal plane to the real globe keeping the same resolution, to run fullpos
and create nested subdomain (modified e927 configuration), to put the fields
back to the ideal plane with constant Coriolis factor and another
configuration to be results from the new nested integration plotable.
- Starting from middle of December some test to several semi-Lagrangian
interpolators were done. Unfortunately even the nested model with the higher
resolution was insensitive to the kind of interpolator used. It turned out
that the ideal cyclogenesis is probably too geostrophic to create
"trouble" to semi-Lagrangian scheme during the interpolation across
the front because most of the backward trajectories didn't cross the frontal
quasi discontinuous area. To make this situation more sensible to different
semi-Lagrangian interpolators used it was decided to add some diabatic process
to create a cross front component of wind. The vertical diffusion seemed to be
the easiest solution to fulfil this requirement.
- Mark ZAGAR : "Prediction of small-scale events through second level
dynamic adaptation of the planetary boundary layer"
- The possibility of the local small-scale wind forecast using dynamic
adaptation has been investigated in detail and results prove the feasibility
of the method.
- Field of vertical velocity in lower to middle troposphere also results
from the dynamic adaptation. As the forced lift of humid air is contributing
most to the precipitation in mountainous regions it looks promising to connect
the fields of vertical velocity and precipitation in i) deterministic way
(available in theory, to be applied in practice), ii) to find an empirical
(statistical) relation between the two and iii) to apply a 1D model, using the
dynamically adapted momentum field as the forcing and resulting in
precipitation field. The three approaches are to be validated against the
full-range integrated high-resolution model with full physics.
- The model runs for the last mentioned purpose have been already done and
reasonable results have been obtained by hydrostatic model at 2.5 km
resolution, in two heavy-precipitation cases, one governed by convective
processes, while in the other dynamic forcing was the main cause of intense rain.