Interface diabatických procesů s pohybovými rovnicemi v numerickém předpovědním modelu počasí ALADIN

• Název práce v češtině: Interface diabatických procesů s pohybovými rovnicemi v numerickém předpovědním modelu počasí ALADIN
• Název v anglickém jazyce: Interface of diabatic processes with equations of motion in numerical weather prediction model ALADIN
• Akademický rok vypsání: 2009/2010
• Typ práce: diplomová práce
• Jazyk práce: čeština
• Ústav: Katedra fyziky atmosféry (32-KFA)
• Vedoucí / školitel: RNDr. Radmila Brožková, CSc.

Zásady pro vypracování

- Seznámit se se základními hypotézami dynamiky a termodynamiky použitých pro odvození řídících rovnic numerického předpovědního modelu počasí.
- Analyzovat problematiku projekce diabatických zdrojů/propadů na dynamickou část systému, a to v případě relaxace hydrostatického přiblížení.
- Navrhnout úpravu stávajícího řešení interface diabatických členů s dynamikou modelu.

Seznam odborné literatury

Catry, B., J.-F. Geleyn, M. Tudor, P. Bénard and A. Trojáková, 2007: Flux conservative thermodynamic equations in a mass weighted framework. Tellus, 59A, 71-79.
Piriou, J.-M., J.-L. Redelsperger, J.-F. Geleyn, J.-P. Lafore and F.
Guichard, 2007: An approach for convective parameterization with memory:
separating microphysics and transport in grid-scale equations. J. Atmos.
Sci., 64, pp. 4127-4139.

Předběžná náplň práce v anglickém jazyce

Moving from Hydrostatic Primitive Equations (HPE) systems to fully compressible Euler Equations (EE) ones did put several challenges in front of people developing so-called physical parameterisation schemes:
- rethinking the treatment of the 'governing equations', because EE systems must be used at scales where one otherwise cannot consider the hydrometeors as 'passive tracers' from the physical-dynamical point of view;
- facing the fact that, quitting hydrostatism, the physical reality is that heat sources/sinks act in nature on both temperature and pressure;
- reconsidering the interaction between various parameterisations in order to have solutions covering a maximum of modelled scales.

All this is more or less on good way (Catry et al, 2007, Laprise, 1998 and Piriou et al., 2007) but this opening to new model characteristics is paradoxically lacking a most direct ingredient: "what is the impact of existing physical parameterisations on the new prognostic variable associated with a tri-dimensional view of the wind field in EE models?" .

Indeed, for parameterisation schemes used in HPE systems, the horizontal momentum 'feels' the sub-grid effects of mountain drag, turbulence and convection.

The first of the three processes is here of little importance because it is admitted that most of the various shapes of orographic stress' upward propagation are fully resolved by the model dynamics when one reaches scales where non-hydrostatism matters.

For the second of the three processes, the same down-gradient approach as for horizontal momentum can be applied to vertical momentum. In the case of the ALADIN discretisation there exists however an additional difficulty: the prognostic variable for vertical momentum is not known in the middle of the model layers (alike the horizontal momentum components) but at the interface between layers. A paper study performed some ten years ago however showed that there is a rather simple way to solve the associated problem.

The last of the three processes is the most interesting one in the proposed study. It has received little attention up to now because it was thought that momentum-transporting convection ought to be fully resolved by the model dynamics at scales where non-hydrostatism matters. But there is now mounting evidence (Gerard et al., 2008) that this is not fully the case. Hence the fact that vertical sub-grid convection transports upwards a systematically rising vertical velocity ought to have some direct impact on the model's dynamics via the 3D divergence term, on top of the already considered thermodynamic impact of deep convection in non-hydrostatic conditions. Here also, the vertical staggering issue (see item just above) will have to be handled and the associated 'paper study' shall be the first step of the proposed study.

The aim of the work will hence be to treat the three (or at least two) above-listed 'missing parameterisation terms' in reverse order and to test their impact on simulations with model horizontal mesh-sizes between 2 and 5km.

The study offers a nice introduction to the complexity of the physical parameterisation trade within a rather isolated scope and with the perspective of new findings. The theoretical basis to acquire are rather simple but their translation to a workable algorithm will require to cleanly master a transversal work across several parts of an existing code (ALADIN).