NH equations, academic tests and open questions on dynamics (presentation, March 2008)
Presentations of Pierre Bénard during the 2nd AROME training course, Lisbon, 4-7 March 2008 :
- Equation System
- Starting from (simple) Euler Equations system in z
- leads to (rather more complicated) Euler Equations system in eta coordinate and new prognostic variablesMinutes of the Discussion about dynamics :
- Academic tools versions of AROME exist for testing and validation ("easy to use").
- 1D Discrete linear analysis
- 2D vertical plan in various configurations
- 3D (LMAP=FALSE) academic orographic waves
- LB Coupling
- conflicts at edges : differences between host and guest models
- well- posedness of LB coupling
- VFE in NH framework
- Fireworks and interactions between dynamics and physics
- Horizontal diffusion and precipitating systems
- "Gard 2005" (flood) event 6 september 2005
Four important items on dynamics were mentionned by P.Bénard: only the first (LB coupling) and fourth (Horizontal diffusion and precipitating systems) points were subject to discussions.
- Regarding the LB coupling, an example of heavy (>100mm) unphysical rainfall occuring in the coupling zone of Arome was shown. This problem originating from a conflict between host and guest models was not found to occur only in Arome (over sea and over orography) but also within the Hirlam community (as well when coupling two Hirlam models as than when coupling an Hirlam model with an ECMWF one). Within Hirlam models this problem was associated with low-level convergence at the boundary. Over mountains the difference of orography between host aand guest models seems to cause this conflict . A solution investigated at Météo-France was to relax as well the orography in the coupling zone as done by K. Essaouini who showed positive effects on spurious rains over mountains (but obviously not over sea). Another possible solution (not investigated because technically not straightforward) could be to enlarge the width of the Davies relaxation zone. S. Malardel wondered also if the convergence of physics between host and guest models could not solve these conflicts.
- P. Bénard recalled that in Arome, strength of horizontal diffusion was decreased by a factor of 4 in order to limit firework’s occurence. Once fireworks eliminated it had little impact on model’s output at Météo-France. Y. Wang mentionned that on MAP case, modification of horizontal diffusion did not modify rainfall forecast significantly.