220 likes | 356 Views
A N R. Agence nationale de la recherche. COPS - FRANCE. LA, Toulouse IPSL, Paris CNRM, Toulouse LaMP, Clermont-Ferrand. Meso-NH IOP 8B. COPS - France. Instrumentation on supersite V Lidar Juan Cuesta Radar Joël Van Baelen Reinforcement of the GPS network Cédric Champollion
E N D
ANR Agence nationale de la recherche COPS - FRANCE LA, Toulouse IPSL, Paris CNRM, Toulouse LaMP, Clermont-Ferrand
Meso-NH IOP 8B COPS - France • Instrumentation on supersite V • Lidar Juan Cuesta • Radar Joël Van Baelen • Reinforcement of the GPS network • Cédric Champollion • Safire Falcon / Leandre II operation • Cyrille Flamant • Numerical modelling • Jean-Pierre Chaboureau • Evelyne Richard • High-resolution assimilation • Pierre Brousseau • Olivier Caumont Other case studies …see Chaboureau
DLR Poldirad at Waltenheim sur Zorn 14:40 UTC 15:00 UTC 14:20 UTC 14:00 UTC 15:20 UTC 15:40 UTC 16:00 UTC IOP 8B : 15 July 2007
12:15 – 13:00 13:15 – 14:00 14:15 – 15:00 15:15 – 16:00 Time evolution : Observation versus simulation Montancy radar observations (2dBz contour – PPI 1°) Meso-NH (2km - ECMWF driven) (0.1 mm precip contour)
12:15 – 13:00 13:15 – 14:00 14:15 – 15:00 15:15 – 16:00 Time evolution : Observation versus simulation Montancy radar observations (2dBz contour – PPI 1°) Meso-NH (2km - ARPEGE driven) (0.1 mm precip contour)
12:15 – 13:00 13:15 – 14:00 14:15 – 15:00 15:15 – 16:00 Time evolution (close up view) : Observation versus simulation Montancy radar observations (2dBz contour – PPI 1°) Meso-NH (2km - ECMWF driven) (0.1 mm precip contour)
12:15 – 13:00 13:15 – 14:00 14:15 – 15:00 15:15 – 16:00 Time evolution : Observation versus simulation Montancy radar observations (2dBz contour – PPI 1°) Meso-NH (500m - ECMWF driven) (0.1 mm precip contour)
2km Meso-NH simulation : 13:00 UTCSurface streamlines / CAPE / precipitation ( ) ECMWF ARPEGE
2km Meso-NH simulation : 14:00 UTCSurface streamlines / CAPE / precipitation ( ) ECMWF ARPEGE
Vertical cross sections of reflectivity along storm propagation Meso-NH (2km - ECMWF driven) Meso-NH (2km - ARPEGE driven)
Conlusion • High sensitivity to initial conditions • More convection with ARPEGE CI than with ECMWF CI (… with the truth in between!) • Earlier triggering with ARPEGE CI (too early) but a more realistic life cycle • Major role of the low-level convergence • No significant improvement when : • Resolution is increased from 2km to 500m (Triggering occurs too much upwind • 3D turbulence is accounted for (very weak impact) More on that topic : Trentmann (U. Mainz) and Cardwell (U. Manchester)
12:15 – 13:00 13:15 – 14:00 14:15 – 15:00 15:15 – 16:00 15/07/07 Montancy Radar Observations Meso-NH simulation (ARPEGE) Meso-NH simulation (ECMWF)
12:15 – 13:00 13:15 – 14:00 14:15 – 15:00 15:15 – 16:00 Time evolution : Impact of the turbulence scheme (1D/3D) Meso-NH (2km – 1D turbulence) (0.1 mm precip contour) Meso-NH (2km – 3D turbulence) (0.1 mm precip contour)
12:15 – 13:00 13:15 – 14:00 14:15 – 15:00 15:15 – 16:00 Time evolution : Impact of the initial / forcing conditions Meso-NH (2km - ECMWF driven) (0.1 mm precip contour) Meso-NH (2km - ARPEGE driven) (0.1 mm precip contour)
12:15 – 13:00 13:15 – 14:00 14:15 – 15:00 15:15 – 16:00 Time evolution : Impact of the resolution (2km/ 500m) Meso-NH (500m – ECMWF driven) (0.1 mm precip contour) Meso-NH (2km - ECMWF driven) (0.1 mm precip contour)
Meso-NH Forecasts http://mesonh.aero.obs-mip.fr/mesonh/cops/ • 3 domains (32, 8, and 2 km) with 2-way interaction. • Vertical grid with 50 levels up to 20 km with a grid length varying from 60 m to 600 m. • Initial and coupling fields with ECMWF operational forecasts • 30 h forecast starting at 00 UTC • Parameterization schemes: • o 1.5-order turbulence scheme • o ECMWF radiation package • o ISBA surface scheme • o Mixed-phase bulk microphysics: cloud, rain, ice, snow, graupel, and hail (hail is simulated for the inner model only) • o Deep and shallow convection scheme for the 32 and 8 km models only
Outline • Basic model evaluation • Raingauges precipitation measurements • Meteosat IR observations • An example of an isolated thunderstorm forecast • 15 July
24h precipitation (P30h – P06h): 04 July 2007 (J185) • Precip evaluation : • BW stations • MF automatic stations • Model fields interpolated at rain gauge location Thanks to M. Kunz and P. Limnaios
Time evolution of the spatially averaged 24h precip. P30h-P06h (mm) Julian day (July/August)