PowerPoint-presentatie

1
8.4 Single-column modeling versus mesoscale
modeling for an episode with three diurnal cycles
in CASES99 Cisco de Bruijn (1), Sander Tijm(1),
Gert-Jan Steeneveld(2) 1 Royal Netherlands
Meteorological Institute 2Meteorology Air
Quality Group Wageningen Univ. 17BLT, 24 May
2006, San Diego
2
Contents

• Model setup of HIRLAM
• Idealized case (GABLS2)
• Realistic case (Mesoscale intercomparison)
• Improved 1D run

3
HIRLAM High Resolution Limited Area Model
4

• HIRLAM physics
• Radiation (Savijärvi)
• Convection STRACO (Sass et al.)
• CBR turbulence (E-l) (Cuxart et al.)
• Surface scheme ISBA (Noilhan et al.)

5

• Prescribed initial profile
• No radiation scheme
• No surface scheme
• Large scale synoptic divergence w-0.005 m/s
• Geostrophic wind u3 v-9 m/s

6

• Prescribed Skin temperature
• No surface scheme

7
Wind speed 36h Valid 24 Oct99 0200 LT
40 level 1D Hirlam, CBR 6.3.8
Wind direction 36h Valid 24 Oct99 0200 LT
8
Wind speed 36h Valid 24 Oct99 0200 LT
100 level 1D Hirlam, CBR 6.3.8
Wind direction 36h Valid 24 Oct99 0200 LT
9
Mesoscale experiment HIRLAM

• 406 by 306 points
• Grid 11km
• Timestep 180 s
• LBC every 6h

Leon (Kansas)
10
Wind speed 36h Valid 24 Oct99 0200 LT
Wind direction 36h Valid 24 Oct99 0200 LT
11
Conclusions 1D GABLSII

• Agreement of 1D output with observations is
  rather poor
• Agreement of 1D model with other models is much
  better (not shown)
• GABLSII recipe is too restrictive to get the most
  out of the model

12
Improved 1D-model run

• Radiation scheme ON (Sarijarvi)
• Surface scheme ON (ISBA)
• Initial profile from 3D model
• Large scale synoptic divergence w-0.005 m/s
• Geostrophic wind NOT constant

13
Geowind
Prescribed every hour 3D model, geopotential at
925 hPa Wind above ABL from 3D model (top ABL)
14
Wind speed 36h Valid 25 Oct99 0200 LT
Wind direction 36h Valid 25 Oct99 0200 LT
15
With Galperin/Perov adjustments to CBR
Wind speed 36h Valid 25 Oct99 0200 LT
16
Theta 36h Valid 25 Oct99 0200 LT
BUT with temperature advection 0.005 K/s
17
Conclusions

• 1D-model good diagnosis tool, very nice
  development tool
• Boundary and initial conditions 1D-model very
  important
• Selection case study Geowind should not vary too
  much, dynamic forcing should be small, good study
  of radiosondes and/or model analyses

18

• Outlook
• Use of radiosonde archive at KNMI for De Bilt to
  search weak dynamic forcing cases
• Strong daily cycle (dry and/or spring/summer)
• Use Cabauw observations (tower, profiler, other
  measurements) to validate further CBR developments