Parameters values : z0_over_height, height_displacement, ct_karman

[dsolyga]

Jan suggested to modify the following parameters values :

• z0_over_height : from 0.0625 to 0.046
• height_displacement : from 0.75 to 0.67
• ct_karman : from 0.35 to 0.40

To note: z0_over_height and height_displacement are externalized, so if needed these parameters can be reset before validation.
See Branches/MergeHydro/logz0_note.

Tests

Sensitivity tests need to be done for evaluation and validation (off-line/coupled).

[aducharne]

ct_karman is the von Karman constant, established from turbulence similarity analysis.
Its value is usually assumed to close to 0.4, and often set to 0.41

Note that increasing ct_karman will decrease the aerodynamic resistance, thus increase evapotranspiration (ET). It might be a good thing since ET is smaller in ORCHIDEE than in other LSMs in intercomparison projects. It could also help reducing the warm bias over land in teh extra-tropics.

Replying to dsolyga:

Jan suggested to modify the following parameters values :

• z0_over_height : from 0.0625 to 0.046
• height_displacement : from 0.75 to 0.67
• ct_karman : from 0.35 to 0.40

To note: z0_over_height and height_displacement are externalized, so if needed these parameters can be reset before validation.
See Branches/MergeHydro/logz0_note.

Tests

Sensitivity tests need to be done for evaluation and validation (off-line/coupled).

[jgipsl]

height_displacement and ct_karma were changed in revision [3524]. They are now:

height_displacement = 0.66 
ct_karman = 0.41

In the same commit, z0_over_height were extened to be dependent of the pft. It is now:

  REAL(r_std), PARAMETER, DIMENSION(nvmc) :: z0_over_height_mtc = &   ! Factor to calculate roughness height from 
  & (/  0.0, 0.0625, 0.0625, 0.0625, 0.0625, 0.0625, 0.0625,  &       !! vegetation height (unitless)   
  &  0.0625, 0.0625, 0.0625, 0.0625, 0.0625, 0.0625  /)

[jgipsl]

z0_over_height is only used in subroutine condveg_z0cdrag which is only called when rough_dyn=false.

[fcheruy]

For now the default for the computation of the roughness lengths is rough_dyn=y and these values are not used. The request could be reconsidered if the default is changed in Orchidee.
The Von Karman constant has been updated already. However in LMDZ it is set to 0.40, I suggest to do the same in Orchidee.

[jgipsl]

Decided during ORCH meeting today:

• Frederique has done simulations with Ionela Musat. She'll add more information in this ticket.
• for short terme: a patch to be done in LMDZ so that q2m never becomes negative
• for longer terme: Nicolas Vuichard would like to look into ROUGH_DYN parametrization to see what can be done. For this part, the ticket is changed to milestone 4.0