Opened 8 years ago

Closed 6 years ago

#237 closed defect (fixed)

The effect of snow on surface roughness is only accounted when OK_freeze is activated whatever the snow model used.

Reported by: maignan Owned by: cottle
Priority: major Milestone: IPSLCM6.v1
Component: Physical processes Version: trunc
Keywords: Cc:

Description (last modified by maignan)

The effect of snow on surface roughness as it was implemented in a previous version of Wang Tao's code and as it is described in Wang et al., 2013, is not implemented in the current version of the trunk. Only the parametrization tested by I. Gouttevin, consisting in the reduction of surface roughness by a prescribed factor (snowfact =10.), is used when the soil freezing is accounted (OK_freeze = TRUE).
Given that the fraction of snow has been implemented, a better model (as the one proposed in ISBA-ES and tested by Wang et al., 2013, which will compute smoother variations of the roughness in transition periods) could be used. Then, the surface roughness could be calculated with the following equation, instead of reducing z0 by a constant factor whatever the snow fraction (z0/snowfact, Gouttevin et al.,'s model)
surface roughness with snow
where fsg is snow cover fraction, z0t is surface roughness length after considering snow cover, z0 is the vegetation or surface roughness length (m), z0n is the snow surface roughness length baseline value (0.001 m), and zr is the blending height (10 m in ORCHIDEE).

Catherine

Attachments (1)

SnowFact.png (2.7 KB) - added by maignan 8 years ago.
surface roughness with snow

Download all attachments as: .zip

Change History (7)

Changed 8 years ago by maignan

surface roughness with snow

comment:1 Changed 8 years ago by maignan

  • Description modified (diff)

comment:2 Changed 8 years ago by maignan

As a first step we could start with replacing in diffuco_aero:

IF (snow(ji).GT.snowcri .AND. ok_snowfact)  snowfact=10.

with:

IF (snow(ji).GT.snowcri .AND. ok_snowfact .AND. ok_explicitsnow)  snowfact=10.

or with:

IF (snow(ji).GT.snowcri)  snowfact=10.

Catherine

comment:3 Changed 8 years ago by nvuilsce

In order to account for the impact of the snow on the roughness as proposed by Catherine, we will use the snow cover fraction, frac_snow_veg. It is computed in condveg_snow as a function of the snow depth (snowdepth) and snow density (snowrho_ave). In this routine, snowrho_ave is computed as:

snowrho_ave=sum(snowrho,2)/nsnow

The second dimension of snowrho being the snow layers and nsnow, the number of snow layers. But I think that the snow layers may have different depths. If this is true, one should probably weight by the thickness of each layer (snowdz):

snowrho_ave=sum(snowrho*snowdz,2)/sum(snowdz,2) 

comment:4 Changed 6 years ago by jgipsl

  • Owner changed from somebody to cottle
  • Status changed from new to assigned

comment:5 Changed 6 years ago by cottle

  • Status changed from assigned to accepted

As discussed at the retreat, it is better to modify the calculation of the cdrag in condveg to account for snow smoothing. The parametrisation is well done in condveg_z0cdrag_dyn, the same modifications have to be done in condveg_z0cdrag, they are proposed in the ticket 62.

comment:6 Changed 6 years ago by maignan

  • Resolution set to fixed
  • Status changed from accepted to closed

Commit done in r4654
See ticket #62

Note: See TracTickets for help on using tickets.