parameters of condveg_soilalb

[nicolasviovy]

we only pass soilalb_dry,soilalb_wet to subroutine condveg_soilalb
whereas we also calculate soilalb_moy that is not passed as argument (and was passed before).
Its work since soilalb_moy is a global variable but it seems strange and others dry and wet are passe and could lead to futur bugs

[jpolcher]

I do not understand how this was coded ! It makes the code unnecessarily complicated.

I would propose to suppress the variable soilalb_moy and just for the case of alb_bare_model=TRUE define soilalb_dry = soilalb_wet. This way the soil moisture does not affect any more the bare soil moisture.

So the flag "alb_bare_model" is only needed in the subourtine condveg_soilalb and the entire condveg module simplified.

[nvuilsce]

Replying to jpolcher:

I do not understand how this was coded ! It makes the code unnecessarily complicated.

I would propose to suppress the variable soilalb_moy and just for the case of alb_bare_model=TRUE define soilalb_dry = soilalb_wet. This way the soil moisture does not affect any more the bare soil moisture.

So the flag "alb_bare_model" is only needed in the subourtine condveg_soilalb and the entire condveg module simplified.

It could be done as suggested by Jan.
If so, I would suggest to use temporary variables to store vis_dry, nir_dry, vis_wet and nir_wet and to initialize them as followed, at the top of the condveg_soilalb subroutine:

 IF ( alb_bare_model .EQ. TRUE) THEN
    vis_dry_temp=vis_dry
    nir_dry_temp=nir_dry
    vis_wet_temp=vis_wet
    nir_wet_temp=nir_wet
 ELSE
    vis_dry_temp=(vis_dry+vis_wet)/deux
    nir_dry_temp=(nir_dry+nir_wet)/deux
    vis_wet_temp=(vis_dry+vis_wet)/deux
    nir_wet_temp=(nir_dry+nir_wet)/deux
 ENDIF

and to use these temporary variables after within the code of condveg_soilalb.

Concerning the comment made by Nicolas, would it be ok if we suppress from the arguments list of condveg_soilalb the 2 variables soilalb_dry and soilalb_wet because they are, like soilalb_moy, module variables.
About this routine, how do you understand the setting of the 4 variables when there is not point associated into the soilcolor map.

        soilalb_dry(ib,ivis) = (SUM(vis_dry)/classnb + SUM(vis_wet)/classnb)/deux
        soilalb_dry(ib,inir) = (SUM(nir_dry)/classnb + SUM(nir_wet)/classnb)/deux
        soilalb_wet(ib,ivis) = (SUM(vis_dry)/classnb + SUM(vis_wet)/classnb)/deux
        soilalb_wet(ib,inir) = (SUM(nir_dry)/classnb + SUM(nir_wet)/classnb)/deux

Why do we mix dry and wet classes for defining both dry and wet albedo ?

[jgipsl]

soilalb_dry,soilalb_wet are currently not used. Philippe and Vladislav will go through the code and maybe do some cleaning without remove the code.

[nvuilsce]

Indeed, as suggested by Jan, soilalb_dry and soilalb_wet have been removed from the argument list of condveg_soilalb because they are global variables. This has been done in rev r2581