From: Julien Jouanno <jjlod@locean-ipsl.upmc.fr>
To: "francois.colas" <francois.colas@locean-ipsl.upmc.fr>, 
	christophe.hourdin@locean-ipsl.upmc.fr
Date: Fri, 18 Oct 2013 14:11:38 +0200

J'ai aussi quelques pixels dans la conf avec forcages climatologiques.
Mais rien de grave...

Peut etre augmenter un peu la diffusion horizontale si vous utilisez pas
UBS.

Pour info quelques éléments sur la config :
- bilap sur la dynamique avec shema en vector form 
- laplacien isoneutral sur les traceurs
- ldf_tra_c2 activé
- ln_zdfevd   = .true. !  convection: enhanced vertical diffusion (T) or not (F)
- nn_evdm     =   0       !  enhanced mixing apply on tracer (=0) or on
tracer and momentum (=1)
- rn_avm0     =   1.e-4   !  vertical eddy viscosity   [m2/s]          
- rn_avt0     =   1.e-5   !  vertical eddy diffusivity [m2/s]          
- cycle diurne activé

Saludos
J/



&namzdf_tke    !   turbulent eddy kinetic dependent vertical diffusion
("key_zdftke")
!-----------------------------------------------------------------------
   rn_ediff    =   0.1     !  coef. for vertical eddy coef.
(avt=rn_ediff*mxl*sqrt(e) )
   rn_ediss    =   0.7     !  coef. of the Kolmogoroff dissipation
   rn_ebb      =   60      !  coef. of the surface input of tke
   rn_ebbice   =   3.75    !  coef. of the surface input of tke under
ice

   nn_havti    =    1      !  horizontal shape for avtb (=1) or not (=0)
under ice
   rn_emin     =   1.e-6   !  minimum value of tke [m2/s2]
   rn_emin0    =   1.e-4   !  surface minimum value of tke [m2/s2]
   rn_bshear   =   1.e-20  !  background shear (>0)
   nn_mxl      =   3       !  mixing length: = 0 bounded by the distance
to surface and bottom
                           !                 = 1 bounded by the local
vertical scale factor
                           !                 = 2 first vertical
derivative of mixing length bounded by 1
                           !                 = 3 same criteria as case 2
but applied in a different way
   nn_pdl      =   1       !  Prandtl number function of richarson
number (=1, avt=pdl(Ri)*avm) or not (=0, avt=avm)
   ln_mxl0     = .true.    !  mixing length scale surface value as
function of wind stress (T) or not (F)
   rn_lmin     =   0.001   ! orig 0.001   !  interior buoyancy lenght
scale minimum value
   rn_lmin0    =   0.01    ! orig 0.01    !  surface  buoyancy lenght
scale minimum value
   nn_etau     =   1       !  exponentially deceasing penetration of tke
due to internal & intertial waves
                           !        = 0 no penetration ( O(2 km)
resolution)
                           !        = 1 additional tke source (rn_efr *
en)
                           !        = 2 additional tke source (rn_efr *
en) applied only at the base of the mixed layer
                           !        = 3 additional tke source (HF
contribution: mean of stress module - module of mean stress)
   nn_htau     =   1       !  type of exponential decrease of tke
penetration
                           !        = 0  constant 10 m length scale
                           !        = 1  0.5m at the equator to 30m at
high latitudes
                           !        = 2  30 meters constant depth
penetration
                   ! option used only id nn_etau /= 0
   rn_efr      =   0.05    !  fraction of surface tke value which
penetrates inside the ocean
                   ! option used only if nn_etau = 3
   rn_addhft   =  -1.e-3   !     add offset   applied to the "mean of
stress module - module of mean stress" (always kept > 0)
   rn_sclhft   =   1.      !     scale factor applied to the "mean of
stress module - module of mean stress"
   ln_lc       = .true.    !  Langmuir cell effect
   rn_lc       =   0.15    !  coef. associated to Langmuir cells