[Mitgcm-support] RE: about AIM

[mitgcm-support at dev.mitgcm.org]

all,

the coupled configuration uses c46 aim as its
base - the modifications affect how
fluxes are passed around and accumulated, but 
not internal physics or dynamics.

configuration and mods are at

http://mitgcm.org/projects/ESMF/JMClbaseline

chris

-----Original Message-----
From: Jean-Michel Campin [mailto:jmc at gulf.mit.edu]
Sent: Sunday, September 22, 2002 5:31 PM
To: support at mitgcm.org
Cc: jamous at mit.edu; jmc at gulf.mit.edu
Subject: about AIM


Hi,

I have few questions about AIM:
(Daniel, if you have some idea about the 1rst one ...Thanks)

1) Does any one remember why we set FWIND0 to .42 instead of 
the value Franco use in his paper (=0.6)
It's set in INPHYS, line 118:

C
C---  3. Constants for surface fluxes (common SFLCON):
C
      FWIND0 = 0.6 _d 0
cchdbg *****************************************************
      FWIND0 = 0.6 _d 0*0.7 _d 0
cchdbg ******************************************************   

Since FWIND0 is use to compute the wind at the surface from
the wind at the 1rst level, and used after in the stress,
this can contribute to the strong surface wind we found in the
runs.

2) with full-cell and no Shapiro on Q, the global average of P-E 
balance the surface correction term. The 2 are different by less 
than 10^-4 mm/day.

3) Concerning Temperature, it does not work:
the same run with full-cell and no Shapiro on theta :
net heat flux (top) - (surface) = 0.9 W/m2  => warm the atmosphere
surface correction on Theta     = 0.5 W/m2  => warm the atmosphere
latent heat due to exess P-E    = 0.3 W/m2  => warm the atmosphere

My impression is that there is a problem in Heat conservation:
the dynamics conserves theta but the physics conserves the absolute 
temperature: 
If there is a radiative flux Fr from level 1 --> 2,
  Cp*dT1 = -Fr ; Cp*dT2 = +Fr   ==> T is conserved.
but 
  PI_1*dTheta_1= -Fr ; PI_2*dTheta_2 = +Fr  ==> Theta is not conserved
because PI_1 & PI_2 are differents (PI = Cp*(p/Po)**kappa)

And since the dynamics advects theta, only theta can be conserved, 
its my impression.

4) for Chris: to put partial cell in AIM, can I start with the code in 
the repository or would be better to use the coupled version ? 

Jean-Michel