[MITgcm-support] Stratification simulations

Martin Losch Martin.Losch at awi.de
Tue Feb 12 02:17:05 EST 2008


David,
the tRef (and sRref) 1D-arrays have two uses:
1. provide the default initialisation (to be overwritten if you set  
the hydrogThetaFile, see ini_theta.F)
2. reference temerature (and salinity) for the linear equation of  
state rho = rhoNil*( 1 - talpha * ( theta - tref ) + sbeta * ( salt -  
sref )), see find_rho.F

So if you have eostype = 'LINEAR' (which is the default, see  
set_defaults.F), and you specify the temperature gradient only via  
tref, your initial temperature is the same as your reference  
temperatue, so RHOAnoma is in your case rho-rhoConst= - 
rhoConst*talpha (tref - tref), if rhoConst=rhoNil, which is usually  
the case unless you set then to something different.  A simple test  
would be to use tref=const and specify the gradient via an input  
field (read in as hydrogThetaFile). Alternatively you could use one  
of the non-linear equations of state, then tref is only used for  
initialisation.

the tempForcing flag should have no effect in this context (unless  
you want to use some external/surface forcing)
when you set tempStepping to false, the temperature field will not  
change, is that what you want?
calc_gw.F only computes the terms in the w-equation that are not used  
for hydrostatic computations. The gravity is part of the hydrostatic  
pressure, right? It's computed in calc_phi_hyd.F

Martin

On 11 Feb 2008, at 20:31, David Hebert wrote:

> Hi all,
>
> I have run two identical simulations, one with a constant  
> temperature field, and one with temperature gradient specified via  
> tRef array. The simulations are run in non-hydrostatic mode. When I  
> compute the vertical velocity fields, and consequently vertical KE  
> = 0.5 w^2, the results are identical.
>
> I have checked the temperature field and indeed there is a  
> temperature gradient. Also, the diagnostic DRHODR and there is a  
> value there for each vertical level of -0.0408. However, the  
> diagnostic RHOAnoma and RHOANOSQ are all zeros. It seems as if the  
> density stratification does not affect the flow. I have turned on  
> the tempForcing flag. Are there any flags in the datafile that  
> should be turned on?
>
> Other notes/questions:
> 1) I turned tempStepping off because I wanted a constant gradient  
> just to demonstrate effects due to stratification. Should this be  
> turned on?
> 2) Do I need to have an initial temperature field to obtain a  
> RHOAnoma value, since it is RHOAnoma is used in the code
> 3) My domain is small, 480x160x160m^3. Grid resolution 1m. Velocity  
> max = 0.2 m/s. Simulations integrated with dt=1s for 10800 seconds.
> 4) I have been looking through src/calc_gw.F and don't see a  
> gravity term. Could someone point me to where is the gravity term  
> introduced into the model? Is it in correction_step.F? I see  
> something in calc_buoyancy.F, but I do not see where CALC_BUOYANCY  
> is being called from.
>
> Any help in setting up the model for stratification is greatly  
> appreciated. Thank you much,
>
> David
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support




More information about the MITgcm-support mailing list