[MITgcm-support] pressure from MITgcm

[Jean-Michel Campin]

Hi Sergey,

On Mon, Nov 21, 2005 at 11:28:31AM -0500, Sergey Vinogradov wrote:
> Hi All!
> 
>    Jean-Michel: I attached my data file. I experimented with changing 
> eos from poly3 to linear and it had no noticeable changes in the 
> results. However, I didn't try anything else. I guess my offline density 
> computations can have some flaws: I don't interpolate vertical density 
> integral; I use constant density for the whole layer instead. I use some 
> old seawater eos subroutines.

It seems that you are using kind of standard parameters:
You are not using the "Finite Volume" form (constant density
for the whole layer) but the "energy conserving" form
where simple average density between level k & k-1 is used
in the integration of the hydrostatic pressure (see equation 2.92
of the manual) (and multiplied by drC(k)).

I have 2 suggestions:
1) you can try to use the same discretisation as the model uses
  (might not be negligible).
2) with linear EOS, you could check not only the variability
  but also the absolute value of the hydrostatic pressure
  (might be easier to see what is missing)

>    Following Dimitris suggestion, I have added Pbar(t) to the pressure 
> computation from density, also replacing the full sea-surface height 
> with just the dynamic signal (dynamic signal = full ssh minus inverted 
> barometer response). This changed the picture (see attached plot), 
> bringing the offline pressure computation closer to one that was 
> computed from totPhiHyd. Still, there is a significant difference in HF. 

Yes, the atmospheric loading is already included in totPhiHyd, I think.

> I'm not sure if this would be the nonlinearity effect of  hydrostatic 
> equation, as Martin suggests. 

Should not be present with linear EOS, POLY-3, or JMD95Z.

Cheers,

Jean-Michel