[MITgcm-support] instantaneous density anomaly diagnostics

[Patrick Heimbach]

Sergey,

On Thu, 2006-01-12 at 10:33, Sergey Vinogradov wrote:
> Hello everyone,
> 
> I need to look at density anomaly instantaneous, not time-averaged, 
> output (created in the same manner as Eta.*, T.*, etc in write_state.F, 
> with specified dumpFreq).
> 1. Is there any way I can add the line into write_state.F such as 
> something like CALL WRITE_FLD_XYZ_RL('RHOA.', suff, 
> alphaRho,myIter,myThid) ?
> If done as shown above, it does not know the variable alphaRho (or any 
> other "rho-related" variable names) as it is not in the list of 
> DYNVARS.h. So it gives an error.

I would instead use diagnostics (our option 2).

> 2. My next option is to use RHOAnoma var in diagnostics pkg. In 
> data.diagnostics, is it possible to specify the frequency(1) in a 
> meaning of dumpFreq, not taveFreq, as I don't need time-averaging done? 
> I can always set frequency(1)=1 and retain only files that correspond to 
> dumpFreq but it is not very efficient approach.

Put negative frequencies, that will give snapshosts
instead of time-means.

> 3. I tried to include diagnostics pkg in my setup. It compiled ok but it 
> didn't save any output that I specified in data.diagnostcics. I searched 
> through our mailing list but I didn't see such problem being mentioned. 
> I included data.diagnostics, changed data.pkg and packages.conf (all 
> attached). Does anyone see the reason why it wouldn't do the output? 
> (the standard output files are still created).

Did you set in data.pkg
 useDiagnostics = .TRUE.
?

-p.

> Thank you,
-- 
--------------------------------------------------------
Patrick Heimbach   Massachusetts Institute of Technology
FON: +1/617/253-5259                  EAPS, Room 54-1518
FAX: +1/617/253-4464             77 Massachusetts Avenue
mailto:heimbach at mit.edu               Cambridge MA 02139
http://www.mit.edu/~heimbach/                        USA