[MITgcm-support] 2d and 3d-diagnostics in one netcdf file?

[Andrea Molod]

hi martin,

On Tue, 23 Jan 2007, Martin Losch wrote:

> Hi all,
>
> I am embarrassed to ask because it may have been discussed already and I just 
> cannot find it:
> 1. Time-averaging with the diagnostics package: the documentation says that 
> frequency > 0 produces averages, frequency < 0 produces snapshots. What does 
> averagingFeq do?

martin i would have to look into this a bit, but averagingFeq
has to do with the periodic averaging option. for example, 
multi-year january average from a multi-year run, or monthly mean
diurnal cycles. i don't remember exactly how this works, ie,
exactly which freq is specified in averagingFeq.

>
> 2. Is it possible in the diagnostics package to have a 2D and 3D field in the 
> same output stream, e.g.,
>
> frequency(5) = 43200.,
> filename(5)  = 'state_diags',
> fields(1,5)  = 'ETAN','THETA','SALT','UVEL','VVEL','WVEL',
>
> ? With these parameters I seem to be getting the level1 of ETAN (which is 
> correct), but then I also get only the first levels for THETA, SALT, etc. 
> here's the header of the netcdf file, note, that both Zmd000001 and Zld000001 
> = 1. Is that to be expected?


yeah, martin. when you don't specify the levels in data.diagnostics
it uses the minimum of the number of levels of all the diagnostics
in the output stream (see diagnostics_set_pointers) - if you specify
the whole list of levels in data.diagnostics it will write ETAN and
then a bunch of zeros. worth a try?

andrea