NcML for GEM, ROMS biogeochem and FVCOM-WASP

[jlehrter]

Hi Rich, Do you guys have guidance for how to generate the NcML, or even a script already developed for translating our model state variables to the CF format? Thanks, John

[rsignell-usgs]

@jlehrter , which models are we talking about here?
Can you point me toward an example?

[jlehrter]

@rsignell-usgs, GEM to start with, but ultimately for ROMS biogeochem and FVCOM-WASP too. The list of state variables for GEM is attached. If you want an example netcdf, I can send that via email.

[rsignell-usgs]

I've copied the sample file you provided by email to comt.sura.org:/data/comt_2/gom_hypoxia/EPA_GEM/sample/output.1Day.AllVariables.nc

which makes it viewable on OPeNDAP at http://comt.sura.org/thredds/catalog/comt_2_full/gom_hypoxia/EPA_GEM/sample/catalog.html?dataset=comt_2_full/gom_hypoxia/EPA_GEM/sample/output.1Day.AllVariables.nc

I'll take a look at the NcML now

[rsignell-usgs]

@jlehrter , can you please upload a 2 time step file that contains real lon/lat and depth values? You should be able to do sftp sura_ftp@comt.sura.org

[rluettich]

We have a document that is supposed to outline how to get compliant model results into the COMT archive. It would be great to use this as an opportunity to assess the adequacy of this document since, in the long run, having such a document is critical for the COMT:

https://github.com/ioos/comt/wiki/COMT2%20Inventory%20and%20Upload%20Guidelines#prepare

[jlehrter]

Thanks, and yes we've looked at these instructions. We are looking for guidance about how to create the NcML file.

[rluettich]

Hopefully we can add a section to the instruction that addresses NcML use as well.

[brianmckenna]

@rluettich @rsignell-usgs

The best I can think we can provide is very barebone templates for models we have handled and know about, assuming the output remains the "default"...eg. ADCIRC, FVCOM, etc

In this particular case, I don't believe we've ever seen the NCOM_GEM model, so some manual interaction would be necessary to map the default model output variables to CF compliant fields.

As an example, not sure if it's even relevant to COMT objectives, but there's no automated way to know that NCOM_GEM output variable Qn1 (Phytoplantkton(sp?) group 1 nitrogen quota) is something like CF standard_name mass_concentration_of_phytoplankton_expressed_as_chlorophyll_in_sea_water or whichever is the appropriate field.

It would help greatly if each COMT project had a strict list of required model output variables (eg. water level for inundation) and even desired model variables, which can be used to create barebone NcML templates.

[jlehrter]

I think a standard list of model variables for each project is a good idea. Especially in the biogeochemical models, there will be differences in the way various processes are represented. Some vars, like Qn in the example above, are specific to our model but are used to obtain the phytoplankton growth rate, which is generic across all the biogeochem models. I'll talk with the folks on our project about putting together a list.