Introducing GMAO ODAS obs sources

[Dooruk]

Added currently available ODAS observations to SWELL. Reminder that in the long term geos_ocean will go away and I will take out inactive obs. YAMLs. SOCA naming sprint changes are included.

Important changes:

Two changes regarding get_observations.py that are important to emphasize (FYI @rtodling @gmao-wgu @mranst @ashiklom ):

• R2D2 fetch method now includes a ignore missing: True flag, so it just creates a warning if the observation file is missing. This is necessary because we want Variational Bias Coeff.s to propagate regardless of the missing obs. This will need further testing by ADAS (@rtodling)

• Since there could be multiple providers for observations (odas and gdas_marine (EMC) for marine) I switched obs_provider key to be a check-list so get_observations.py can browse multiple locations. There could be better handling of this in the future with R2D2v3 but this is where we are for now.

Some Quick Plots:

Showing scatter plots for ARGO data jedi vs h(x) salinity and temperature (black dots failed the QC).

Salinity	Temperature

FYI @DavidSRussell, you should be able to test with these YAMLs (by setting odas as the obs provider) and in theory it should be easy to add SWOT altimeter data to the list.

• Filters in YAMLs need to be modified according to our needs, @lren20 can help with this part.

[rtodling]

Quite nice!!!

[DavidSRussell]

I'm getting a cannot find field sea_ice_area_fraction error in job.out. I noticed there is a variable called sea_area_fraction elsewhere; maybe that's a conflict?

[Dooruk]

I'm getting a cannot find field sea_ice_area_fraction error in job.out. I noticed there is a variable called sea_area_fraction elsewhere; maybe that's a conflict?

Hmm, can you check your fields_metadata.yaml in the geos_ocean folder? There should be sea_area_fraction

[DavidSRussell]

I'm getting a cannot find field sea_ice_area_fraction error in job.out. I noticed there is a variable called sea_area_fraction elsewhere; maybe that's a conflict?

Hmm, can you check your fields_metadata.yaml in the geos_ocean folder? There should be sea_area_fraction

There is sea_ice_area_fraction under ice state variables with no categories and sea_area_fraction under built-in derived variables that you don't need to worry about. By the way, I'm pretty sure I got a similar error before whenever sea_ice_area_fraction wasn't commented out in the observation yamls.

[Dooruk]

There is sea_ice_area_fraction under ice state variables with no categories and sea_area_fraction under built-in derived variables that you don't need to worry about. By the way, I'm pretty sure I got a similar error before whenever sea_ice_area_fraction wasn't commented out in the observation yamls.

Derived variables one is the GeoVals, and sea_area_fraction in particular is something that we have been using quite a bit. Just to make sure, you are not pointing to your own SwellStatic files correct?

[Dooruk]

Ok, on a second look, that GeoVal (sea_ice_are_fraction) is used by sst_merra2 and that would require a sea_ice state input. So you are right, if you turn that filter off it should run.

[DavidSRussell]

OK, after commenting out those lines and changing the analysis variable names in the override yaml, it seems to be working fine.