Example Ocean Model data sets for TAMOC testing

[ChrisBarker-NOAA]

As discussed when @socolofs was in Seattle, we'd like a couple (or more?) test data sets to exercise TAMOC / GNOME coupling.

As a start, let's use the Navy ESPC model (formerly know as "HYCOM").

I'm going to suggest the Gulf -- but not Deepwater Horizon exactly -- maybe in about 1000m of water?

Alternatively -- the Hebron Oil Field in Canada?

Two cases:

• Surface only -- as it's a common case for us, and it can excercise providing teh depth provfiles separately.

• 3D -- then we can get the depth profiles from the ocean model results.

We'll need:

• Currents
• Temperature
• Salinity

Anything else??

Time:
3 days?

Size of region:

Big enough to cover three days of transport -- depends??? We can always start small and make it bigger if we need to.

@AmyMacFadyen, @RachaelMueller: I think you started on this, but then got side-tracked by libgoods, which should be able to make this easy, but maybe doesn't quite do it yet?

Once we have set of files, we can move all this forward ....

We can put them in the GDrive folder we set up.

[RachaelMueller]

Hi Chris, I ought to be able to pick this back up either late this week or next week. It will be good practice with LibGOODS. @amy MacFadyen - NOAA Federal ***@***.***>, can LibGOODS handle the 3D case? Cheers, Rachael

[RachaelMueller]

re: location: I had chosen the SW gulf for the test case because of a confluence of depth and domain. Depth, because the continental shelf slopes off closer to shoreline. The domain is more limited down there but I *think* it covered 1000 meters. @chris Barker - NOAA Federal ***@***.***>, what is more important: depth (1000m) or model product?

[ChrisBarker-NOAA]

@chris Barker - NOAA Federal @.***>, what is more
important: depth (1000m) or model product?

@socolofs: your thoughts on this?

My thought is that we want to be deep enough that that the plume will "trap" well below the surface.

And the more "interesting" the profiles are, the better. -- that is if everything is pretty depth uniform, then we're not really exercising the ability to use the profile from the model.

I think the ESPC model is about it for options.

I"d like to test an unstructured grid model, but I don't know of a 3D one that meets our other needs -- e.g. the NGOGF2 is pretty much only on the shelf, so not very deep -- maybe deep enough for a smaller release?

@socolofs: one other question:

Should we use the same model/case for the surface only example? Or somewhere else?

[socolofs]

Usually, the plume will trap within 250 to 350 m of rise, so 500 m depth is probably the minimum that will give us a subsurface intrusion. I agree, it would be nice if the profiles are not entirely uniform.

I think it would be good to do at least one case in Global HYCOM (did you all say that model changed names?). This is because in some locations, there may be not other data available.

To me, the surface-only case could be the same case, but I would like to eventually test as many different datasets as are practical to make sure the coupling code is robust. Since I will be using the GNOME grid current object, I'm not sure how important that is, though. The interface to my model should be the same through that object regardless of the input data.

[ChrisBarker-NOAA]

Global HYCOM (did you all say that model changed names?)
yes, that's now called:

Navy Earth System Prediction Capability (ESPC)

They are releasing results re-gridded to the same grid that they used to use, so it look to end users, very similar.

[ChrisBarker-NOAA]

Since I will be using the GNOME grid current object, I'm not sure how important that is, though. The interface to my model should be the same through that object regardless of the input data.

Yes, that's the idea -- though good to test for real, as you may be using stuff that we haven't used in GNOME yet (e.g. 3D temp and salinity) -- but that's our problem.

[AmyMacFadyen]

Global HYCOM (did you all say that model changed names?)
yes, that's now called:

Navy Earth System Prediction Capability (ESPC)

They are releasing results re-gridded to the same grid that they used to use, so it look to end users, very similar.

Just to be pedantic, the numerical model is still HYCOM. The forecasting system being run by NAVO has changed names from GOFS 3.1 to ESPC.

[RachaelMueller]

Criteria:

• depth >500 m
• ESPS (HYCOM)
• NGOFS (FVCOM) -> at least, we talked about a 2-model "best case scenario"

Attached is an image of the region I have in mind. The Navy model is shown in, well, navy of course, and NGOFS is shown in purple. The location of the blue marker is 1000 meters (~3500 ft). If we estimate a max 1 knot drift speed (is that a generous enough upper bound?), then I think a few clicks more than 1 degree in meridional and zonal directions will suffice for a 3-day run.

Sound good? Any changes/preferences to differ from the above?

Thanks @jay-hennen for the new mapping feature! :-)

[RachaelMueller]

correction: NGOFS is FVCOM and not ROMS...though the grid shown is not a triangular mesh, which I find confusing. The purple grid is...what...an interpolated mesh of the NGOFS grid???

[dylanrighi]

correction: NGOFS is FVCOM and not ROMS...though the grid shown is not a triangular mesh, which I find confusing. The purple grid is...what...an interpolated mesh of the NGOFS grid???

did you get it from WebGNOME? cuz the NGOFS available there is a version interpolated onto a regular grid. (oh wait...you are in WebGNOME in the image clip...so yes?

Use libgoods or the OFS subsetter if you want a version not interpolated to a regular grid.

[AmyMacFadyen]

I even put it in bold 🤣 [image: image.png]

…

On Tue, Sep 9, 2025 at 2:57 PM Dylan Righi ***@***.***> wrote: *dylanrighi* left a comment (NOAA-ORR-ERD/PyGnome#197) <#197 (comment)> correction: NGOFS is FVCOM and not ROMS...though the grid shown is not a triangular mesh, which I find confusing. The purple grid is...what...an interpolated mesh of the NGOFS grid??? did you get it from WebGNOME? cuz the NGOFS available there is a version interpolated onto a regular grid. (oh wait...you are in WebGNOME in the image clip...so yes? Use libgoods or the OFS subsetter if you want a version not interpolated to a regular grid. — Reply to this email directly, view it on GitHub <#197 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABHEUJZIONIWAWPKLYS4TYT3R5ELZAVCNFSM6AAAAACFOY45L2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTENZSGQYDSOBXGU> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[RachaelMueller]

The issue with the NGOFS domain is just a graphical one. The domain shown in the graphic does NOT show the model extent. NGOFS doesn't extend that far east. It's a part of this graphical display that I don't really care for, but I am working around it by looking at the current vectors instead of the grid.

Turns out that I was wrong about focusing on a part of the shoreline with a relatively short continental slope. Looking off the coast of LA now as I think that will get us where we want to go with all the above criteria plus an ability to run with multiple models (if wanted).