Changes to enable coupling with WW3 using 3d vortex formulation

[uturuncoglu]

The WW3 changes are merged with UFS WM level PR: ufs-community/ufs-weather-model#2396. These are the remaining changes in the SCHISM side to enable coupling with 3d vortex formulation. This is still draft and before merging it, we need to following,

• Sync UFS Coastal with the changes in UFS WM (@uturuncoglu)
• Sync schism-esmf and schism under UFS Coastal and also sync this development branch (@uturuncoglu)
• Test NC Duck case (@mansurjisan, @yunfangsun)

If everything goes as expected, I could make this PR ready for review. JFYI, @janahaddad

[josephzhang8]

Thx @uturuncoglu.
@danishyo @platipodium: could you please review? Thx

[uturuncoglu]

@josephzhang8 JFYI, I have just synced UFS coastal with UFS WM. This also brings changes related to WW3. I am testing now by running RTs. Along my testing I also tried to update SCHISM but RTs are started to fail with following error.

10:    2: ABORT: INIT: USE_ATMOS must use nws=4

At this point we are using nws = 2 for our regression tests along with -DUSE_ATMOS=ON. I wonder if we could solve it by simply updating our param.nml for the recent version of the SCHISM.

[josephzhang8]

Yes, you can simply change nws=4. Let me know if you still have errors. Thx

[platipodium]

The code looks fine to me, thanks for adhering to the conventions set in the code base. I am not so happy with the naming of the variables 'Sw_*', but that can be fixed at a later stage and might be a consequence of the coupling. Please go ahead and merge.

[platipodium]

Thanks for making the naming changes, as well. Please go ahead and merge. We might have to think about auto-generating documentation while we expand the field dictionary....

[josephzhang8]

Thx @platipodium. I've merged and closed the PR.

[uturuncoglu]

@josephzhang8 Again. this was a draft PR and I supposed to make it ready to review and then merge it after testing. I think merging PRs without testing is not good practice. It is just my feeling. @platipodium We could change the field naming since field dictionary allow us to define aliases.

[platipodium]

I agree with Ufuk, first test it!. It was by accident that I converted from draft PR to PR ... always discovering new pitfalls when on GitHub

[uturuncoglu]

@josephzhang8 @platipodium Okay. I have already found an issue with this PR. If you try to compile ESMF interface without building model with USE_WW3 option. The model is failing. I think we need to add this CPP flag also to ESMF interface. Anyway, the fix will come soon.

[uturuncoglu]

@josephzhang8 @platipodium Okay. there are two solution for this issue.

1. We could remove USE_WW3 from the get_ww3_arrays routine in the model since compute_wave_force_lon has no that flag. But, this also related to the data structures. It seems that data structures are defined in src/Core/schism_glbl.F90 without using USE_WW3. So, it would be safe to remove it from the model.
2. In this solution, rather then removing USE_WW3 from the model, we could introduce it to ESMf interface around the SCHISM_StateImportWave3dVortex call.

Anyway, I just wonder about your preference. If you want to go with (2) then I think also introducing USE_WW3 to compute_wave_force_lon would be more consistent since that one also related with wave.

[josephzhang8]

Option 1 is fine. I added that CPP to make the code clear but it's not needed as you mentioned. Thx

[uturuncoglu]

@josephzhang8 Okay. Let me try. I'll update you about it.

[josephzhang8]

Sorry, I mixed this with another PR discussion.

[yunfangsun]

Hi @uturuncoglu,

For the mediator, I see https://github.com/oceanmodeling/CMEPS/blob/ed32f3ed12efc5fe224dc5305dfa27beec34ada6/mediator/esmFldsExchange_ufs_mod.F90 is updated, and https://github.com/oceanmodeling/CMEPS/blob/ed32f3ed12efc5fe224dc5305dfa27beec34ada6/mediator/esmFldsExchange_coastal_mod.F90 is not.

Are we using esmFldsExchange_ufs_mod.F90 or esmFldsExchange_coastal_mod.F90 for the coupling?

Thank you!

[uturuncoglu]

@yunfangsun I'll check it and let you know. I think I forgot to merge the CMEPS branch. Thanks for reminder.

[yunfangsun]

Hi @uturuncoglu ,

My /work2/noaa/nosofs/yunfangs/ufs-3d/ufs-weather-model_cdep_01052025/CMEPS-interface/CMEPS/mediator/esmFldsExchange_coastal_mod.F90 on Hercules is the most updated version.

Best

[uturuncoglu]

@yunfangsun Okay. It is done. Please that with the recent version of the UFS coastal. BTW, do not forget to add new PIO options for the WW3 (an example is in here - https://github.com/oceanmodeling/ufs-weather-model/blob/711f55369d880c32548465a2443b4f31bc468df8/tests/parm/ufs.configure.coastal_datm_ocn_wav.IN#L57). You might also need to create mod.def from scratch using ww3_grid tool. I did it for the RTs. Let me know if you have any issues.

[yunfangsun]

Hi @uturuncoglu

Thank you! I will try it

[yunfangsun]

Hi @uturuncoglu ,

My DOCN is located at /work2/noaa/nosofs/yunfangs/ufs-3d/ufs-weather-model_cdep_01052025/CDEPS-interface/CDEPS/docn . Could you please take a look at it? Thank you!

[uturuncoglu]

@yunfangsun did you modified docn_datamode_copyall_mod.F90?

[uturuncoglu]

@yunfangsun Okay. I found it. I have just pushed the mods to CDEPS and updated UFS Coastal. Could you try. I did not push the change related with the So_omask since it also breaks other tests. If you could have issue with it we could look at it.

[yunfangsun]

Hi @uturuncoglu ,

Thank you! I will try it