MITgcm
Yoshi Nakayama code for tangential component of tides
What changes does this PR introduce?
Feature: This is Yoshi Nakayama's code for adding tangential component of tides at obcs.
What is the current behaviour?
Currently only the normal component of tides at open boundaries is allowed.
What is the new behaviour
This adds the tangential component.
Does this PR introduce a breaking change?
I think the only potential break would be when the user is currently using obcs as control variables (in adjoint mode) and also using tides. If we really want to be sure, we can add an ifdef flag into OBCS_OPTIONS.h to enable this tangential component only when the user chooses to. However, I do not really know because I do not run the adjoint with obcs as control, and I do not know of any user who is currently using both obcs as ctrl and using tides.
Other information:
This piece of code has been distributed to several users who run regional domains and want to apply tides. I'd like it to be in the main branch and am willing to do the needed testing to get it in if possible, so that we don't have to privately exchange the code and constantly needing to bring it up-to-date with every new check point.
This looks like a great addition to me. It’s hard to see how the existing tidal boundary conditions give good tides except for f=0.
Is there a paper to cite for this change?
@antnguyen13 any chance there is a simple a setup that uses this? Its the sort of thing where a simple test would be useful to check and make sure it behaves as expected etc....
Also - @jrscott maybe you could help get the OBNAmtFile file parameters etc.... into http://mitgcm.readthedocs.io/en/latest/phys_pkgs/obcs.html?
Thanks,
Chris
@jrscott a reminder!!!
@christophernhill Do you know of a verification experiment that uses obcs and tides (and read in files at obcs)? Then I can add this to it.
I'm currently using it for my high-res regional set up, and both Yoshi @menemenlis and if i'm correct @mmazloff are using it for their setups..
OK - I was afraid you would ask that. I think we only have 2d examples at moment, which is not so useful for tangential terms! @jm-c and I will chat about this.
@menemenlis could we create a little test experiment for this from one of the cutout setups?
@antnguyen13 I will get together with Dimitris about this next week.
P.S. @mmazloff and @menemenlis we are still looking for a test setup for this.....
P.S. @mmazloff and @menemenlis we are still looking for a test setup for this.....
P.S. @mmazloff and @menemenlis we are still looking for a test setup for this.....
Hi Chris
Could we add it to one of the current setups? If so, which would you recommend? Options may be: dome, exp4, internal_wave, obcs_ctrl, seaice_obcs, so_box_biogeo, tutorial_plume_on_slope
@mmazloff not sure!
Either obcs_ctrl or dome seems like best choices to me. obcs_ctrl currently only runs tests within adjoint cycle. What do you prefer?
Alternatively we could add something based on present California work, if that was easier for you? It might be useful to have something for California region in general?
Thoughts on others
internal_wave and tutorial_plume_on_slope are 2d, so not so good I think?
exp4, so_box_biogeo very large resolution, not sure it makes sense?
seaice_obcs has sea-ice that adds complexity.
If you give me a couple of weeks, I could come up with something coarse and idealized. i.e. a box with an oblong bump in the middle. I think at some point I promised to demonstrate that the tangential component is important, so that could be the demonstration case. Are python init scripts OK?
Sounds good to me! Thanks, Dimitris
On Sep 18, 2018, at 8:50 AM, Jody Klymak [email protected] wrote: If you give me a couple of weeks, I could come up with something coarse and idealized. i.e. a box with an oblong bump in the middle. I think at some point I promised to demonstrate that the tangential component is important, so that could be the demonstration case. Are python init scripts OK?
@jklymak that would be great, python init is great. Be great to have some tides related reference experiments that are connected to your work.
Yes, we discussed with Jean-Michel in Austin and I promised to add something that tests tangential component of tides under seaice_obcs I plan to replace input.tides/output.tides.txt to include tangential component.
Dimitris Menemenlis
On Nov 5, 2018, at 12:00 PM, christophernhill [email protected] wrote:
@jm-c https://github.com/jm-c should we add this to seaice_ibcs
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/MITgcm/MITgcm/pull/73#issuecomment-436014979, or mute the thread https://github.com/notifications/unsubscribe-auth/AE_xSQpwxTUUDiTCNexOuFUEpcV5VBXMks5usJjigaJpZM4S5o6e.
My Thanksgiving project ...
Dimitris Menemenlis
On Nov 19, 2018, at 12:53 PM, christophernhill [email protected] wrote:
@menemenlis https://github.com/menemenlis any news on this
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/MITgcm/MITgcm/pull/73#issuecomment-440038098, or mute the thread https://github.com/notifications/unsubscribe-auth/AE_xSREOQM-aRSxFswxWX3QWWm1CsJBiks5uwxpggaJpZM4S5o6e.
@menemenlis everybody (including Matt) is waiting for your addition to experiment seaice_obcs/input.tides so that this code could finally be merged.
@antnguyen13 I tried to run verification experiment "seaice_obcs", but the secondary test (using "input.tides") did not run. It's a bit unfortunate that this set of changes does not allow previous set-up with useOBCStides=T to just run as they use to, without having to change "data.obcs". Otherwise, I saw how these tidal amplitude and phase are generated for this test (seaice_obcs.tides) so I could just generate a new set of files for the Tangential component of the flow.
@antnguyen13 I updated the OB test with tides to also test this new feature. @christophernhill waiting for your review now. Also, I am going to open an issue to list few things that can be improved in this OB tides code, but this would be for an other PR.