Change SORRMv3 ocean initial condition #6758
Conversation
|
Here are results from 30 year B-cases: Baseline with previous ocean initial condition: Run with this ocean initial condition: Model vs. Model comparison: Time series of OHC anomaly, which shows smaller magnitude/drift with this ocean IC (black) vs. previous ocean IC (red): |
|
I'd like to request @maltrud document on this PR, however briefly, the method used to obtain these results, so that we have that information clustered with this application. |
|
will do ASAP |
|
@maltrud still waiting for docs. |
1 similar comment
|
@maltrud still waiting for docs. |
|
@maltrud still waiting for you to provide requested docs. |
|
here is an outline of how I generated the initial condition. |
|
Thanks @maltrud for that document. It's really helpful and the procedure makes sense. There's one part you don't mention that I'd appreciate having you add. How was the highly non-z-level vertical coordinate handled near and under ice shelves in both meshes? Did you treat it as if it were z-level? Or did it get some special treatment? |
|
Thanks, @maltrud! It all was very clear. I noticed that you said the vertical interpolation was linear; was the horizontal interpolation also (bi)linear? |
|
@cbegeman and @xylar thanks for your questions. I did nothing special to handle ice shelf cavities, just did "horizontal" extrapolation based on vertical index. my plan was to see if this initial method caused problems, and if necessary to modify the procedure to handle cavities, but it seems to have worked ok without special treatment. and, yes, the horizontal interpolation was bilinear. I'll add this info to the document. |
|
@maltrud, okay, I'm fine with that. While the vertical grids under ice shelves are pretty unrelated to the z-level grids in the open ocean (and thus the interpolation won't necessarily be very accurate), the flushing time for cavities is probably on the order of a few years even for the largest cavities so these differences shouldn't have a lasting impact. |
|
@cbegeman or @proteanplanet this still needs an approval from one of you. |
There was a problem hiding this comment.
Approving on the basis of @darincomeau's simulations, particularly the smaller drift in OHC, as well as @maltrud's methodological justification. Thanks @maltrud for all your work generating this IC!
|
updated document that describes the creation the new initial condition which incorporates answers to the questions raised by @cbegeman and @xylar . |
There was a problem hiding this comment.
Thank you for the documentation @maltrud and for the simulations @darincomeau proving the utility.
Change SORRMv3 ocean initial condition Changes ocean initial condition for B-cases with the SOwISC12to30E3r3 ocean mesh. The previous default initial condition is from a one-month G-case restart from cold-start (standard default). The initial condition here has been interpolated from year 701 of the SORRM v2.1 1950 control simulation, in an effort to bypass/speed-up ocean spinup. New ocean initial condition file is staged in public inputdata repo, world-readable. [BFB] for all currently tested configurations
|
Passes:
merged to next |
|
merged to master |
Changes ocean initial condition for B-cases with the
SOwISC12to30E3r3ocean mesh. The previous default initial condition is from a one-month G-case restart from cold-start (standard default). The initial condition here has been interpolated from year 701 of the SORRM v2.1 1950 control simulation, in an effort to bypass/speed-up ocean spinup.New ocean initial condition file is staged in public inputdata repo, world-readable.
[BFB] except
[non-BFB] for B-cases with the
SOwISC12to30E3r3ocean mesh (not in current testing).