Ww3 gse2 #1143
Ww3 gse2 #1143
Conversation
|
@aronroland thank you for submitting this PR! As soon as the testing information from ERDC is added we'll start the tests and review. |
| JGS_BLOCK_GAUSS_SEIDEL = .TRUE. | ||
| JGS_USE_JACOBI = .TRUE. | ||
| JGS_LGSE = .FALSE. |
There was a problem hiding this comment.
Please add descriptions of the new terms to model/inp/ww3_grid.inp in the UNST namelist section: https://github.com/NOAA-EMC/WW3/blob/develop/model/inp/ww3_grid.inp#L318
There was a problem hiding this comment.
@aronroland once this is done I can try to do some tests of this feature, but will need additional information before I can do so. We'll wait to run regression tests once we have testing information included in the PR. Lastly, you checked the box for "If a new feature was added, a regression test for testing the new feature is added", but I don't see the new regression test included yet, so we'll need that as well before this can be merged. That being said, as soon as I have enough information to run a test on my end I can do that.
There was a problem hiding this comment.
Thanks for adding updates to ww3_grid.inp !
|
Hi Jessica,
as I see things we would like to proceed with the small global case. I have no time to work on the synthetic test case now but I promise to do that in the 2nd half of this year. We will inform u when the full pull request is ready for review.
Cheers
Aron
From: Jessica Meixner ***@***.***>
Sent: Mittwoch, 10. Januar 2024 17:33
To: NOAA-EMC/WW3 ***@***.***>
Cc: Aron Roland ***@***.***>; Mention ***@***.***>
Subject: Re: [NOAA-EMC/WW3] Ww3 gse2 (PR #1143)
@JessicaMeixner-NOAA commented on this pull request.
_____
In model/src/w3gridmd.F90 <#1143 (comment)> :
JGS_BLOCK_GAUSS_SEIDEL = .TRUE.
JGS_USE_JACOBI = .TRUE.
+ JGS_LGSE = .FALSE.
@aronroland <https://github.com/aronroland> once this is done I can try to do some tests of this feature, but will need additional information before I can do so. We'll wait to run regression tests once we have testing information included in the PR. Lastly, you checked the box for "If a new feature was added, a regression test for testing the new feature is added", but I don't see the new regression test included yet, so we'll need that as well before this can be merged. That being said, as soon as I have enough information to run a test on my end I can do that.
—
Reply to this email directly, view it on GitHub <#1143 (comment)> , or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB2S43WILPUJIW7KX2DFRELYN27FBAVCNFSM6AAAAABAXURM7OVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMYTQMJTGYZTQMZZGY> .
You are receiving this because you were mentioned. <https://github.com/notifications/beacon/AB2S43UE7EWZW6JFJJCQ4QTYN27FBA5CNFSM6AAAAABAXURM7OWGG33NNVSW45C7OR4XAZNRKB2WY3CSMVYXKZLTORJGK5TJMV32UY3PNVWWK3TUL5UWJTTMDHWPY.gif> Message ID: ***@***.*** ***@***.***> >
|
|
@aronroland Any updates on this? |
|
Dear @JessicaMeixner-NOAA, we are writing the paper now on the GSE alleviation method we have implemented in WW3. As for the pull requests we need init_bugs2 1st before this goes in ... |
|
@aronroland it seems like there are some outstanding issues with init_bugs2, can you help me understand why we cannot move forward with this PR? Is the other fixing a related bug? |
|
@thesser1 has commented on that and we r working actually on it. We should be soon done with the review of the changes. |
Pull Request Summary
Description
Solves the anisotropic diffusion equation following Booij and Holthuijsen (1987, hereafter denoted as BH87). We apply a simple Galerkin scheme for the non-diagonal terms and treat the diagonal terms using linear shape functions. The scheme has been tested in various global setups. We have further developed the filter functions introduced by Hendrik, e.g., we fade out the diffusion coefficients when waves reach shallow water. This is crucial, as we have much higher resolution near the coastlines within the multiscale approach. An implicit implementation could also be a future option, but this, from my point of view, is not fruitful. For the interested reader, the justification for this is given by Peter Janssen in his JCP paper. Why use higher order schemes and then degrade them by adding diffusion? To keep it short, if the resolution goes to lim -> 0, the diffusion coefficient goes to infinity, which renders this approach not really optimal. We must keep in mind that all of this is only done since the model is limited by the finite bandwidth approximation, which was chosen to discretize the wave spectra. For example, when increasing the spectral resolution, the GSE effect vanishes. In conclusion, we can say that all this is not needed if the spectral resolution is high enough, which was not affordable at the time when this was developed. Nowadays, we must reevaluate the whole approach.
Please also include the following information:
none
enhancement
No!
Issue(s) addressed
#930
Commit Message
adding gse alleviation for ug grids.
Check list
This needs further attention by Jessica, we need to chose, which test we want.
Testing
Global runs with significant GSE.
Not yet see above.
Will be done by ERDC
None
Will follow ...