MOSART negative channel storage stopped v3.LR simulation #6622
Comments
wlin7
added a commit
that referenced
this issue
Sep 17, 2024
…#6623) Revert PR #6313 aborting run on negative channel storage The enforced abort whenever MOSART has negative channel storage may prevent v3.LR simulations from finishing, if the code base that includes to ensure new v3.LR simulations that need to use the latest commits to master are free from this issue. Fixes #6622. [BFB]
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The master code base that includes the changes via PR #6313, can cause v3.LR.piControl simulation to stop running due to negative channel storage that is no more tolerated. This was not an issue for the original v3.LR.piControl simulation with v3.0.0 tag.
Similar issue was also reported in the comment section of #6313.
For debugging, a branch run based off year 0416-01-01 of v3.LR.piControl can reproduce the issue in 8 months (occur on 0416-08-25), using master branch that has #6313 in place (make sure not being reverted)
Since its occurrence is unpredictable, and v3.LR simulation campaigns still have a high number of long simulations (single forcings over historical period, SSPs) to be completed with commits in the latest master branch (until v3.0.1 is tagged), we may have to revert #6313. The revert can then be reverted after v3.0.1 tag is created. And after that, a physics-based fix will still need to be implemented to either prevent the negative storage from occurring, apply a fixer, or make it more tolerant.
@peterdschwartz @hydrotian , please weigh in.
The text was updated successfully, but these errors were encountered: