In the traditional one­-way nesting, an error occurred ww3_shel.

[nYifeiLee]

Hi everyone！
My large scale grid has a resolution of 10′ and a small scale grid has a resolution of 6 ′.The computer I'm using is a virtual machine running 34.8GB of RAM.When I run the ww3_shel on a small scale grid, WW3 gets an error.In another previous simulation, WW3 also prompted this error due to my too fine meshing.I don't know if it's because my computer is running too little memory or something else that caused this error.
Below is a picture of the WW3 error prompting, along with my input file.I hope you can give me some advice, thank you!

ERROR:

Large scale grid:
ww3_grid.inp.txt
ww3_prnc.inp.txt
ww3_shel.inp.txt
Small scale grid：
ww3_grid.inp.txt
ww3_prnc.inp.txt
ww3_shel.inp.txt

[ukmo-ccbunney]

Hi @nYifeiLee
Is there any other output to the screen?
Sometimes the actual wave model error message can be hidden amongst a lot of those w3srervmd message lines...
MPI_ABORT is being called with error code 12, which could help to narrow down the internal error (there are a few modules that use that error code).

[nYifeiLee]

It's a pity that I didn't see a "*** WAVEWATCH III ERROR ... ".This screenshot is the content of "log.ww3".

The CPU of my computer is i7-8700 and it is too old.I'm going to run this model with a better computer.

[aliabdolali]

Just as a sanity check, run your smaller scale domain w/o boundary/nest file and see if it runs or not. If it runs, the problem is with thr boundary. Usually,you would see problems with boundary if you have open boundary nodes in shallow depth while they are not well captured in the large scale parent simulation.
The easy fix is to exclude those open boundary nodes from your small scale domainz

[nYifeiLee]

Dear aliabdolali
Thanks for the suggestion.I ran a small area with no nested borders and it worked correctly.So, as you predicted, the problem should arise at the boundary point.Could you please elaborate on the meaning of the term "open boundary nodes in shallow depth" so that I can better understand this issue.What does "open boundary nodes in shallow depth" refer to?
Below is a screenshot of the latest error report.

[aliabdolali]

well, I think the depths at open boundary nodes are not ~similar between the parent and child domains, due to the course resolution of the parent domain, so you can either exclude those points as open boundary nodes (which is a safe option, usually nothing gets to your domain from there and if it gets there, it is not representative) or do not use the spec files from those points from the large scale simulation.

[nYifeiLee]

I'd like to ask you how nested boundaries should be set up is prescriptive.I read the technical manual before, and my understanding was too superficial.I think that keeping the increments (DX, DY) of the large output boundaries consistent with the small range of resolutions will do the nesting smoothly.But so far, my understanding should be incorrect.I would like you to explain in detail how to set a large range of output boundaries and a small range of input boundaries to ensure that nesting goes smoothly.