WRF completing prematurely with adaptive time stepping

[jamesrup]

Hi all,

I had an issue with WRF completing ("successfully") long before the specified end time in the namelist (and corroborated in namelist.output) and found
this thread on the matter WRF Stops Prematurely if adaptive time step is used (RESOLVED)

I tried applying that fix listed there and it didn't change the problem for me. I'm using WRFv4.5.2

Is there another thread I might be missing that's still open on this issue, describing some other fix?

Cheers,
James

 

[kwerner]

Hi James,
I'm not aware of any other open issues on this. First, as a test, can you try with WRFV4.6.1 just to see if it makes any difference. I don't really expect it to, so if the same thing happens, will you attach your namelist.input file, as well as the run/error logs (if you have several rsl files, you can package them together in a single *.tar file and attach that). Can you also share your wrfbdy/wrfinput files so I can test with your input? Those files will most likely be too large to attach, so see the home page of this forum for instructions on sharing large files. Thanks!

 

[jamesrup]

Hi kwerner,

No problem. One point first - I forgot to mention that this is with WRF-ideal, specifically em_convrad. It seems to not be a problem with em_real, as I use that option for another simulation setup with no problem.

I've now repeated the same problem with WRFv4.6.1. I'm attaching namelist.input here along with rslout.tar.gz since it's small. I uploaded wrf_input for this idealized run under the name ruppert_wrfinput.tar.gz

Thanks for your help!

P.S. I'll check for other threads on this but I also just tried compiling em_real with the other project I mentioned using the latest WRFv4.6.1, which has a bug fix (reflectivity output) that I want to take advantage of. For some reason it's failing to compile using config option #50 on Derecho, even though that worked for em_convrad, and with WRFv4.6.0 worked for em_real.

 

[kwerner]

Hi,
I want to let you know I've been able to repeat the issue you're seeing, but it's going to take me a bit of time to dig into the code. I assume the simulation would run too slowly if you didn't use the adaptive time-step option?

 

[jamesrup]

Hi kwerner, so I could take the approach of learning from adaptive test runs what an ideal time step is...that could be a viable approach for now. But yes ideally I'd like to be able to rely on the internal adaptive_dt system for determining this based on what's happening in the simulation. Thanks for your help checking into this!