troubles, could not find trapping x locations

[nlybarger]

Hello,
I am trying to run the WRF function real.exe on Derecho. I have produced WRF-I files using NCL scripts that I've had success with before (on Cheyenne), run geogrid.exe on my wanted domain, and run metgrid.exe on the WRF-I files, successfully producing met_em files. When I run real.exe on these met_em files, however, I am met with the following error:

FATAL CALLED FROM FILE: <stdin> LINE: 6241
troubles, could not find trapping x locations

Some quick googling leads me to believe this is due to some NaNs in the input data, but I'm not particularly familiar with parsing the rsl.error.0000 and rsl.out.0000 log files for pertinent information. I have attached these error files here. Please let me know if there's any more information I can provide you to help me out here. The only NaNs I see in rsl.out.0000 is the "target pressure and value" (not sure what this means).

I also see a strange entry in the log for "Using sfcprs3 to compute psfc using the new automatic levels program" where one of the output lines is:
25416.39 9097.176 49 31


Again, I'm not wildly familiar with troubleshooting WRF, so any help would be greatly appreciated! I've attached my rsl.error.0000, rsl.out.0000, namelist.input, and namelist.wps files.

Thanks,
Nick

 

[Ming Chen]

What data did you ungrib? Can you tell me where your met_em files are located at Derecho? Thanks.

 

[nlybarger]

I'm using NCL scripts that take in CESM2-LE data and convert to WRF-I files directly. I've done this in the past to run WRF on Cheyenne.
met_em files can be found at

/glade/derecho/scratch/nlybarger/data/CESM2-LE/cesm2_wrfi/LahontanValley/hist/1.accum/001_1171.009_19851217_19851220/WRF/run/

 

[Ming Chen]

Hi Nick,
I looked at almost all the variables in your met_em files, and they all look fine to me. Basically I don't think the input data is a concern for this case. Sorry that I haven't figured out yet what is wrong in the REAL program, but I am highly suspicious that the vertical level setting may not be reasonable. Would you please run the tests I list below:

(1) add the following options to namelist.input, then rerun this case. Please let me know whether it works.
max_dz = 1000.
auto_levels_opt = 2
dzbot = 50.
dzstretch_s = 1.3
dzstretch_u = 1.1
e_vert = 40

(2) if (1) doesn't work, please explicitly set the vertical levels (remember to delete auto_levels_opt, dzstretch_s etc):
e_vert = 40
eta_levels = 1.00000 , 0.99307 , 0.98348 , 0.97105 , 0.95551 ,

0.93651 , 0.91363 , 0.88644 , 0.85460 , 0.81855 ,

0.77877 , 0.73579 , 0.69016 , 0.64246 , 0.59329 ,

0.54573 , 0.50104 , 0.45908 , 0.41972 , 0.38281 ,

0.34824 , 0.31589 , 0.28563 , 0.25735 , 0.23096 ,

0.20635 , 0.18343 , 0.16209 , 0.14226 , 0.12384 ,

0.10677 , 0.09095 , 0.07633 , 0.06282 , 0.05036 ,

0.03889 , 0.02835 , 0.01868 , 0.00983 , 0.00000

Then rerun REAL.

Please keep me updated of the results.

By the way, when you run REAL using the same/similar met_em files as input in cheyenne, do you run the same WRF version? Can you find your namelist.input used in cheyenne?

Thanks.

 

[nlybarger]

Hello Ming,
I have rerun real.exe using the settings you suggested, but alas, it still fails with the same error without producing the wrfinput or wrfbdy files. I have attached to this response a tarball containing the namelist and rsl files corresponding to these attempts. Your option (1) above is labeled NEW_AUTO_LEVEL_SETTINGS and your option (2) is labeled EXPLICIT_LEVELS.

As for my previous success on Cheyenne, I have attached that namelist file outside the tarball. I believe it was run on WRF version 4.4.1.

Thanks for helping me out!
Nick

If it helps diagnose anything, the domains I'm trying to run are highly mountainous. Is it possible there's some issue defining pressure levels that are below topography?

 

[nlybarger]

Update:
I was able to get real.exe.to run by changing the levels to which the hybrid level data from CESM2 is interpolated to. That is, I defined different pressure levels for the WRF-I data, ranging from 850mb to 5mb. However, 850mb is rather high in the atmosphere for much of my domain, which spans from sea level, where surface pressure is ~1000mb, to mountainous terrain, where the surface pressure is ~750mb, so I would like to include pressure levels below 850mb.

 

[Ming Chen]

Hi Nick,
I am still looking at the issue. It will take some time. I will get back to you once I fix the issue. Thank you for your patience.

 

[nlybarger]

Hello,
I was finally able to get real.exe to run by adding the option
sfcp_to_sfcp = .true.
to my namelist.input domain options

 

[Ming Chen]

Hi, Nick,
Thank you for the update. I am glad it works.
However, I am a little confused by this case. The option sfcp_to_sfcp = .true. should be turned on when the met_em data only has surface pressure and terrain, but not sea level pressure. However, in your met_em file, I did find the variable "PMSL". And the values of PSFC (=94948.48) and
PMSL(= 101652.4) both look reasonable at the point (i=281, j=45), where the code crashed. So it remains mysterious for me why the option sfcp_to_sfcp = .true. should be turned on and why REAL works with this option.
Anyway, thank you so much for the update and I will continue to look at this issue. Hopefully I can find an explanation.