Error while reading namelist dynamics

[francescomaicu]

HI everyone,
my issue actually concerns the running of WRF3.5.1, starting from a namelist.input that I previously used to run WRF4.2.
I was not sure it could work, since new modifications occurred between the two releases.
Firstly I deleted the Physics_suite CONUS, non implemented in 3.5.1 .
After several aborted runs, following the indications of the error file, I added the requested empty namelists like &dfi_control, &tc, &scm ..... (never requested in 4.2 version), but still WRF does not run and returns the error :

------ ERROR while reading namelist dynamics ------
-------------- FATAL CALLED ---------------
FATAL CALLED FROM FILE: <stdin> LINE: 9651
ERRORS while reading one or more namelists from namelist.input.
-------------------------------------------
Abort(1) on node 0 (rank 0 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0

I attached also the namelist.input file View attachment namelist.input and the namelist.outputView attachment namelist.output.txt

Thank you in advance

 

[Ming Chen]

Please delete the two lines below:

hybrid_opt = 2,
gwd_opt = 0,

then try again.

 

[francescomaicu]

Oh thank you so much, now it runs....only for few minutes of the simulation time.
I see in the Users Manual that the introduction of the hybrid vertical coordinates is one of the most important improvement of the versions 4.* , and I did not realized that before.
This means that probably the input file generated by the program real.exe using the vertical hybrid coordinates , are not completely suitable for WRF 3.5.1.
This should explain the crash of the simulation with error message in the attached file.
Am I correct?
Actually I was trying to run the simulation with WRF3.5.1 using the domain and boundary conditions generated with the 4.1 suite (WPS+WRF).
I understand the issue related to the boundary condition, I'm wondering now whether or not to restart from the beginning of the simulation chain.
I also posted a related topic here :
https://forum.mmm.ucar.edu/phpBB3/viewtopic.php?f=30&t=9793
Thank you so much
F

 

[francescomaicu]

Hi all,
I think this is a cross-cutting issue for my WRF/WPS 3.5.1 installation.
I made a test running WRF with previous boundary conditions generated with real.exe (with hybrid vertical coordiantes, versione from WRF4.2).
WRF crashes after few minutes for physical reasons I guess (see attached rsl.error.0002.wrf), but still wrf is not able to open and write the output with the initial condition and some time records (rsl.out.0000.wrf). I only get a file core.3301 sizing 7GB.

So I tried to avoid the crash re-generating the boundary conditions with the program real.exe.
Same issue: real is not able to write to wrfinp file.

Bachward on the modelling chain also metgrid has the same issue.
Here it is documented: https://forum.mmm.ucar.edu/phpBB3/viewtopic.php?f=30&t=9793#p19367 .

So the point is that recently I compiled both WRF4.2 and WRF3.5.1 .
The first works , the latter no.
Both codes were successfully compiled.
I attached the configure and logs.
Guess it could be a problem of netcdf libraries (https://forum.mmm.ucar.edu/phpBB3/viewtopic.php?f=39&t=9416)?....look the .bashrc with the loaded modules.
thanks a lot

 

[francescomaicu]

Dear all,
this issue has been solved.
PLease take a look to the related post:
https://forum.mmm.ucar.edu/phpBB3/viewtopic.php?f=30&t=9793#p19376
Bye