SIGSEGV, Segmentation fault occurred while running WRF 4.4 for a nested, idealized tropical cyclone case

[6575]

Greetings all,

I'm trying to run a nested, idealized tropical cyclone case on WRFV4.4, but the error occured when the integration had begun for only 3 minutes.
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
wrf.exe 0000000003298AE5 Unknown Unknown Unknown
wrf.exe 0000000003296707 Unknown Unknown Unknown
wrf.exe 000000000322D3B4 Unknown Unknown Unknown
wrf.exe 000000000322D1C6 Unknown Unknown Unknown
wrf.exe 00000000031B5BA6 Unknown Unknown Unknown
wrf.exe 00000000031BC6F0 Unknown Unknown Unknown
libpthread.so.0 0000003F0F20F130 Unknown Unknown Unknown
wrf.exe 0000000002D86C92 Unknown Unknown Unknown
wrf.exe 0000000002D817B8 Unknown Unknown Unknown
wrf.exe 0000000002D7F5BB Unknown Unknown Unknown
wrf.exe 00000000025B7EE4 Unknown Unknown Unknown
wrf.exe 0000000001DF2E50 Unknown Unknown Unknown
wrf.exe 000000000164D974 Unknown Unknown Unknown
wrf.exe 0000000001440E18 Unknown Unknown Unknown
wrf.exe 000000000053F6F3 Unknown Unknown Unknown
wrf.exe 000000000053FCDC Unknown Unknown Unknown
wrf.exe 0000000000410BA1 Unknown Unknown Unknown
wrf.exe 0000000000410B5F Unknown Unknown Unknown
wrf.exe 0000000000410AFE Unknown Unknown Unknown
libc.so.6 0000003F0EE21B45 Unknown Unknown Unknown
wrf.exe 0000000000410A29 Unknown Unknown Unknown

I tried 'ulimit -s unlimited' but it did not work and there is no 'CFL' in the rsl files. The error won't occur if I only run one domain. Here is my namelist.

I need your help urgently and I appreciate you taking the time to read and try to sort this out for me. I am an undergraduate student majored in atmospheric science in China, so I think my English may be poor. Apologies for the inconvenience.

Thanks in advance.

 

[6575]

I have tried modifying my namelist for several times, but it did not work

 

[6575]

I try to make the model output every 3 minutes and I have found that the computation become very unstable (although there is no CFL error) , but reducing the timestep do not work. So how to solve the problem?

 

[kwerner]

Hi,
Can you package all of your rsl files into a single *.TAR file and attach that? Thanks!

 

[6575]

Hi,
Can you package all of your rsl files into a single *.TAR file and attach that? Thanks!

Hi,

There is no other information in the rsl files about the error so I did not save them. I ask a classmate for his namelist used in nested runs and copy the dynamic options into my namelist. It worked although there are still CFL errors after 5 days of integration because of the noise induced by lateral boundary condition and I finally solve this problem by using periodic boundary condition instead of open.

However, I still wonder what dynamic option of my original namelist (I have attached it previously) is not suitable. Here is my namelist which can successfully complete the simulation.

 

[kwerner]

I'm glad to hear you were able to get past the problem. It's difficult to say exactly what caused the issue. These are the variables from your original namelist that differed from those in the namelist that worked (some variables differed, but were set to the default, so if they weren't present in one of the namelists, they were set to the same values, regardless).

Original Namelist

&dynamics
 diff_opt            = 2,      2,      2,
 time_step_sound     = 6,6,
 mix_full_fields     = .true.,
 /

Working Namelist

&dynamics                          
 diff_opt      = 1,      1,
 diff_6th_opt  = 1,      1,
 gwd_opt       = 1,      0,
 /

If you're interested in the variable(s) that made the difference, you could run some different tests to try to narrow it down.