Segmentation fault - points exceeded cfl=2

[Andrea-ARPAPUGLIA]

Dear all,

my goal is to execute WRF model ver. 4.2 on 2 nested domains with resolution respectively of 5000, 1000 (See namelist.input attacched) using analysis ECMWF as input global model for 181 houres (144 CPUs).

After about 27 hours of run, the execution stops with error "Program received signal SIGSEGV: Segmentation fault - invalid memory reference" and same rls.error files reports the message "points exceeded cfl=2" (see files attacched).

Following your suggestions I edited:
-time_step value;
- smooth_cg_topo=.true.
- epssm = 0.2
- w_damping = 1

Do you have any suggestions on what other parameter I can change to avoid the same error?

Thank you so much for your support,
Andrea

 

[Ming Chen]

Hi,
I take a look at your nameless.input. Below are a few suggestions:
(1) please turn off adaptive time step (use_adaptive_time_step = .false.)
(2) set time_step = 25
(3) run with RRTMG radiation (ra_lw_physics = 4, 4, ra_sw_physics = 4, 4)
(4) radt = 5, 5
feedback =1
(5) diff_opt = 2, 2
km_opt = 4, 4
w_damping = 1
Epssm = 0.6, 0.6

if possible, can you also turn off grid nudging?

Please try and let me know whether this case can work.

 

[Andrea-ARPAPUGLIA]

Thank you very much. Your suggestion was decisive even if the execution time has more than tripled (was not necessary turn off grid nudging).

Can you tell me what could be the difference in output due to having used RRTMG radiation?

 

[Ming Chen]

RRTMG is a more advanced radiation scheme. It considers radiative impact of ozone and has climatological aerosol input options. This is why I recommend using RRTMG.

 

[Andrea-ARPAPUGLIA]

Thanks a lot,
Andrea