CFL error and Segmentation Fault error

[Khin Myat]

Hi ,
I am trying to run the nested WRF simulation with the grid size of 1 km to 37m. However, during the simulation, the model crashed with the CFL error and segmentation fault error.
Although I tried to solve CFL error with the ways mentioned in the forum Segmentation Faults - Helpful Information ,the model still stopped with the CFL error in the innermost domain.
I tired to solve the problem with the following setting,
1.w_damping = 1,
2. epssm = 0.8
3. smooth_cg_topo = .true
4.ulimit -s unlimited
5.diff_6th_opt =2 , but it did not solve the problem.
What else can I try to solve the CFL error in this domain setup? Sometimes, the model stopped without CFL error. In that case, how should I consider or distinguish whether the model stop with CFL error or insufficient memory error?
Thank you in advance for your time. Any comments or advice are appreciated.

 

[Ming Chen]

Hi,

Please modify the options below:

(1) turn off cumulus scheme for all the domains, i.e.,
cu_physics = 0, 0, 0, 0
(2) turn off gravity wave drag,
gwd_opt = 0, 0, 0, 0
(3) time_step =6
(4) parent_time_step_ratio = 1, 3, 3, 3

Let me know whether these changes can make this case work. Thanks.

 

[Khin Myat]

Dear Ming Chen,

Thank you very much for your advice and I will try it again following your advice.
I will inform you again after I have tried.


Best,
Khin

 

[qianhe]

Hi， did you solve this problem？

 

[Khin Myat]

Dear @Ming Chen ,

I am really sorry for my late reply. Although I turned on the options you mentioned, the model still crashed at the beginning of the simulation.

Thank you in advance for your time.

Best regards,
Khin

 

[Khin Myat]

Hi， did you solve this problem？

I still could not solve the problem.

 

[Ming Chen]

I am sorry that these modified options didn't work.
The high resolution of 37m used in your case is definitely a concern. At this resolution, LES mode should be activated. While WRF works for ideal LES simulations, real-time LES simulation is much more complicated. For example, eddy structure may be permanently missing the lateral buffer zone of the 37-m domain while eddies develop in the interior of the domain. Such kind of inconsistency could lead to issues during the model integration.
WRF doesn't have specific codes to handle troubles for real-time LES simulations. Please refer to the literature and hope you can find some helpful information. The following paper applies WRF for LES simulation, yet the highest resolution used in this study is 111-m.

Real-time WRF large-eddy simulations to support uncrewed aircraft system (UAS) flight planning and operations during 2018 LAPSE-RATE, by James O. Pinto et al. (Real-time WRF large-eddy simulations to support uncrewed aircraft system (UAS) flight planning and operations during 2018 LAPSE-RATE)​

 

[Khin Myat]

Dear Ming Chen,

Thank you very much for your reply and information.
Have a good day.

Best,
Khin