Increasing time intervals in output file

[Setareh Rahimi]

Dear all,

I am running WRF-Chem V.4.4 using chem-opt = 202 (MOZART-MOZAIC). Once checked the rsl.error.0000 file, I noticed that the model gives output for each second and this ways, it takes a long time to finis the simulation.

*******************************************************************************************

Timing for main: time 2021-01-14_00:00:01 on domain   3:  130.27022 elapsed seconds
Timing for main: time 2021-01-14_00:00:02 on domain   3:   40.59900 elapsed seconds
Timing for main: time 2021-01-14_00:00:03 on domain   3:   18.22544 elapsed seconds
Timing for main: time 2021-01-14_00:00:03 on domain   2:  349.17075 elapsed seconds
Timing for main: time 2021-01-14_00:00:04 on domain   3:   18.71079 elapsed seconds
Timing for main: time 2021-01-14_00:00:05 on domain   3:   18.93999 elapsed seconds
Timing for main: time 2021-01-14_00:00:06 on domain   3:   18.76395 elapsed seconds
Timing for main: time 2021-01-14_00:00:06 on domain   2:   72.12794 elapsed seconds
Timing for main: time 2021-01-14_00:00:07 on domain   3:   17.82852 elapsed seconds
Timing for main: time 2021-01-14_00:00:08 on domain   3:   19.16705 elapsed seconds
Timing for main: time 2021-01-14_00:00:09 on domain   3:   18.43423 elapsed seconds
Timing for main: time 2021-01-14_00:00:09 on domain   2:   72.55254 elapsed seconds
Timing for main: time 2021-01-14_00:00:09 on domain   1:  663.78632 elapsed seconds
Timing for main: time 2021-01-14_00:00:10 on domain   3:   17.60782 elapsed seconds
Timing for main: time 2021-01-14_00:00:11 on domain   3:   18.83975 elapsed seconds
                                                                                               

 ******************************************************************************************                                                                     

So how can I increase time intervals, to finish the simulation in a reasonable time period. I increased time step up to 54s, but after a while the model stopped. I also attached my name-list.

Many thanks in advance.

Kind regards,

[Gabriele Pfister]

Hi Rahimi

you have a timestep of 9 sec for d01 and a ratio of 3 for d02 and d03. This means your inner domain timestep is 1 sec. 

However, why did you set a timestep of 9 sec for d01 when the resolution is dx=9 km? The general recommendation is setting a timestep that is 6 times dx (i.e. 54 sec). The timestep only needs to be reduced when your run fails,  e.g. experiencing CFL errors. 

Gabi

--

================================
Gabriele Pfister

Atmospheric Chemistry Observations & Modeling

National Center for Atmospheric Research (NCAR)

Email: pfi...@ucar.edu

Temporary Home Office Phone: +1 303 241 1329

Work Phone: +1 303 497 2915

Web: https://staff.ucar.edu/users/pfister

[Setareh Rahimi]

Dear Gabbriele,

Thanks for your response. I already set 54s for the time step, but  after a short time the simulation stopped with no error messages. I also set it in the 30s, 20s and got the same result.

So what is your recommendation in this case?

Kind regards,

--

S.Rahimi

[Gabriele Pfister]

First, set your debug level higher so that you get more information on why the run failed. Look for CFL errors in the log files. 

Does it run with 9 sec? If so, then run maybe for a day and then switch to a higher time step. it could be that your setup is not stable and needs a period to spin up. 

Gabi