Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

Help Needed: Optimal Time Step for Model Stability

K

Kartik

New member
Hello everyone,

I am running the WRF model with the following configuration:
  • Horizontal Resolutions:
    • Domain 1: 9 km
    • Domain 2: 3 km
    • Domain 3: 1 km
  • Time Step Used: I tried 54 seconds initially and later reduced it to 32 seconds, but the model is still blowing up.
  • Vertical Levels: 54
  • parent_time_step_ratio = 1, 2, 2,

I have ensured my boundary and initial conditions are correct and that I am using GFS data at 6-hour intervals. Despite trying multiple time steps, the model still blows up.
I would appreciate advice on what time step I should try for this configuration or any additional changes to stabilize the model.

Thank you!
 
sure.. also explain how to calculate time step based on different resolution in the domain.
 

Attachments

  • files.zip
    431 KB · Views: 1
Last edited:
Bash: 
[workhorse@localhost files]$ grep -i FATAL rsl.*

grep -i error rsl.*

grep -i SIGSEGV rsl.*

grep -i cfl rsl.*
rsl.error.0118:-------------- FATAL CALLED ---------------
rsl.error.0118:FATAL CALLED FROM FILE:  <stdin>  LINE:     613
rsl.error.0129:-------------- FATAL CALLED ---------------
rsl.error.0129:FATAL CALLED FROM FILE:  <stdin>  LINE:     613
rsl.error.0130:-------------- FATAL CALLED ---------------
rsl.error.0130:FATAL CALLED FROM FILE:  <stdin>  LINE:     613
rsl.error.0131:-------------- FATAL CALLED ---------------
rsl.error.0131:FATAL CALLED FROM FILE:  <stdin>  LINE:     613
rsl.error.0140:-------------- FATAL CALLED ---------------
rsl.error.0140:FATAL CALLED FROM FILE:  <stdin>  LINE:     613
rsl.out.0118:-------------- FATAL CALLED ---------------
rsl.out.0118:FATAL CALLED FROM FILE:  <stdin>  LINE:     613
rsl.out.0129:-------------- FATAL CALLED ---------------
rsl.out.0129:FATAL CALLED FROM FILE:  <stdin>  LINE:     613
rsl.out.0130:-------------- FATAL CALLED ---------------
rsl.out.0130:FATAL CALLED FROM FILE:  <stdin>  LINE:     613
rsl.out.0131:-------------- FATAL CALLED ---------------
rsl.out.0131:FATAL CALLED FROM FILE:  <stdin>  LINE:     613
rsl.out.0140:-------------- FATAL CALLED ---------------
rsl.out.0140:FATAL CALLED FROM FILE:  <stdin>  LINE:     613
rsl.error.0118:forrtl: error (76): Abort trap signal
rsl.error.0129:forrtl: error (76): Abort trap signal
rsl.error.0130:forrtl: error (76): Abort trap signal
rsl.error.0131:forrtl: error (76): Abort trap signal
rsl.error.0140:forrtl: error (76): Abort trap signal
rsl.error.0130:d01 2024-07-08_00:00:32            4  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:00:32 hours
rsl.error.0130:d01 2024-07-08_00:00:32  MAX AT i,j,k:          288         288           3  vert_cfl,w,d(eta)=   2.471375     -0.2601430      5.0002337E-04
rsl.error.0130:d01 2024-07-08_00:00:32           34  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:00:32 hours
rsl.error.0130:d01 2024-07-08_00:00:32  MAX AT i,j,k:          288         286           3  vert_cfl,w,d(eta)=   7.450485      0.8678455      5.0002337E-04
rsl.error.0130:d01 2024-07-08_00:01:04          192  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.error.0130:d01 2024-07-08_00:01:04  MAX AT i,j,k:          288         288           3  vert_cfl,w,d(eta)=   181.7314      -40.44104      5.0002337E-04
rsl.error.0130:d01 2024-07-08_00:01:04          751  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.error.0130:d01 2024-07-08_00:01:04  MAX AT i,j,k:          288         286           3  vert_cfl,w,d(eta)=   7326.644       2389.802      5.0002337E-04
rsl.error.0130:d01 2024-07-08_00:01:04          270  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.error.0130:d01 2024-07-08_00:01:04  MAX AT i,j,k:          303         283           3  vert_cfl,w,d(eta)=   176.9564      -51.52967      5.0002337E-04
rsl.error.0131:d01 2024-07-08_00:01:04           13  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.error.0131:d01 2024-07-08_00:01:04  MAX AT i,j,k:          308         283           3  vert_cfl,w,d(eta)=   3.831510     -0.1998033      5.0002337E-04
rsl.error.0131:d01 2024-07-08_00:01:04          109  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.error.0131:d01 2024-07-08_00:01:04  MAX AT i,j,k:          307         283           3  vert_cfl,w,d(eta)=   16.18935      -6.973195      5.0002337E-04
rsl.error.0140:d01 2024-07-08_00:01:04           14  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.error.0140:d01 2024-07-08_00:01:04  MAX AT i,j,k:          239         316           3  vert_cfl,w,d(eta)=   3.548797     -0.3824556      5.0002337E-04
rsl.error.0140:d01 2024-07-08_00:01:04          165  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.error.0140:d01 2024-07-08_00:01:04  MAX AT i,j,k:          239         314           3  vert_cfl,w,d(eta)=   25.92192       2.994892      5.0002337E-04
rsl.out.0130:d01 2024-07-08_00:00:32            4  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:00:32 hours
rsl.out.0130:d01 2024-07-08_00:00:32  MAX AT i,j,k:          288         288           3  vert_cfl,w,d(eta)=   2.471375     -0.2601430      5.0002337E-04
rsl.out.0130:d01 2024-07-08_00:00:32           34  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:00:32 hours
rsl.out.0130:d01 2024-07-08_00:00:32  MAX AT i,j,k:          288         286           3  vert_cfl,w,d(eta)=   7.450485      0.8678455      5.0002337E-04
rsl.out.0130:d01 2024-07-08_00:01:04          192  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.out.0130:d01 2024-07-08_00:01:04  MAX AT i,j,k:          288         288           3  vert_cfl,w,d(eta)=   181.7314      -40.44104      5.0002337E-04
rsl.out.0130:d01 2024-07-08_00:01:04          751  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.out.0130:d01 2024-07-08_00:01:04  MAX AT i,j,k:          288         286           3  vert_cfl,w,d(eta)=   7326.644       2389.802      5.0002337E-04
rsl.out.0130:d01 2024-07-08_00:01:04          270  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.out.0130:d01 2024-07-08_00:01:04  MAX AT i,j,k:          303         283           3  vert_cfl,w,d(eta)=   176.9564      -51.52967      5.0002337E-04
rsl.out.0131:d01 2024-07-08_00:01:04           13  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.out.0131:d01 2024-07-08_00:01:04  MAX AT i,j,k:          308         283           3  vert_cfl,w,d(eta)=   3.831510     -0.1998033      5.0002337E-04
rsl.out.0131:d01 2024-07-08_00:01:04          109  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.out.0131:d01 2024-07-08_00:01:04  MAX AT i,j,k:          307         283           3  vert_cfl,w,d(eta)=   16.18935      -6.973195      5.0002337E-04
rsl.out.0140:d01 2024-07-08_00:01:04           14  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.out.0140:d01 2024-07-08_00:01:04  MAX AT i,j,k:          239         316           3  vert_cfl,w,d(eta)=   3.548797     -0.3824556      5.0002337E-04
rsl.out.0140:d01 2024-07-08_00:01:04          165  points exceeded cfl=2 in domain d01 at time 2024-07-08_00:01:04 hours
rsl.out.0140:d01 2024-07-08_00:01:04  MAX AT i,j,k:          239         314           3  vert_cfl,w,d(eta)=   25.92192       2.994892      5.0002337E-04

So looking at your rsl logs you are getting a lot of CFL violations.

This post by the admins can explain how to fix that.
forum.mmm.ucar.edu

Segmentation Faults and CFL Errors

Segmentation faults can be difficult to track down. As there isn't usually a clear error message, it can take some trial and error to figure out the problem. 1) A segmentation fault is often the result of using too many or too few processors, or a bad decomposition. Take a look at this FAQ...
forum.mmm.ucar.edu forum.mmm.ucar.edu


As for your second question time steps have a maximum of 6*dx. So for 9km you will have 6*9=54 or somewhere near or around that. Each nest takes the domain ratio and goes smaller. So for easy example say time step was 45s for d01 then d02 would be 15s and d03 would be 5s based on the 1:3 ratio.
 
i am keeping the time step as 45, 15, 5 this time but model still blows up. can you tell me what combinations should i try to fix this
 
Time step that is required for the model to be numerically stable is case dependent. It is not possible to tell; sometimes 3*dx is not short enough, sometimes 20*dx will still work.

parent_time_step_ratio = 1, 2, 2,
That does not work in favor of numerical stability. This means, that nest will have only 2x smaller time step than its parent, while grid dx is 3 times smaller.

You should set at least:
parent_time_step_ratio = 1, 3, 3,

or even more like:
parent_time_step_ratio = 1, 4, 4,

if your runs crash due to instability in nests (which is usually the case because of larger topography slopes there compared to parent domain).

Edit: although from Will's answer I see that in your case it seems to be in d01 so its time step should be reduced. Also, if it is not already, w_damping should be = 1, and epssm should be 0.5 or higher. I hope that helps.... :)
 
Kartik said:
i am keeping the time step as 45, 15, 5 this time but model still blows up. can you tell me what combinations should i try to fix this
Click to expand...
Apologies for the long delay in response while our team tended to time-sensitive obligations. Thank you for your patience. If you are still experiencing this issue, will you please attach your most recent namelist.input file, as well as the rsl* files (you can package those together into a single *.tar file, and attach that). Thanks!
 
You must log in or register to reply here.
Top