About changing physic suite on restart case

[duydoanmanh]

Hi, I'm Duy,
I'm planning a real experiment to examine how different physics suites affect simulation outcomes, especially when the target event is sensitive to physics choices.
Here’s the approach:

1. Run a CONTROL case using the original physics suite that successfully produces the event (which occurs around +18 hours from the initial time). Save the wrfrst file for restarting.
2. At +17h (or +17:30h — just before the event), restart the simulation using a different physics suite.

This setup is technically feasible, but I have a concern:
Will this restart behave like a warm-start run, meaning I can skip the spin-up period?
If not, what would be the best way to create a fair environment to compare different physics suites (beyond relying solely on idealized test cases)?

 

[kwerner]

Hi,
Apologies for the delay in response and thank you for your patience. The restart files will serve in place of the wrfinput* files, so you will not need to run another spin-up period. The only thing to be careful about is that if you are changing to an LSM that uses a different number of soil levels, you'll need to change that setting in the namelist.

 

[duydoanmanh]

Hi,
Apologies for the delay in response and thank you for your patience. The restart files will serve in place of the wrfinput* files, so you will not need to run another spin-up period. The only thing to be careful about is that if you are changing to an LSM that uses a different number of soil levels, you'll need to change that setting in the namelist.

Thank you for your reply.

I'd like to clarify my idea. During the initial 18-hour spin-up period, the model is run using a stable microphysics (MP) scheme (WSM5) to allow the storm to develop. A restart file is written at +18:00, and the simulation is then restarted using a different MP scheme (WSM7).

The issue I’m concerned about is whether, within about one hour after the restart, the model can reach a new stable equilibrium state (e.g., transitioning from 5 to 7 hydrometeor phases)?

 

[kwerner]

Thanks for clarifying. I'm not sure about that. Did you try testing to see?

 

[ztx]

Thank you for your reply.

I'd like to clarify my idea. During the initial 18-hour spin-up period, the model is run using a stable microphysics (MP) scheme (WSM5) to allow the storm to develop. A restart file is written at +18:00, and the simulation is then restarted using a different MP scheme (WSM7).

The issue I’m concerned about is whether, within about one hour after the restart, the model can reach a new stable equilibrium state (e.g., transitioning from 5 to 7 hydrometeor phases)?

In some mp_physics code, the scheme will only check weather the input conforms to physical laws (for example, the QRAIN is very large while the QNRAIN is very small, this will cause the rain drop very very big, which is not comforms to physical laws). But I'm not sure how WSM7 deal with that. Maybe read the code of the mp_physics scheme may help you.