Fields output for consecutive time step and/or time derivatives of prognostic variables

[Kazim Hussain]

Dear members,
Is it possible to output in a single WRF files which I save let's say 1 hour apart, some fields at just the previous model time step along with the current time step which obviously is outputted generally. I need these in order to work with the time derivatives of certain quantities like the prognostic variables over the entire domain which are of immense importance to me.

 

[Ming Chen]

If I understand your question correctly, I. guess you can use the run-time IO option to output variables you are interested.
Please see details here how to use this option.

 

[Kazim Hussain]

Thank you for the response. The link here seems to be to this same page or maybe it is broken for me. Anyway I have some question on the available fields in the registry EM Common file related to this. The number of time levels for the prognostic variables is two. But they have the same name, so in the run-time IO text file I add what to get the two different time levels? something like +:h:0:U_1,U_2 and so on? I ask because they are not named differently in the registry files it's just that their NumTLevs=2 but in the history there is just one time level. Also I want to know if these two time levels are separated by 1 dt or are they from some intermediate step in the Runge Kutta time stepping?

 

[Ming Chen]

u_1 is the U component of wind at the previous time step, and u_2 is the U component of wind at the current time step. These are the variables used to create restart files. They are separated by 1 time step (dt).

You can simply output 'u', which is x-wind component. The same for 'v', the y wind component.

 

[Kazim Hussain]

Thank you for your response, However I still did not manage to output U_1 or any prognostic variable on the first time level. Adding simply 'u','v' (by adding +:h:0:u,v and so on in iofields file) does not generate the prognostic variables on two levels in default history file like they are generated in restart files. I want to know if I can add a line in the iofields file that may output these variables on both time levels in the history file.
Thank you very much for your help, much appreciated.

 

[Ming Chen]

You cannot output 2-time-level variables (e.g., U_1, U_2, V_1, V_2 etc.) because they are not defined as state variables. Instead, you can only output these variables at the specified model time.
Is there any special reason you want to have variables at two consecutive time steps?

 

[Kazim Hussain]

Yes, I need the time derivatives of the prognostic variables as calculated by the model for my study. I don't know how else I can have it, I also can't get the restart file for my high resolution domains, WRF crashes when it has to generate it I think because of memory requirements. Is there any way for me to get these time derivatives then?

 

[Ming Chen]

Unfortunately WRF has no capability to output 2-time-level variables. I will talk to our software engineer and see whether there is an easy way to modify the codes to output such variables.

 

[islas]

I would recommend looking at share/output_wrf.F and seeing how to make the regular output mimic when switch .EQ. restart_only for the particular fields of interest. That said, as the desire is to have something similar to the restart file it may be easier to try to resolve your issues with restart files and use those in lieu of output instead of modifying output to have similar multiple time step writes of variables.