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

Reset runoff accumulation

Hi All,

I am performing some long climate simulations and I need to reset the accumulated runoff at the beginning of each restarted run.
I was wondering if it is correct to put to zero these variables in the wrfrst files?

SFROFF:description = "SURFACE RUNOFF" ;

I looked at the code and, if I am not wrong, the accumulated runoff is always a left elements, namely it is newer used to compute other variables.
Thanks into advance for help,
Are you looking to reset the runoff in the hydro restart files, the land-surface model, the wrf model itself? Each has their own restart file.

If you are in fact using wrf-hydro, then in your hydro.namelist you can set the following variable:


This will set the accumulated variables in the hydro restart file to 0 after each restart.
The full description of the options in the hydro.namelist file can be found here.

If you're just looking to eliminate the runoff from the wrf simulation entirely, I would try to set SFROFF and UDROFF to 0 manually in the restart files and see if that sticks. If it does stick, you should not only see that in subsequent output files, but your simulation should diverge slowly from one using an unedited restart file. It is possible that the LSM will ignore those values in the wrf restart file and read them in from the LSM restart file instead, which may not be desired behavior in your case. I am a bit unfamiliar with restart file behavior in wrf-hydro.
Last edited:
Thanks for the quick reply.
I also had in mind to set SFROFF and UDROFF to 0 manually in the restart files, so I wonder if it is correct or affects the restarted simulations in some way?! Looking at the code it seems to me that SFROFF and UDROFF are only used as diagnostic variables, but a cross check here would be highly appreciated.
I'm sure someone with more knowledge would be able to answer that rather quickly, but the way to check yourself would be to run two parallel simulations, one where you modify the SFROFF and UDROFF variables as described and another where you don't. If they are purely diagnostic, then resetting them to 0 should have no impact at all on the simulation and you will get the same results (besides different values for SFROFF and UDROFF, of course). If the runs diverge then you know the SFROFF and UDROFF variables are being used to do prognostics.


I believe it is fine if you mannualy set sfroff and unroff to zero in wrfrst files. In this case the newly generated sfroff and unroff in your wrfout files will be the values from the restart time to the end time of this specific run.