How to output history at 0 hour

[YongjunZHENG]

Description

Data assimilation updates u and v in the RESTART file *.fv_core.res.tile[1-6].nc. Since u and v are grid-relative wind, to check the corresponding zonal and meridional wind, I set model_configure as follows (output_fh=0 1) to ouput history.tile[1-6].nc which contain zonal (ugrd) and meridional (vgrd) wind.

model_configure:
nfhout: 1
nfhmax_hf: 0
nfhout_hf: 1
nsout: 0
output_fh: 0 1

But ncdum -v time history.tile1.nc show that there is NO 0 hour output

Output

ncdum -v time history.tile1.nc

data:

 time = 1 ;

[junwang-noaa]

Is this a cold start run or IAU run? If it is a cold start run, model does not output purely fh=000 history files since no integration has been done; model will output the first time step history files and mark it as fh000. You can set:

output_fh: $first_timestep_in_hour 0 1

$first_timestep_in_hour is a real number.

[YongjunZHENG]

Is this a cold start run or IAU run? If it is a cold start run, model does not output purely fh=000 history files since no integration has been done; model will output the first time step history files and mark it as fh000. You can set:

output_fh: $first_timestep_in_hour 0 1

$first_timestep_in_hour is a real number.

Thank you for the solution. Even only one step forward, the wind has been updated which is not what I want. Is there a subroutine in the source code of UFS for rotating the grid-relative wind to the corresponding zonal and meridional wind?

[gspetro-NOAA]

Hi @YongjunZHENG ,

You might have more luck asking your question about the specific subroutine in the fv3atm repository or the atmos_cubed_sphere repository, since the ufs-weather-model is just the umbrella repo that assembles the source code for various components, including the atmospheric component. @jkbk2004 @BrianCurtis-NOAA Any suggestions for where this subroutine would be located and if it exists?

You could also try running your case with an offset--i.e., prepare initial conditions for a slightly earlier time, and then check at the current f000 time for those values. Otherwise, it seems like the f000 values for u- and v-wind should just be included in the initial conditions you're using since no integration has been done at f000, as @junwang-noaa mentioned.

I will convert this issue to a Discussion since it is a question, not a bug report or feature request.

Best,
Gillian Petro

[jkbk2004]

@gspetro-NOAA @YongjunZHENG what about https://github.com/NOAA-GFDL/GFDL_atmos_cubed_sphere/blob/4f1a5ef4b83461a79a62c8d965ebc2b26cf5eba8/tools/test_cases.F90 ?