Repository navigation
Set up a twin like RTOFS v2.5 #58
Description
-
NCODA configured SAME way.
-
Use (clean) initial conditions.
-
Prepare forcing fields that match those used in RTOFS v2.5.
-
Diagnose timing info on wcoss. Is MOM slow? Where? Which components?
-
Extend to using CICE ice/increment.
Further testing:
- Daily mean (archive) as background.
- FGAT. (HYCOM always runs with it ON; MOM6 there is/was some problem/development).
- No FGAT.
Activity
sanAkel commentedon Apr 28, 2025
Use this tool to decode fit to observations
sanAkel commentedon Jun 24, 2025
From @zgarraffo
Path to experiment on hera:
expt 14.1 /scratch2/NCEPDEV/marine/Zulema.Garraffo/FV3_RT/expt_14.1This is an experiment:
Restarts from RTOFS converted to MOM6 and CICE6:
/scratch2/NCEPDEV/marine/Zulema.Garraffo/FV3_RT/expt_14.1/rtofs.20250401_v2.5.aTaking restarts from ⬆ run (36 hours):
/scratch2/NCEPDEV/marine/Zulema.Garraffo/FV3_RT/expt_14.1/rtofs.20250401Next day (did not finish):
/scratch2/NCEPDEV/marine/Zulema.Garraffo/FV3_RT/expt_14.1/rtofs.20250402/Path source code:
/scratch4/NCEPDEV/marine/Zulema.Garraffo/ufs-weather-model_2025zgarraffo commentedon Jul 18, 2025
Expt 14.1 failed at day 7. Rerunning it with a shorter thermal time scale, but have to check the incremental update creation utility (possible cause of failure).
Started expt_14.2, similar to previous experiments for 2023: same 20250401 initial conditions than 14.1, instantaneous background, no ice assimilation, incremental update hours -6 to 0 from the analysis time, incremental update archive in the RTOFS-DA analysis levels.
After the 1st assimilation, expt_14.1 and 14.2 look almost identical, but noisier than the previous MOM6 experiments.
The initial forecast, before assimilation, 20250401, is similar to after assimilation (it was done with the default time steps, but will be chedked for DT_THERMAL=300. The hycom state is changed going to MOM, before assimilation
sanAkel commentedon Jul 19, 2025
@zgarraffo would you be able to plot differences between 14.2 and v2.5.a (parallel)? And post here?
zgarraffo commentedon Jul 22, 2025
-14.1 (daily mean background, -12 to 12 hours incremental update in layers, dt thermal=300s, day 20250407, noisy but didn't blow up
-14.2 (instantaneous background, -6 to 0 hours incremental update in z, dt thermal=300 day 20250407, similar noise.
-Hae-Cheol run a forecast only. Noticeable changes vs the initial rtofs archive in SSH off Brazil-Argentina in the forecast and in 14.1 and 14.2.
Will do some difference plots.
@zgarraffo
Can you please add a table of what's in each experiment? How they differ? And what's their status? A link to presentation/ plots?
sanAkel commentedon Oct 13, 2025
Results from @zgarraffo's experiments can be found here.
sanAkel commentedon Oct 29, 2025
@zgarraffo reports completion of the following plots:
She concludes: