Repository navigation
hybrid coordinate wrf #1148
Description
hybrid coordinate:
"It is mandatory to use the terrain following coordinate
system (hybrid_opt=0)"?
#404 (comment)
DART/models/wrf/tutorial/README.rst
Lines 21 to 28 in 80502a8
Lines 288 to 290 in 80502a8
Apparently the "It is mandatory to use the terrain following coordinate
system" is only when using WRF CV3 perturbation to generate the boundary condition perturbed ensemble.
Needs to be explained in the docs.
Came in in d85b7a0
part of #1030
Originally posted by @hkershaw-brown in #741 (comment)
Activity
braczka commentedon Jul 13, 2026
The statement in the documentation seems fairly clear, "Using the sigma hybrid coordinate can lead to adverse effects
when generating ensemble spread leading to poor forecast performance." However, the link to the DART Issue needs to be updated from Issue #650 to #807, where I performed a detailed analysis on the negative impact of using hybrid-coordinates with pert_wrf_bc. I will issue a documentation fix for this.
hkershaw-brown commentedon Jul 13, 2026
My question is on this statement:
"It is mandatory to use the terrain following coordinate system (hybrid_opt=0)"
Is it mandatory to use the terrain following coordinate with WRF in DART in general, or just mandatory when following the WRF-DART tutorial?
braczka commentedon Jul 13, 2026
It is mandatory when using the WRF-DART tutorial approach to perturb the boundary conditions -- which perturbs either the GFS or ECMWF state and applies it to the BC. If you have your own custom perturbed BC ensemble apriori (like Peter Jan) then you don't have to use this standard perturbation approach. The WRF-DART tutorial approach (for performing BC perts) is still how we recommend running WRF-DART in general -- the shell scripting used for the tutorial is identical to general WRF-DART. We only port in the severe storm information for the tutorial. In my experience the user having their own custom ensemble BC is not common.
hkershaw-brown commentedon Jul 13, 2026
Ok thanks, it would be good to have what is mandatory and when documented.
Currently the bash version of the scripting models/wrf/shell_scripts/gen_retro_icbc.sh enforces "hybrid_opt must be 0"
The models/wrf/shell_scripts_csh/gen_retro_icbc.csh does not seem to enforce this.
The docs appear to say that hybrid_opt 0 must be used for WRF-DART, but really this is just for when using a generated perturbation bank.
braczka commentedon Jul 13, 2026
I will provide a more detailed explanation in the tutorial docs, and also update the WRF model docs -- which are a bit out of sync with the WRF-DART tutorial