[Ncep.list.emc.fv3gfs_tickets] FV3GFS Ticket #21: update fv3gfs workflow to use the latest model with Write Component and Unified Post and to include NSST in 2d history output for DA
FV3GFS Trac Ticket
ncep.list.emc.fv3gfs_tickets at noaa.gov
Mon Sep 25 03:59:29 UTC 2017
#21: update fv3gfs workflow to use the latest model with Write Component and
Unified Post and to include NSST in 2d history output for DA
-----------------------------+--------------------------
Reporter: fanglin.yang@… | Owner: fanglin yang
Type: enhancement | Status: new
Priority: major | Milestone:
Component: component1 | Version:
Resolution: | Keywords:
-----------------------------+--------------------------
Comment (by fanglin.yang@…):
-------------------------------
'''Status of workflow development'''
-------------------------------
Branch
https://svnemc.ncep.noaa.gov/projects/fv3gfs/branches/fanglin.yang/tic21
can now be used for running following types of experiments using the fv3
branch
https://svnemc.ncep.noaa.gov/projects/fv3/branches/fanglinyang/tic26.
Users should check out tic26 from NEMSfv3gfs branch from
https://svnemc.ncep.noaa.gov/projects/nems/apps/NEMSfv3gfs/branches/jun/tic26.
Both the workflow and model code have been updated to its respective
trunk. Note that only Rocoto-based workflow is supported. The old
workflow based on pend/psub will not be maintained from this version
onward.
1) Cycled experiments with DA, either with or without NSST in the
forecast, with NSST turned off in analysis, still using nemsio_regrid for
remapping. (I have tested it on Luna for many cycles)
2) Forecast only experimen using the current NEMS GSM physics, either with
or without NSST, either using the Write Component or GFDL FMS for I/O
The Write Component is almost ready (thanks to Jun Wang). There are still
a couple of minor issues. I am working with Jun to resolve them before the
fv3 tic26 branch can be committed to the trunk.
The workflow has been extensively modified to include NSST, GFDL MP and
Write Component. The GFDL MP is yet to be tested. Please see
https://svnemc.ncep.noaa.gov/trac/fv3gfs/diff?new_path=%2Fbranches%2Ffanglin.yang%2Ftic21&old_path=%2Fbranches%2Ffanglin.yang%2Ftic21&new_rev=&old_rev=
for the detail of the changes.
---------------------
A few important items
---------------------
1) GMTED2010 terrain is set as the default
2) A new config.nsst has been added to set up NSST model. "NONST" is
added to config.base.
3) A new diag_table2 has been created to run with the Write Component.
Jun and Fanglin are still working on unifying field names used in the
model, POST and analysis.
4) Find in config.base
# I/O QUILTING, true--use Write Component; false--use GFDL FMS
# if quilting=true, chose OUTPUT_GRID as cubed_sphere_grid in netcdf or
gaussian_grid
# if gaussian_grid, set WRITE_NEMSIOFILE for nemsio or netcdf
export QUILTING=".true."
export OUTPUT_GRID="gaussian_grid"
export WRITE_NEMSIOFILE=".true."
5) set_bitmap has been added to downstream grib2 product generation
(thanks to Wen Meng)
6) Write Component needs ESMF lib esmf/7.1.0bs34
7) ush/drive_makeprepbufr.sh was updated by Xu Li to get NSST bufr data.
8) Forecast parameters in exglobal_forecast are set to those used in the
forecast-only experiment prfv3l65d (see Fanglin's recent presentation at
the FV3GFS weekly meeting).
9) GFDL_MP namelist parameters have been added to exglobal_fcst script
(Thanks to Rusty Benson).
...... and many other tweaks.
For model updates, please see
https://svnemc.ncep.noaa.gov/trac/fv3/ticket/26
--
Ticket URL: <https://svnemc.ncep.noaa.gov/trac/fv3gfs/ticket/21#comment:12>
fv3gfs <https://svnemc.ncep.noaa.gov/trac/fv3gfs>
NGGPS FV3GFS Development
More information about the Ncep.list.emc.fv3gfs_tickets
mailing list