[Ncep.list.emc.fv3gfs_tickets] FV3GFS Ticket #20: Update global_cycle to run on FV3 cubed-sphere grid

FV3GFS Trac Ticket ncep.list.emc.fv3gfs_tickets at noaa.gov
Fri Aug 25 17:43:44 UTC 2017


#20: Update global_cycle to run on FV3 cubed-sphere grid
-----------------------------+----------------------------
  Reporter:  george.gayno@…  |      Owner:  george.gayno@…
      Type:  task            |     Status:  new
  Priority:  major           |  Milestone:
 Component:  component1      |    Version:
Resolution:                  |   Keywords:
-----------------------------+----------------------------

Comment (by george.gayno@…):

 '''r96885'''

 Update `CHGRES` to read lat/lon from the grid file (not the orography
 file) as is done by `CYCLE`. The grid file is double precision and the
 orography file is single precision. This can lead to potential differences
 in computed surface fields between the two programs (via sfcsub.f).  A
 quick C384 test using r96885 and r96719 showed floating point differences
 between some fields.  And some categorical fields were different as well.
 One point on tile 1 had a different slope type, and one point on tile 3
 had a different vegetation type.

 The regional model (NAM) does not have this issue because it computes all
 static fields once and places the output in the fixed directory.  Need to
 consider this paradigm for the global physics.

--
Ticket URL: <https://svnemc.ncep.noaa.gov/trac/fv3gfs/ticket/20#comment:10>
fv3gfs <https://svnemc.ncep.noaa.gov/trac/fv3gfs>
NGGPS FV3GFS Development



More information about the Ncep.list.emc.fv3gfs_tickets mailing list