Skip to content

Instantly share code, notes, and snippets.

@sanAkel

sanAkel/rtofs-atm-forcings.md

Last active January 8, 2026 23:21
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save sanAkel/449fcb46f01d4e6184be64adb21dfec9 to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save sanAkel/449fcb46f01d4e6184be64adb21dfec9 to your computer and use it in GitHub Desktop.
Download ZIP
Raw
rtofs-atm-forcings.md

Sources:

  • MOM6 from Zulema's experiment: /lfs/h2/emc/couple/noscrub/zulema.garraffo/FV3_RT/DATM_INPUT/GFS.20230301-20231231.nc_hCk_positive.nc
  • RTOFS my experiment: /lfs/h2/emc/couple/noscrub/santha.akella/prod/com/rtofs/v2.5/rtofs.20241208/
Field name Available to use in MOM6 Used in RTOFS Long name of field (units)
slmsksfc Y ? sea-land-ice mask (0-sea, 1-land, 2-ice) (1)
icecsfc Y ? sea-ice coverage/fraction, proportion (1)
--- --- --- ---
psurf Y ? surface pressure (Pa)
pres_hyblev1 Y ? layer 1 pressure (Pa)
hgt_hyblev1 Y ? layer 1 height (m)
--- --- --- ---
fprecp Y ? surface snow precipitation rate (kg/m**2/s)
totprcp_ave Y ? surface precipitation rate (kg/m**2/s)
precp Y ? surface rain precipitation rate (kg/m**2/s)
--- --- --- ---
DLWRF Y ? surface downward longwave flux (W/m**2)
DSWRF Y ? averaged surface downward shortwave flux (W/m**2)
vbdsf_ave Y ? Visible Beam Downward Solar Flux (W/m**2)
vddsf_ave Y ? Visible Diffuse Downward Solar Flux (W/m**2)
nbdsf_ave Y ? Near IR Beam Downward Solar Flux (W/m**2)
nddsf_ave Y ? Near IR Diffuse Downward Solar Flux (W/m**2)
ULWRF Y ? surface upward longwave flux (W/m**2)
shtfl_ave Y ? surface sensible heat flux (W/m**2)
lhtfl_ave Y ? surface latent heat flux (W/m**2)
--- --- --- ---
dusfc Y ? surface zonal momentum flux (N/m**2)
dvsfc Y ? surface meridional momentum flux (N/m**2)
u10m Y ? 10 meter u wind (m/s)
v10m Y ? 10 meter v wind (m/s)
ugrd_hyblev1 Y ? layer 1 zonal wind (m/s)
vgrd_hyblev1 Y ? layer 1 meridional wind (m/s)
--- --- --- ---
t2m Y ? 2m temperature (K)
q2m Y ? 2m specific humidity (kg/kg)
spfh_hyblev1 Y ? layer 1 specific humidity (kg/kg)
tmp_hyblev1 Y ? layer 1 temperature (K)

Analysis quality forcings used in RTOFS:

See files: /lfs/h2/emc/couple/noscrub/santha.akella/prod/com/rtofs/v2.5/rtofs.20241208/rtofs_glo.anal.*.b

airtmp
glbrad
lwdflx
precip
presur
radflx
shwflx
surtmp
vapmix
wndewd
wndnwd
wndspd

analysis cycle forcing (thanks to Dan):

/lfs/h1/ops/prod/com/gfs/v16.3/gdas.20241101/00/atmos/gdas.t00z.sfluxgrbf00*grib2

RTOFS DA uses hours = 1, 2, 3, 4, 5, 6 from any DA cycle. Hour "0" from previous cycle.

See listflx.dat in RTOFS dir

@sanAkel
Copy link
Author

sanAkel commented Jan 5, 2026 •
edited
Loading

Mapping between grib2 ➡️ netcdf

GDAS to DATM Variable Mapping & Transformation Reference

List of variables needed by DATM has been gathered from datm.streams, see stream_data_variables01. This is a listing of many fields, for each field, the format is: var_name_in_forcing_file_DATM_reads_in and CDEPS_internal_variable_name; you need to write out the NetCDF file with the former.

This table defines the surgical extraction of GRIB2 messages and the subsequent NCO (NetCDF Operators) processing required to produce UFS-compatible forcing for RTOFS/ocean coupling.

Physical Constants & Assumptions

  • Freezing Point of Seawater ($T_f$): $271.35,K$ ($-1.8^\circ C$)
  • Latent Heat of Vaporization ($L$): $2.5 \times 10^6,J/kg$
  • Sign Convention: All fluxes are transformed to positive downward (into the ocean).

Variable Mapping Table

GRIB2 Search Key (Input) Intermediate NC Name Final DATM Name Physical Fixes / Sign Flips / Derivations
TMP:2 m above ground tair Sa_t2m T-Star Adjustment: If $T_{sfc} \le T_{f}$, then $Sa_t2m = T_{f} + tair - T_{sfc}$; else $Sa_t2m = tair$.
TMP:surface tsfc Null Trigger for $Sa_t2m$ adjustment logic.
SHTFL:surface shtfl Sa_shflx Sign Flip: $Sa_shflx = float(shtfl \times -1.0)$.
LHTFL:surface lhtfl Sa_lhflx Sign Flip: $Sa_lhflx = float(lhtfl \times -1.0)$.
UFLX:surface uflx Sa_taux Sign Flip: $Sa_taux = float(uflx \times -1.0)$.
VFLX:surface vflx Sa_tauy Sign Flip: $Sa_tauy = float(vflx \times -1.0)$.
LHTFL:surface lhtfl evap Derivation: $evap = float(Sa_lhflx / L)$. Check: where(evap < 0.) evap=0.
PRATE:surface prate_all Faxa_rain Split: $prate_all \times (1.0 - cpofp/100.0)$. Check: where < 0 = 0.
CPOFP:surface cpofp Faxa_snow Split: $prate_all \times (cpofp/100.0)$. Check: where < 0 = 0.
DSWRF:surface DSWRF Faxa_swdn Check: where(Faxa_swdn < 0.) Faxa_swdn=0.
DLWRF:surface DLWRF Faxa_lwdn Check: where(DLWRF < 0.) DLWRF=0.
VBDSF:surface vbdsf_ave Faxa_swvdr Check: where < 0 = 0. (Visible Beam Downward)
VDDSF:surface vddsf_ave Faxa_swvdf Check: where < 0 = 0. (Visible Diffuse Downward)
NBDSF:surface nbdsf_ave Faxa_swndr Check: where < 0 = 0. (Near-IR Beam Downward)
NDDSF:surface nddsf_ave Faxa_swndf Check: where < 0 = 0. (Near-IR Diffuse Downward)
HGT:1 hybrid level Sa_z Sa_z Mid-layer height of lowest model level.
TMP:1 hybrid level Sa_tbot Sa_tbot Mid-layer temperature of lowest model level.
SPFH:1 hybrid level Sa_shum Sa_shum Check: where(Sa_shum < 0.) Sa_shum=0.
UGRD:1 hybrid level Sa_u Sa_u Zonal wind at lowest model level.
VGRD:1 hybrid level Sa_v Sa_v Meridional wind at lowest model level.
PRES:surface Sa_pslv Sa_pslv Surface Pressure.
LAND:surface Sa_mask Sa_mask Land/Sea Mask.
ICEC:surface icec Null Check: where(icec < 0.) icec=0. (Sea Ice Conc.)

Processing Workflow

  1. Extraction: wgrib2 filters raw GRIB2 files using exact colon-delimited match strings.
  2. Renaming: ncrename aligns wgrib2 default NetCDF names with UFS/CDEPS conventions.
  3. Arithmetic Pass: A single-pass ncap2 command handles sign-flipping, ice-air temperature adjustments, and precipitation splitting.
  4. Aggregation: ncrcat merges individual files into the final forcing time-series.

@sanAkel
Copy link
Author

sanAkel commented Jan 8, 2026

Next iteration follows.

Table of mapping variables from GFS (sflux) GRIB2 to UFS DATM NetCDF files

How was it generated?

1. Start with what is at the ufs-weather-model

2. Then modify it based on what is actually used by datm_datamode_gefs_mod.F90

Notes:

- GFS pressure levels are defined here; see HGT:1.

GRIB2 Search Key (Input) NetCDF forcing file variable name Name in datm_datamode_gefs_mod.F90 Checks and/or changes Avg forecast
:LAND:surface: slmsksfc Sa_mask None False
:HGT:1 hybrid level: hgt_hyblev1 Sa_z None False
:UGRD:1 hybrid level: ugrd_hyblev1 Sa_u None False
:VGRD:1 hybrid level: vgrd_hyblev1 Sa_v None False
:TMP:1 hybrid level: tmp_hyblev1 Sa_tbot None False
:PRES:1 hybrid level: pres_hyblev1 Sa_pbot None False
SPFH:1 hybrid level: spfh_hyblev1 Sa_shum where < 0 = 0. False
:UGRD:10 m above ground: u10m Sa_u10m None False
:VGRD:10 m above ground: v10m Sa_v10m None False
:TMP:surface: N/A N/A None False
:TMP:2 m above ground: t2m Sa_t2m where (tmp_sfc <= 271.35) t2m = 271.35 + t2m - tsfc False
:SPFH:2 m above ground: q2m Sa_q2m where <0 = 0. False
PRES:surface psurf Sa_pslv None False
DLWRF:surface DLWRF Faxa_lwdn None False
PRATE:surface N/A N/A Total precip. where < 0 = 0. True
CPOFP:surface N/A N/A Used to calculate partition of liquid/frozen precip
N/A precp Faxa_rain PRATE * (1-CPOFP*0.01) True
N/A frecp Faxa_snow PRATE * CPOFP*0.01 True
VBDSF:surface vbdsf_ave Faxa_swvdr where < 0 = 0. True
VDDSF:surface vddsf_ave Faxa_swvdf where < 0 = 0. True
NBDSF:surface nbdsf_ave Faxa_swndr where < 0 = 0. True
NDDSF:surface nddsf_ave Faxa_swndf where < 0 = 0. True

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment