Eta Model Parallel Change Log
New Mesoscale Branch Parallel Web Page
DEC 01
- 20/12Z: Changes to Eta parallels:
- ETAX turned off after this cycle
- 32-km EDAS control switched to run with
operational 3DVAR code and new cloud
microphysics and precipitation assimilation.
- 15/00Z: 12km/50level run in ETAX turned off. Started
12km/60level run in ETAX with fix to test for
vertical CFL violation in the vertical advection
routine.
- 08/12Z: Eta parallel status:
- ETAX : Start of 12 km / 50 level run;
EDAS cycle started from GDAS with
soil states from 32 km ETAV parallel
- ETAY and ETAL turned off; ETAV still running
with old (Zhao's) cloud scheme and
original precipitation assimilation
algorithm
NOV 01
- 29/18Z: Change to 32 km / 60 lev parallel:
- ETAL : New cloud microphysics with new
precip assimilation, with fix to reduce high
soil moisture bias.
- 27/12Z: Eta-12 parallel turned off; implemented
into operations with new cloud microphysics,
modified precip assimilation, and modified
3DVAR analysis.
- 27/12Z: Start of 32 km / 60 lev parallels:
- ETAV : Control run with old cloud microphysics
and old precip assimilation
- ETAY : New cloud microphysics and new
precip assimilation
- ETAL : New cloud microphysics with no
precip assimilation
OCT 01
- 20/00Z: Changed number of 3DVAR interations in
the Eta-12 parallel from 100 to 50.
- 13/00Z: Added new cloud microphsyics to the
Eta-12 parallel.
- 11/12Z: NEXRAD radial wind test in Eta-10
parallel (Central US domain) turned off
after this cycle.
SEP 01
- 18/12Z: Began test of 10-km Central US nest:
- Perform 3dvar reanalysis on 10-km grid with
NEXRAD radial winds and NOAA-16 radiances
- 10/06Z: ETAX parallel turned off after this cycle.
- 06/18Z: Turned off NOAA-16 radiances in ETAX parallel;
still running new 3DVAR code for a sanity check.
AUG 01
- 23/12Z: Start of twice/daily Eta-12 parallel runs
(parallel designated ETAX12). Other differences
from production:
- Use of NOAA-16 radiances
- Vertical diffusion of cloud water
- Twice the number of iterations in SLP routine
- 06/12Z: Began test of NOAA-16 radainces in the ETAX
parallel.
JUL 01
- 02/18Z: Due to a bug in GRDETA which caused the
fixed fields of vegetation type, soil type,
and surface slope type to be set to zero, the cycled
EDAS in the ETAX parallel was restarted
from the operational EDAS with a cold start
for the new non-frozen soil moisture variable.
JUN 01
- 11/00Z: All 48 km parallels (ETAV, ETAL, ETAY)
turned off after this cycle
MAY 01
- 23/12Z: ETAX parallel now running with assimilation
of observed precipitation, new land-surface
physics, and new 3DVAR analysis.
- 22/12Z: New 3dvar analysis added to assimilation
of observed precipitation in ETAX parallel.
- 21/18Z: Pulled new land-surface physics from
ETAX parallel due to persistent failures.
- 20/12Z: Pulled the new 3DVAR analysis from
the ETAX parallel due to persistent failures.
Restarted cycled EDAS in ETAX parallel from
operational EDAS.
- 17/18Z: Added new 3DVAR analysis to the 22-km
ETAX parallel along with assimilation of
hourly precipitation analyses and the
new land-surface physics.
- 16/18Z: Added new land-surface physics to the 22-km
ETAX parallel along with assimilation of
hourly precipitation analyses.
- 03/12Z: Test of new 3DVAR analysis code restarted
in ETAL parallel with retuned parameters and
ans new background errors. Cycled EDAS restarted
from ETAV control run.
- 02/12Z: All eta-10 runs now initialized using
soil moisture/temperature from the cycled
EDAS. Due to a code error they were using
GDAS soil moisture.
APRIL 01
- 25/00Z: Test of new 3DVAR analysis code restarted
in ETAL parallel with retuned parameters.
Cycled EDAS restarted from ETAV control run.
- 24/12Z: Began test of new land-surface physics in
the ETAY parallel. Cycled EDAS restarted from
ETAV control run.
- 04/12Z: Test of new 3DVAR analysis code started in
ETAL parallel. Cycled EDAS restarted from ETAV
control.
MARCH 01
- 29/18Z: Cycled EDAS in ETAX parallel restarted from
operational EDAS.
- 28/12Z: First run with assimilation of observed precipitation
in the 22 km ETAX parallel. Cycled EDAS restarted from
operational EDAS.
- 26/12Z: All 10 km nested runs now initialized from
the production parallel Eta-22.
- 20/18Z: First EDAS run with new test version of the
Eta 3DVAR code in the ETAX parallel with improved
balance constraint. Cycled EDAS restarted from
operational EDAS.
- 05/00Z: All Eta parallels are now running on system B.
- 04/12Z: Cycled EDAS in 22-km ETAX parallel restarted
from operational EDAS due to shutdown of development
queue on ASP.
FEBRUARY 01
- 24/00Z: Changed parameter (COAC) for horizontal diffusion
in all 48 km parallels from 0.125 to 0.06667
- 17/00Z: EDAS running in ETAL parallel restarted from ETAV
control.
- 15/00Z: ETAL parallel changed so that there is no assimilation
of observed precipitation anywhere that the model is
predicting snow to fall. EDAS running in ETAL parallel
restarted from ETAV control.
- 13/18Z: Cycled EDAS from ETAX put into operational Eta-22
- 13/18Z: All Eta parallels now using production observation
error table. They were using a table which had
different temperature/wind error covariances for
ob types 187,287 (METAR w/no valid station pressure)
JANUARY 01
- 31/12Z: First run of the 10-km Hawaii nest and the 10-km
Central U.S. nest. The Central U.S. nest will run
at 12Z, the Hawaii nest at 00Z and 12Z.
- 31/00Z: New 2DVAR SST analysis (implemented in
the operational Eta at 12Z 1/30) turned on in all Eta
parallels.
- 31/00Z: First run of the Alaska 12-km nested Eta forecast. It
will run at 00Z only.
- 24/12Z: Switched to running the Eastern and Western U.S.
nested Eta-10 runs at alternate times. The East10
forecast will run at 00Z, and the West10 forecast
will run at 12Z.
- 08/12Z: First run of the western U.S. (West10) 10 km nested
Eta forecast. Initial and boundary conditions
provided by the ETAX 22 km parallel.
- 06/00Z: High-resolution SST analysis used in ETAX, East10
parallel changed to use inhomogeneous correlation
length scales.
- 03/00Z: Switched East10 run to use the same SST which
is used in the ETAX parallel. Previously, it was
using the low-resolution 1 degree lat/lon SST
which is in the operational Eta-22.
DECEMBER 00
- 28/12Z: First run of the eastern U.S. (East10) 10 km nested
Eta forecast. Initial and boundary conditions
provided by the ETAX 22 km parallel.
- 21/18Z: Replaced global 1 degree lat/lon SST analysis with
OMB/CMB 2-d variational 0.5 deg lat/lon SST analysis
in the ETAX parallel.
- 21/00Z: Disabled use of high-resolution SST analysis
in all 48 km parallels; defaults to 1 degree lat/lon
global (Reynolds) SST analysis.
- 19/00Z: Disabled use of high-resolution SST analysis
in ETAX parallel; defaults to 1 degree lat/lon
global (Reynolds) SST analysis. This change was
implemented into production at 18Z 12/20/00.
- 08/12Z: The 18Z 11/21 change to the ETAX parallel
was not correctly implemented; this has been
done as of this cycle. EDAS cycle in ETAX restarted
from operational EDAS.
- 07/12Z: ETAY parallel resumes with test of Kain-Fritsch
cumulus parameterization scheme. Cycled EDAS started
from ETAV control.
NOVEMBER 00
- 21/18Z: Changed 22-km ETAX parallel to a test of the
operational 10-m wind algorithm in SFCDIF, but
with further modifications designed to decrease
roughness length (and increase 10-m wind) over
over higher terrain. Restarted cycled EDAS from
operational Eta-22.
- 14/12Z: Restarted 22-km ETAX parallel with test of new
4-point wind averaging in subroutine TURBL.
- 07/12Z: ETAL parallel with assimilation of observed
precipitation turned back on after the etafcst code
was fixed to prevent underground temperatures from
being used.
- 06/12Z: Put new 3DVAR code into 48 km parallels. The new
code fixes a bug so that the non-linear quality
control algorthim will be correctly applied to data.
- 01/12Z: Turned off ETAL parallel w/precipitation
assimilation pending fix of bug in the etafcst code.
OCTOBER 00
- 30/12Z: Began test of assimilation of precipitation (using
multi-sensor analysis) in the ETAL parallel. Cycled EDAS
started from ETAV control.
- 30/00Z: ETAY parallel (test of Kain-Fritsch convection) turned
off after this cycle.
- 24/12Z: This cycle is the last 22-km ETAX parallel run on the
IBM Phase I machine. When the Phase II machine becomes
available in early November ETAX will be turned back on.
- 16/12Z: Changes to the 22-km ETAX parallel
- Restored operational 10-m wind algorithm in SFCDIF, but
with modifications designed to decrease roughness length
(and increase 10-m wind) over higher terrain
- Began running ETAX out to 84-h.
- 05/00Z: Replaced operational NESDIS 0.5 deg lat/lon SST analysis with
CPC 2-d variational 0.5 deg lat/lon SST analysis in the ETAX
parallel.
- 04/12Z: Assimilation of satellite radiances, implemented in the
operational Eta at 12Z 9/26/00, was put into all 48 km
parallels now running (ETAV control and ETAY, K-F convection).
The EDAS for both parallels was restarted from the GDAS in
partial cycling mode.
- 04/00Z: ETAL parallel (assimilation of satellite radiances) was turned
off after this cycle.
- 03/12Z: Eta model changed in ETAX parallel to use original 10-m wind
algorithm in SFCDIF. Cycled EDAS restarted from operational Eta-22
SEPTEMBER 00
- 13/12Z: The Eta model changes in the ETAX parallel (see 00Z 8/3/00 entry)
were accidently removed from the ETAX etafcst code when it was recompiled
at 00Z 9/9/00; they have been reinstated starting with this cycle.
- 12/12Z: All 48 km parallels (ETAV, ETAY, ETAL) have been converted to run
out to 60-h.
- 09/00Z: Restored all 3DVAR parameters back to the values used before the
12Z 9/4/00 changes.
- 04/12Z: Changed 3DVAR parameters in ETAX using numbers derived from Dave Parrish's
tuning experiments.
AUGUST 00
- 24/12Z: Restarted Eta-22 with satellite radiances in the ETAX parallel. Cycle
restarted from GDAS, but using operational Eta-32 soil moisture.
- 15/12Z: Began runs of the Eta-48 with the Kain-Fritsch convective scheme in the
ETAY parallel; cycled EDAS in ETAY restarted from ETAV control run.
- 09/00Z: ETAX (22 km Eta w/radiances) temporarily turned off for tuning of the
3DVAR analysis.
- 08/00Z: Changes to the the 3DVAR analysis in the 22 km ETAX parallel :
- Increase horizontal correlation length for moisture (bshrinkq changed
from 1.5 to 2.2)
- Increase vertical correlation length for moisture (pshrinkq changed
from 0.5 to 1.0)
- Decreased temperature background error (bscalet changed from 3.0 to 2.0)
- Decreased wind background error (bscalepsi changed from 5 to 4)
- 03/00Z: Eta model changes in the 22 km ETAX parallel:
- Added vertical advection of cloud
- Use cloud efficiency (CLDEFI) from the EDAS instead of
reinitializing it at the start of each forecast
- Set convection threshold to zero (was 0.25 in/24 h)
- 03/00Z: The parameter file for the 22 km 3DVAR analysis running in the
ETAX parallel was changed to do the following:
- Decrease height background error (bscaleh changed from 1.5 to 0.5)
- Increase temperature background error (bscalet changed from 1.4 to 3.0)
- Increase wind bacground error (bscalepsi changed from 3 to 5)
- Strengthen mass-wind balance (errgeo0, errhgeo0 changed from 1 to 0.7)
- Reduce raob height and moisture observation error covariances
JULY 00
- 31/12Z: ETAL parallel (48 km w/radiances) now using revised parallel 3DVAR
code which is running in the 22km ETAX parallel; all 48 km
parallels now using "internal quilt" version of the forecast
code.
- 24/12Z: Started using revised parallel 3DVAR code in 22km ETAX parallel,
with use of satellite radiances.
- 07/12Z: Switched back to using the operational version of the 3DVAR
code (without satellite radiances) in the ETAX parallel while
the new parallel version is being debugged.
- 06/12Z: Started running parallel MPI version of 3DVAR code with
Bill Collins' non-linear quality control in the 22 km ETAX
parallel.
JUNE 00
- 29/00Z: Changes to ETAY parallel:
- Revert back to unified sea profiles in BMJ convection
- Change FSS from 0.95 to 1.25. This "inverts" the BMJ
moisture profiles so that the moist limit is 25% drier
than the dry limit, which should increase precipitation
maxima.
- Cycled EDAS in ETAY restarted from ETAV control.
- 27/12Z: Switched to using new parallel 3DVAR code with recursive
filters in the ETAX parallel, with satellite radiances.
- 13/00Z: Turned off processing of NOAA-15 radiances in ETAL parallel.
- 09/12Z: Started running new version of parallel 3DVAR code with
recursive filters in the ETAL parallel.
- 08/00Z: Switched to using operational EDAS/Eta dumps of radiance
data in ETAL parallel
- 06/12Z: 22km EDAS/Eta parallel started in the ETAX parallel slot.
Soil parameters from the operational 32km EDAS were
used to start the cycled EDAS.
- 01/00Z: BMJ convection in ETAY parallel changed to use
land profiles everywhere (in addition to other changes
discussed in 12Z 5/8/00 entry). Cycled EDAS for ETAY
restarted from ETAV control.
MAY 00
- 08/12Z: Began test of the following in the ETAY parallel:
- Vertical advection of cloud
- Use cloud efficiency (CLDEFI) from the EDAS instead of
reinitializing it at the start of each forecast
- Set convection threshold to zero (was 0.25 in/24 h)
- Change FSS from 0.85 to 0.95 (reduce range between "fast"
and "slow" profiles in convection scheme)
- 04/12Z: Due to spurious precipitation bug in the 48 km Eta model (see
5/4/00 entry) the cycled EDAS for both the ETAV and ETAL
parallels was restarted from the GDAS.
- 04/00Z: ETAY test of SSM/I obs turned off after this cycle.
- 04/00Z: Recompiled Eta model forecast executable for ETAV, ETAL, ETAY
parallels using Tom's latest version of the code. The previous
version used had a discrepency between the cuparm file and
the convection subroutine (CUCNVC.F) which led to unrealistic
precipitation maxima.
- 02/00Z: Radiance version of 3dvar code in ETAL modifed to use the sensible skin
temperature from the EDAS first guess instead of the potential
skin temperature.
APRIL 00
- 27/00Z: Dave Parrish fixed a bug in radiance version of 3dvar code running
in ETAL parallel which would occasionally caused the code to
fail to converge.
- 26/12Z: Cycled EDAS in ETAL parallel restarted from ETAV control after an
error in the radiance version of the eta 3DVAR code was fixed so
that the correct observation time for the TOVS radiances is used.
The previous version was using the wrong number which led to almost
all radiance data being excluded from the analysis.
- 26/00Z: Cycled EDAS in ETAL parallel restarted from ETAV control after
a script bug was fixed which prevented the correct bias corrections
from being read into the radiance version of the eta 3DVAR code.
- 18/00Z: Turned SSMI PW data back on in ETAL parallel.
- 17/12Z: Began use of "combined" SSMI data in ETAY; neural net oceanic
surface winds and NESDIS PW data, with observation error
covariance for SSMI PW data set back to 8 g/kg. Restarted
cycled EDAS from ETAV control run.
- 14/12Z: Increased the observation error covariance for SSMI PW data
in the ETAY parallel from 8 g/kg to 16 g/kg. Eliminated
creation of 1 deg lat/lon superobs for the SSMI data, reverting
to original 25 km resolution.
- 11/12Z: Started Eta-48 with GOES/TOVS radiances in the ETAL
slot.
MARCH 00
- 28/12Z: ETAX turned off after this cycle; changes running in ETAX
implemented in operational Eta at 12Z 29 March 2000.
- 20/12Z: Started 48 km parallel test of neural-net SSM/I surface winds
and precipitable water data in ETAY slot.
- 10/12Z: ETAX : Put Geoff Manikin's new versions of SNDP and ETAPOST with
the Bunkers algorthim for storm relative motion.
- 09/12Z: ETAX: Additional changes to CHKOUT.F, CUCNVC.F, PHYS.comm, PHYS2.comm
to remove spurious cloud diagnostics. Also added new DDAMP.F routine
to fix bug which was causing 48-km run to blow up.
- Began 48 km control parallel (ETAV) on IBM.
- 07/12Z: ETAX: New versions of CUCNVC.F and HZADV2.F in eta_etafcst to remove
spurious cloud diagnostics as seen by Manikin/Jascourt.
- 03/00Z: Changed RSTUPT in ETAX parallel so that the SNO array is updated
with the new NESDIS snow analysis once per day (in the 12Z EDAS
at t-12 h) instead of twice/day.
FEBRUARY 00
- 15/00Z: Added new version of 3dvar which accounts for drift of
rawinsonde balloon to the ETAX parallel
- 12/00Z: ETAX (32 km) parallel started on IBM-SP:
- CQC of VAD winds (turned back on in Eta 3dvar analysis)
- Forecast run to 60-h
- Modified BMJ convection (sea profiles only, modified
convection of elevated terrain)
SEPTEMBER 99
- 28/00Z: ETAV/ETAY parallels ended due to cray3 fire
JUNE 99
- 16/12Z: Cycled EDAS in ETAY parallel restarted from ETAV guess
due to break in ETAY cycle on cray5. Put newest 3DVAR analysis
(with fix of underground RH and wind superobs) into ETAV and ETAY parallels.
MAY 99
- 24/12Z: Due to /dm problems on cray3, 80 km control parallel
did not run from 00Z 5/22- 12Z 5/23, breaking the EDAS cycle.
Problems with para on cray3 caused a break in the ETAY parallel.
Therefore, EDAS cycles in both 80km parallels were restarted from
the GDAS.
- 03/12Z: Both 80km parallels (ETAV and ETAY) switched to use
the new 3DVAR analysis without interruption of the EDAS cycle.
Discovered that a script error caused the 00Z EDAS in the ETAY
parallel to be started from a 12h Eta foreast, not the previous EDAS.
APRIL 99
- 01/12Z: Test of new convection (with unified sea profiles in
Betts-Miller-Janjic scheme) started with this cycle in the ETAY
parallel. Modified thermal conductivity to use fixed soil moisture
of 0.26 (subroutine SFLX/TDFCND).
MARCH 99
- 31/12Z: Test of new 3DVAR code in ETAY ended with this cycle.
- 13/12Z: Started test of new 3DVAR code in ETAY parallel on cray5.
- 11/12Z: Restarted ETAV (80km control run) on cray3.
FEBRUARY 99
-
26/00Z:
Changes to the Kain-Fritsch 32km parallel run include:
- Changes designed to suppress deep convective initiation in marginally
unstable environments with dry lower to middle troposphere and to
suppress widespread non-convective "drizzle":
- minimum entrainment (dilution) rate imposed in updraft calculations
- initial updraft vertical velocity limited to no greater than 3 m/s
- shallow convection enabled with TKE closure
- Minimum cloud depth to initiate deep convection allowed to vary as a
function of cloud-base temperature. Minimum cloud depth decreases as
cloud-base temperature decreases, down to minimum value of 2 km for
temperatures below 0 deg C
- Changes to downdraft formulation:
- Downdraft influx occurs in 150-200mb layer above cloud base. Each
model layer in this layer contributes equally.
- Downdraft mass flux dependent on mean RH in source layers.
- Downdraft assumed to be saturated at base of its source layer,
decreases RH by 20%/km below this layer.
JANUARY 99
-
27/00Z: 80km 4DVAR parallel running in the ETAZ slot was turned off after
this cycle.
-
12/12Z: First successful run of 80-km cycled 4DVAR parallel in ETAZ slot;
both ETAZ and control parallel (ETAV) restarted from the GDAS.
Due to script errors ETAZ output files were not saved until 12Z
1/13 cycle.
-
10/00Z: Convection test running in ETAY parallel on cray5 was turned off
after this cycle due to NCO use of cray5 for end-to-end Y2K
testing.
DECEMBER 98
-
24/00Z: CUCNVC in ETAY changed set minimum depth for convection to be
(290 mb * (PSFC/1000)) (the same as ETAV) instead of
200 mb as was done previously.
NOVEMBER 98
-
16/12Z: Unknown problems and a UNIX script error in the ETAV parallel
caused 80 km parallel EDAS' to be restarted from the GDAS