IRS, HIRS and CSBT Experiments Plus Other FY05/06 Results by James A. Jung Tom H. Zapotocny and John Le Marshall Cooperative Institute for Meteorological Satellite Studies University of Wisconsin-Madison
AIRS, HIRS and CSBT Experiments Plus Other FY05/06 Results
by
James A. Jung Tom H. Zapotocnyand
John Le Marshall
Cooperative Institute for Meteorological Satellite Studies
University of Wisconsin-Madison
References
Peer-Reviewed Publications:
Le Marshall, J., J. Jung, J. Derber, M. Chahine, R. Treadon, S. J. Lord, M. Goldberg, W. Wolf, H. C. Liu, J. Joiner, J. Woolen, R. Todling, P. Van Delst and Y. Tahara, 2006: Atmospheric Infrared Sounder (AIRS) observations improve global analysis and forecast capability. Submitted to Bull. Amer. Meteor. Soc.
Le Marshall, J., J. Jung, T. Zapotocny, J. Derber, R. Treadon, M. Goldberg and W. Wolf, 2006: The Application of AIRS Radiances in Numerical Weather Prediction. Submitted to Aust. Meteor. Mag. February 2006.
Zapotocny, T. H., J. A. Jung, J. F. Le Marshall, R. E. Treadon, 2006: A Two Season Impact Study of Satellite and In-Situ Data in the NCEP Global Data Assimilation System. Submitted to Weather and Forecasting May 2006.
Conference Papers/Posters Presented at 14th Conference on Satellite Meteorology and Oceanography or Eighth International Winds Workshop 2006 :
Jung, J. A., T. H. Zapotocny, R. E. Treadon, J. F. LeMarshall, 2006: Atmospheric Infrared Sounder Assimilation Experiments using the National Center for Environmental Prediction’s Global Forecast System.
Le Marshall, J. F., J. A. Jung, J. Derber, R. Treadon, M. Goldberg, W. Wolf, and T. H. Zapotocny, 2006: Assimilation of Atmospheric Infrared Sounder (AIRS) Observations at JCSDA.
Le Marshall, J. F., T. Zapotocny, C. Redder, J. Jung, M. Dunn, R. Seecamp, L. P. Riishojgaard, A. Rea, J. Daniels, D. Santek, C. Velden, W. Bresky, 2006: Error Characterization and use of Atmospheric Motion Vectors at the JCSDA.
Santek, D., J. A. Jung, T. H. Zapotocny, J. Key, C. Velden, 2006: Mechanisms that Propagate Polar Satellite Derived Atmospheric Motion Vector Information into Lower Latitudes.
Zapotocny, T. H., J. A. Jung, R. E. Treadon, J. F. LeMarshall, 2006: Recent MODIS Assimilation Experiments in the NCEP GDAS.
HIRS ExperimentHIRS Experiment
Data Denial OSEData Denial OSE
GOES Imager CSBT ExperimentGOES Imager CSBT Experiment
OSE using AIRSOSE using AIRS
Experiments Presented in This Talk
HIRS HIRS Assimilation Assimilation ExperimentsExperiments
Experiment DesignExperiment Design
T254L64T254L64 Operational version of SSI and GFSOperational version of SSI and GFS AIRS and Aqua AMSU not usedAIRS and Aqua AMSU not used AMSU ice emissivity constantAMSU ice emissivity constant 1 Jan – 15 Feb 20041 Jan – 15 Feb 2004
Assimilation TechniqueAssimilation Technique
Adjusted Assimilation WeightsAdjusted Assimilation Weights Night Land/Ocean Clear TestNight Land/Ocean Clear Test
Abs(11.11Abs(11.11µmµm – 4.00 – 4.00µmµm + 1.0) < 2.0 + 1.0) < 2.0 Abs(11.11Abs(11.11µmµm – 3.76 – 3.76µmµm) < 2.0) < 2.0 Abs(4.00Abs(4.00µmµm – 3.76 – 3.76µmµm + .5) > 0.2 + .5) > 0.2
Polar Inversion TestPolar Inversion Test 7.337.33µmµm – 11.11 – 11.11µmµm > 0.0 > 0.0
Use warmest field of view in boxUse warmest field of view in box If other tests fail or not applicableIf other tests fail or not applicable
ResultsResultsN. Hemisphere 1000 hPa AC Z 20N - 80N
Waves 1-20 1 Jan - 15 Feb '04
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
0 1 2 3 4 5 6 7
Forecast [day]
An
om
aly
Co
rrel
atio
n '
Control HIRS
S. Hemisphere 1000 hPa AC Z 20S - 80S Waves 1-20 1 Jan - 15 Feb '04
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
0 1 2 3 4 5 6 7
Forecast [day]
An
om
aly
Co
rrel
atio
n '
Control HIRS
N. Hemisphere 500 hPa AC Z 20N - 80NWaves 1-20 1 Jan - 15 Feb '04
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
0 1 2 3 4 5 6 7
Forecast [days]
An
om
aly
Co
rrel
atio
n '
Control HIRS
S. Hemisphere 500 hPa AC Z 20S - 80SWaves 1-20 1 Jan - 15 Feb '04
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
0 1 2 3 4 5 6 7
Forecast [day]
An
om
aly
Co
rrel
atio
n '
Control HIRS
Day 5 Average Anomaly CorrelationWaves 1- 20
1 Jan - 15 Feb 2004
0.72
0.74
0.76
0.78
0.8
0.82
0.84
0.86
nh 500 sh 500 nh 1000 sh 1000
control HIRS
Future PlansFuture Plans
T382L64T382L64 Latest version of GSI (June 2006)Latest version of GSI (June 2006) Use hybrid version of GFSUse hybrid version of GFS 2 Seasons2 Seasons Determine separate selection criteria for Determine separate selection criteria for
land and ocean (night only)land and ocean (night only) Re-evaluate observation errorsRe-evaluate observation errors Continue literature search for potential Continue literature search for potential
selection criteria over snow and iceselection criteria over snow and ice
Data Denial Data Denial Observing Observing
System System ExperimentsExperimentsPolar Satellite DenialsPolar Satellite Denials
Experiment DesignExperiment Design Pre - Operational GDAS (Oct 2003)Pre - Operational GDAS (Oct 2003) Control (3_NOAA)Control (3_NOAA)
All operational dataAll operational data NOAA-15, 16, 17 usedNOAA-15, 16, 17 used
Deny NOAA – 15 (2_NOAA)Deny NOAA – 15 (2_NOAA) AMSUAMSU
Deny NOAA – 15, 16 (1_NOAA)Deny NOAA – 15, 16 (1_NOAA) 15 AMSU 15 AMSU 16 AMSU and HIRS16 AMSU and HIRS
N. Hemisphere 500 hPa AC Z 20N - 80NWaves 1-20 15 Jan - 15 Feb 2003
0.5
0.6
0.7
0.8
0.9
1
0 1 2 3 4 5 6 7
Forecast [day]
An
om
aly
Co
rrel
atio
n
3_NOAA 1_NOAA
S. Hemisphere 500 hPa AC Z 20S - 80SWaves 1-20 15 Jan - 15 Feb 2003
0.5
0.6
0.7
0.8
0.9
1
0 1 2 3 4 5 6 7
Forecast [day]
An
om
aly
Co
rrel
atio
n
3_NOAA 1_NOAA
Day 5 Average Anomaly CorrelationWaves 1-20 15 Jan - 15 Feb 2003
0.7
0.72
0.74
0.76
0.78
0.8
0.82
0.84
0.86
0.88
N.H. 1000 N.H. 500 S.H. 1000 S.H. 500
An
om
aly
Co
rre
lati
on
3_NOAA 2_NOAA 1_NOAA
Day 5 Average Anomaly CorrelationWaves 1-20 15 Aug - 20 Sep 2003
0.7
0.72
0.74
0.76
0.78
0.8
0.82
0.84
0.86
0.88
N.H. 1000 N.H. 500 S.H. 1000 S.H. 500
An
om
aly
Co
rre
lati
on
3_NOAA 2_NOAA 1_NOAA
Impact of 3_NOAA vs 1_NOAA
Table: Atlantic Basin average forecast error (km) during 2004. 12.8 29.3 45.9 61.1 70.9 105.7 153.6 196.1 1 NOAA
14.7 29.6 47.6 65.0 80.4 119.3 156.0 214.7 2 NOAA
12.2 31.8 45.6 59.4 70.6 103.0 137.3 201.4 3 NOAA
38 36 31 30 25 19 17 15 # cases
00-hr 12-hr 24-hr 36-hr 48-hr 72-hr 96-hr 120-hr Fcst Hr
Results compiled by Qingfu Liu
Analysis is Analysis is continuingcontinuing
GOES Imager Clear GOES Imager Clear Sky Brightness Sky Brightness TemperatureTemperatureAssimilation Assimilation ExperimentsExperiments
(CSBT)(CSBT)
Experiment DesignExperiment Design
Current Operational version of GFS and Current Operational version of GFS and SSISSI
T382L64T382L64 Observation errors from tableObservation errors from table 1 Dec – 24 Dec 20051 Dec – 24 Dec 2005 Tested 2 scenariosTested 2 scenarios
Water vapor channel only (channel 3)Water vapor channel only (channel 3) The 4 IR channels (channels 2, 3, 4, 5 or 6)The 4 IR channels (channels 2, 3, 4, 5 or 6)
Collected statistics from all channelsCollected statistics from all channels
Assimilation TechniqueAssimilation Technique
Thinned to 180 Km assimilation boxThinned to 180 Km assimilation box Introduced since previous testIntroduced since previous test
Minimum Standard deviation chosen in Minimum Standard deviation chosen in each boxeach box Standard deviation must be less than 0.6Standard deviation must be less than 0.6
Must have 20% or more clear pixelsMust have 20% or more clear pixels Channel 2 used only at nightChannel 2 used only at night Surface channels used only over oceanSurface channels used only over ocean
Channels 2, 4, 5 or 6Channels 2, 4, 5 or 6 Data used only at 6 hour synoptic timeData used only at 6 hour synoptic time
Assimilation ResultsAssimilation Results
More observations used/passed QC than More observations used/passed QC than in previous versionin previous version Better cloud clearingBetter cloud clearing Midnight effect fixedMidnight effect fixed
Neutral in Northern HemisphereNeutral in Northern Hemisphere Slight negative in Southern HemisphereSlight negative in Southern Hemisphere Slight negative in TropicsSlight negative in Tropics Histograms still have cloud Histograms still have cloud
contaminated tail.contaminated tail.
S. Hemisphere 1000 hPa AC Z 20S - 80S Waves 1-20 1 Dec - 24 Dec 2005
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
0 1 2 3 4 5 6 7
Forecast [day]
An
om
aly
Co
rrel
atio
n '
Control CSBT
CSBT assimilation O-B histograms GOES - 10 at 12Z
0
500
1000
1500
2000
2500
3000
3500
4000
Delta T [K]
Co
un
ts
ch 2 ch 3 ch 4 ch 5
Chan 3 OK
Chan 2, 4, 5/6 still have cold tails
Recommendations made Recommendations made to NESDISto NESDIS
Change the standard deviation cutoff Change the standard deviation cutoff factor in the cloud mask from 1.2 to 1.0 factor in the cloud mask from 1.2 to 1.0 for channels 4 and 5 (or 6)for channels 4 and 5 (or 6) Remove more cloud contamination from Remove more cloud contamination from
observationsobservations Increase the precision of the standard Increase the precision of the standard
deviation in the BUFR file from 0.1 to deviation in the BUFR file from 0.1 to 0.010.01 Allow for better selection of observations by Allow for better selection of observations by
the assimilation systemthe assimilation system
Future PlansFuture Plans
Re-evaluate data and assimilation Re-evaluate data and assimilation technique after modifications are technique after modifications are made to the real-time datamade to the real-time data
Switch to latest version of GSI (June Switch to latest version of GSI (June 2006)2006)
Use Hybrid version GFSUse Hybrid version GFS T382L64 if possibleT382L64 if possible Test all 4 channels for impactTest all 4 channels for impact
Observing Observing System System
Experiments Experiments Using AIRSUsing AIRS
Experiment DesignExperiment Design
T254L64T254L64 Operational SSI and GFSOperational SSI and GFS Operational selection criteriaOperational selection criteria
Warmest spotWarmest spot Tests checked every field of viewTests checked every field of view AMSU ice emissivity constantAMSU ice emissivity constant AQUA AMSU on for control and all AQUA AMSU on for control and all
experimentsexperiments
Assimilation TechniquesAssimilation Techniques
All AIRS channels off All AIRS channels off (control)(control)
115 Channels on from 3.76 – 9.13115 Channels on from 3.76 – 9.13µm µm (shortwave)(shortwave)
152 Channels on from 3.76 -15.40µm 152 Channels on from 3.76 -15.40µm (opschan)(opschan)
251 Channels on from 3.76 – 15.40µm 251 Channels on from 3.76 – 15.40µm (longwave)(longwave)
Day 5 Average Anomaly CorrelationWaves 1- 20
1 Jan - 15 Feb 2004
0.73
0.75
0.77
0.79
0.81
0.83
0.85
nh 500 sh 500 nh 1000 sh 1000
control
longwave
shortwave
opscahan
Day 5 Average Anomaly CorrelationWaves 1 - 20
10 Aug - 20 Sep 2004
0.75
0.76
0.77
0.78
0.79
0.8
0.81
0.82
nh 500 sh 500 nh 1000 sh 1000
control
longwave
shortwave
opschan
24-hr 850 hPa Relative Humidity for longwave experiment
24-hr Precipitable Water for longwave experiment
24-hr 500 hPa Temperature for longwave experiment
24-hr 500 hPa Geopotential Height for longwave experiment
24-hr 250 hPa U - Component for longwave experiment
Thank you:Thank you:
Russ Treadon for coding, scripts and Russ Treadon for coding, scripts and verification assistance.verification assistance.
NCEP (Stephen Lord) for computer NCEP (Stephen Lord) for computer
resources and tape space.resources and tape space.
Stacie Bender and Dennis Keyser for Stacie Bender and Dennis Keyser for collecting and processing our collecting and processing our various data streams.various data streams.
The AIRS Science team.The AIRS Science team.