1. Infrared Satellite Data Assimilation at NCAR
Tom Auligné, Hui-Chuan Lin, Zhiquan Liu, Hans Huang, Syed Rizvi, Hui Shao, Meral Demirtas, Xin Zhang
National Center for Atmospheric Research
Work supported by AFWA, NASA, NSF, KMA

2. Outline Introduction to satellite data assimilation at NCAR
Practical issues with AIRS data assimilation
Current developments on infrared radiances

3. Introduction: Data assimilation at NCAR
WRF-ARW: Local Area Model (with global version) + TL/ADJ version
DART: Ensemble Data Assimilation (EnKF, ETKF, …) (no radiance yet)
WRF-Var: Variational Data Assimilation (3DVar, FGAT, 4DVar) + Hybrid system
Community support

4. Satellite DA: WRF-Var capabilities
Retrievals (T / Q profiles)
SATEM (from AMSU)
AIRS retrievals (NASA version 5)
GPS Radio Occultation
Retrieved refractivity from COSMIC
Winds
Retrieved winds: polar MODIS, SATOB
Active sensors: Quikscat
Radiances (BUFR format from NCEP/NRL/AFWA/NESDIS)
HIRS from NOAA16, 17, 18
AMSU-A from NOAA15, 16, 18, EOS-Aqua, METOP-2
AMSU-B from NOAA15, 16, 17
MHS from NOAA18, METOP-2
AIRS from EOS-Aqua
SSMIS from DMSP16

5. Satellite DA: WRF-Var capabilities
Retrievals (T / Q profiles)
SATEM (from AMSU)
AIRS retrievals (NASA version 5): NASA-EOS Project to assess impact over Antarctica
GPS Radio Occultation
Retrieved refractivity from COSMIC
Winds
Retrieved winds: polar MODIS, SATOB
Active sensors: Quikscat
Radiances (BUFR format from NCEP/NRL/AFWA/NESDIS)
HIRS from NOAA16, 17, 18
AMSU-A from NOAA15, 16, 18, EOS-Aqua, METOP-2
AMSU-B from NOAA15, 16, 17
MHS from NOAA18, METOP-2
AIRS from EOS-Aqua: AFWA Project to incorporate in operational 3DVar
SSMIS from DMSP16

6. AIRS Channel Selection: 10hPa model top
Ozone
Solar contamination
RTTOV
(v 8.7)
AIRS
T Jacobians
CRTM
(REL-1.1) T
Surface O3 Q T

7. Observation Error: Tuning of statistics
NCEP
ECMWF
NCEP (and most ECMWF) observation errors statistics consistent with innovations
Error factor tuning from objective method (Desrozier and Ivanov, 2001)
Channel number`

8. Quality Control & Thinning
Pixel-level QC
Reject limb observations
Reject pixels over land and sea-ice
Cloud/Precipitation detection (NESDIS)
Synergy with imager (AIRS/VIS-NIR)
Channel-level QC
Gross check (innovations <15 K)
First-guess check (innovations < 3o).
Thinning Warmest Field of View
Thinning (120km)
345 active data
Warmest FoV
696 active data

9. Bias Correction: Static and Variational
Modeling of errors in satellite radiances:
Predictors:
Offset
mb thickness
200-50mb thickness
Surface skin temperature
Total column water vapor
Scan, scan2, scan3
Parameters
Cost Function
“Offline” bias correction
“Variational” bias correction

10. VarBC: Issues with regional models
No Inertia Constraint
Inertia Constraint
VarBC Timeseries
Innovations for AIRS window channel #787
After BC
Before BC

11. Parameter estimation: in CRTM & RTTOV
g modulates atmospheric absorption to compensate for:
poor knowledge of gas concentrations (CO2, …)
errors in definition of ISRF
errors in mean absorption coefficient
Gamma sensitivity
Timeseries of  estimations
-1 (%)
Analysis cycle

12. Cloud Detection: MMR scheme
AIRS
2378 channels
From « hole hunting » (identifying clear pixels)…
… to identifying clear channels (insensitive to the cloud).
= Radiance calculated in clear sky
RTM
= Radiance calculated for overcast cloud at level k /
Nk3
Nk2
Cloud fractions Nk are ajusted variationally to fit observations.
Nk1
Vertical Level No
Pixel
Channel Number (LW band)

13. Cloud Detection: Initial validation for AIRS
MODIS NASA Level 2 Product
AIRS Cloud Detection
Cloud Top Pressure (hPa)

14. Current Developments: Cloudy Radiances
Cloud Top Pressure

15. <y-H(xb)> (F/ y) Adjoint of WRF-ARW Forecast
Current Developments: Observation Impact
Observation
(y)
Analysis
(xa)
Forecast
(xf)
WRF-VAR
Data Assimilation
WRF-ARW
Forecast
Model
Define
Forecast
Accuracy
Background
(xb)
Forecast
Accuracy
(F)
Observation Impact
<y-H(xb)> (F/ y)
Observation
Sensitivity
(F/ y)
Analysis
Sensitivity
(F/ xa)
Gradient
of F
(F/ xf)
Adjoint of WRF-ARW Forecast
TL Model
(WRF+)
Adjoint of
WRF-VAR
Data
Assimilation
Derive
Forecast
Accuracy
Background
Sensitivity
(F/ xb)
STATUS: DONE
ONGOING
Obs Error
Sensitivity
(F/ eob)
Bias Correction
Sensitivity
(F/ k)
Figure adapted
from Liang Xu

16. More plans… AIRS/AMSU (v.5) Retrievals over Antarctica
Collocate with COSMIC retrievals
Assess impact in AMPS system
AFWA Cloud Analysis
Introduce cloud hydrometeors in control variable
Study background error covariances for clouds
Include cloud microphysics into WRF-ARW TL/ADJ
Assess the accuracy/linearity of radiative transfer in cloudy conditions
IASI, CrIS

17. Thanks for your attention…