AIRS and OMI Ozone

1
AIRS and OMI Ozone
Mike Newchurch/UAH Bill Irion/JPL Sunmi Na/Pusan
U Mike Gunson/JPL
AURA Validation Meeting 22 September, 2005 GSFC
2
AIRS scanning geometry
Spectra from the nine AIRS footprints (upper
right) are used to produce cloud-cleared spectra
within the AMSU footprint (45 km) over which T,
H2O, and Ozone are retreived.
3
Comparison of TOMS, OMI and AIRS column ozone

• Level 2 AIRS V4 compared to gridded TOMS (V8) and
  OMI data.
• Daytime columns compared if center of AIRS/AMSU
  footprint falls within TOMS/OMI gridbox (so more
  than one AIRS measurement is often compared.)
• AIRS uses 9.6 ?m ozone radiances.
• Column data for AIRS can be quantitatively
  unreliable poleward of 50 N or S, and over ice
  and desert regions.

4
AIRS daytime ozone
6/23/2005 (AIRS - OMI) / OMI
12/23/2004 (AIRS - OMI) / OMI
5
AIRS - OMI relative difference vs cloud-top
pressure 6/23/05
1st cloud
2nd cloud
6
AIRS - OMI -TOMS latitudinal comparisonDaytime,
Land Ocean
12/23/04 - 90S - 65N
6/23/05 - 65S - 90N
7
Latitudinal difference vs OMI and TOMS Daytime,
Ocean Only
6/23/05 50S - 50N
12/23/04 50S - 50N
8
AIRS and TOMS Polar Night1/7/2003

9
AIRS vs. sondes

10
ECMWF ozone (AIRS a priori) vs. sondes

11
AIRS profile biases leaking in from ECMWF?

• AIRS used ECMWF ozone profiles to train
  regression.
• Biases in ECMWF profiles have been noted,
  although the total column looks good Dethof and
  Hólm, 2004.
• ECMWF biases leaking into AIRS L2 where AIRS
  lacks sensitivity (e.g. lower troposphere.)?
• Not enough good matchups of ozonesondes with AIRS
  to retrain regression.
• Try using TES profiles to retrain regression
• 3-4 independent pieces in TES, 1-2 in
  troposphere.
• A-train makes matchups easy to find and numerous.

Dethof, A. and E. V. Hólm, Ozone assimilation in
the ERA-40 reanalysis project, Q. J. R. Meteorol.
Soc., 130, 2851-2872, 2004.
12
Summary results
13
Conclusions

• Initial comparisons of total ozone column between
  AIRS and OMI as good as or better than that AIRS
  and TOMS.
• Biases in the AIRS profile compared to
  ozonesondes morphologically similar to biases of
  ECMWF.
• Will try re-training regression with TES ozone
  profile and AIRS L2 radiances.