Infrared Atmospheric Spectroscopy using AIRS

1
Infrared Atmospheric Spectroscopy using AIRS
Sergio De-Souza Machado, Scott Hannon, L.
Larrabee Strow and Howard Motteler Physics
Department and Joint Center for Earth Systems
Technology University of Maryland Baltimore
County, Baltimore, MD 21250
AIRS High Spectral Resolution and Broad IR
Coverage Provides Ability to Improve existing
infrared molecular spectroscopy
Water vapor continuum coefficients

• Uplooking AERI data collected for several years
  at various sites (thanks, Dave Turner and rest of
  U-Wisc folks!)
• Dedicated radiosonde launches, aircraft campaigns
  to validate AIRS
• RAL laboratory data in the 1300-1600 cm-1 region
• ALL suggest that CKD2.4 is incorrect in some
  spectral regions
• Based on the above, MT-CKD v1.0 has been
  developed
• We have further tuned MT-CKD 1using AIRS data

• AIRS has low noise, high radiometric accuracy
  (0.2 K for 250 K scene temperature)
• AIRS has high reolution (v/dv 1200)
• Physical retrieval accuracy of AIRS products
  depend on radiative transfer model
• Validation of RTA done using dedicated
  radiosonde launches and ECMWF model fields.
• Early analysis showed deficiencies in the CO2
  and H2O regions
• This was confirmed by re-examining lab spectra,
  and studying AERI downwelling measurements at
  DOE-ARM sites
• We have since improved the 4 um CO2 spectroscopy,
  and water vapor continuum
• This should improve clear sky spectroscopy for
  infrared remote sensing community
• Using window channels, can possibly differentiate
  between spieces of scatterers in atmosphere
  (ice/water versus silicates, and possibly between
  different silicate spieces

Optical Depth
CO2 P,R line mixing in the 4 um band
Self (CS) and Foreign(CF) continuum coeffs at 300
K

• Lorentz lineshape severely underestimates
  transmission
• GENLN2 uses Cousin lineshape, which is a chi
  function on top of a Voigt lineshape
• Many order Perturbation theories provide further
  agreement with data, but are complicated
• kCARTA/AIRS RTA use LineMixing lineshape uses
  two parameters, simpler to use

Checking against instrument campaigns and AIRS
data
Comparing different lineshapes for 15 um and 4 um

• RAL lab data enabled us to estimate 4 um P/R line
  mixing parameters
• Apply same parameters in 15 um region, with Q
  branch mix parameters from D. Tobin
• Thousands of AIRS spectra vindicated use of
  linemxing line shape

(clear FOVs only)
March 2003 night time statistics
AIRS spectra

• Tuning for new AIRS Fast Model includes fixes to
  CO2 and water vapor
• New model shows LESS variablilty for wide ranges
  of water amount
• New model also shows less variability for scan
  angles
• Uncertainities in surface emissivity, surface
  temperature and possible cloud contamination
  could affect results

• AIRS indicated some problems still existed eg at
  2380-2400 cm-1 bandhead
• Use this information to tune the optical depths
  to improve agreement check with lab data and
  radiances

Thanks to Dave Tobin and Dave Turner and U-Wisc
folks AIRS Project at JPL AIRS Science Team
GSFC DAAC and BAE for a great instrument!
Summary and Conclusions

• High resolution of AIRS, combined with low noise
  and high radiometric accuracy, has helped us
  identify shortcomings in existing spectroscopy
• Large datasets from AIRS, as well as analyzing
  AERI datasets, and lab data, has improved CO2 and
  H2O continuum spectroscopy
• Tuning some channels has improved obs-calcs
  further
• Further validating spectroscopy using dedicated
  radiosonde launches
• Window channels can help differentiate aerosol
  particles