Direct fitting of NO2 from GOME1, GOME2, SCIAMACHY, and OMI K' Chance1, T'P' Kurosu1, R'V' Martin1,2

1
Direct fitting of NO2 from GOME-1, GOME-2,
SCIAMACHY, and OMIK. Chance1, T.P. Kurosu1,
R.V. Martin1,2, T. Beck3, and S.
Kondragunta31CfA-SAO 2Dalhousie U. 3NOAA/NESDIS
2
Abstract

We have implemented direct fitting of radiance
spectra from the GOME-1, SCIAMACHY, OMI, and
GOME-2 satellite spectrometers. We are attempting
to analyze the spectra from these instruments in
as identical a fashion as possible, with the most
complete treatment of underlying algorithm
physics, in order to separate instrumental,
algorithmic, and temporal differences in
satellite measurements. Initial efforts
concentrate on NO2 slant column measurements.
3
OMI NO2
Tropospheric Column NO2 (Sector Method) July 2005
Total Column NO2 (geometric AMF) July 2005
Cloud screening cloud fraction 20
4
OMI NO2
5
OMI NO2
6
Limb-nadir matching orbits 2509-2510
Largest value over Athens, Greece 5.4?1015 cm-2
vertical column density
7
GOME/SCIAMACHY/OMI/GOME-2
8
Best fitting 2?10-4 FS
9
Algorithm Overview

• Requires precise (dynamic) wavelength (and often
  slit function) calibration, Ring effect
  correction, undersampling correction, and proper
  choices of reference spectra (HITRAN!)
• Best trace gas column fitting results come from
  direct fitting of radiances, (except for OE
  tropospheric ozone and SO2)

10
Algorithm Overview

• Slant column abundances from direct fitting of
  radiances by nonlinear least-squares fitting
• - Simple Ring effect formulation (no induced
• Fraunhofer structure or induced
  wavelength mismatch)
• - No distortion of measured data (no high-pass
  filtering)
• Correction for
• - Wavelength calibration
• - Instrument transfer (slit) function
• - Ring effect
• - Spectral undersampling (GOME-n, SCIA, OMI,
  OMPS do not Nyquist sample the spectra)
• Division by air-mass factor (AMF) using LIDORT
  radiative transfer model and GEOS-CHEM 3-D
  tropospheric chemistry and transport model to
  determine vertical column abundances
• - Tropospheric residuals may require further
  adjustment (e.g., for
• NO2)
• Ozone profiles Direct fitting for profile using
  optimal estimation

11
Fitting approach Nonlinear least-squares fitting
of radiances with lots of optimization
Radiance R is fitted directly (BOAS fitting) as
N.B. ??!
Other approaches Division by I0
Further manipulation, for Beers law fitting
gives
But note It is NOT a linear fitting problem!
DOAS fitting adds high-pass filtering (H) to
give
12
Kurucz sun ?
NO2
Reference spectra for determination of NO2
from SCIAMACHY
Ring
O3
O2-O2
H2O
H2O Ring
Undersampling
13
High resolution solar reference spectrum
14
Top-of-atmosphere solar spectral irradiance

• The high resolution solar spectral irradiance is
  critical in analyzing atmospheric trace gases
• Solar lines are source of accurate wavelength
  calibration (?0.0003-0.0004 nm for GOME!) Our
  method now used operationally on GOME, SCIAMACHY,
  OMI, and OMPS
• Determination of the Ring effect
• Improved knowledge of instrument slit functions
• Partial correction for spectral undersampling
• Photochemistry of Schumann-Runge system
• A space-based determination would be an ideal
  support mission for 12 international atmospheric
  missions!
• Range 240-1000 nm
• FWHM 0.01 nm or better
• Ideal FTS Space Shuttle experiment Canadian
  experiment

15
Ring effect calculation Multiple-scattering
versions and molecular interference by successive
orders of absorption implemented
K. Chance and R.J.D. Spurr, Ring effect studies
Rayleigh scattering, including molecular
parameters for rotational Raman scattering,
and the Fraunhofer spectrum, Appl. Opt.
36, 5224-5230, 1997. (www.cfa.harvard.edu/atmosphe
re/) New, improved, version now
available! Better absolute accuracy Less gas
interference Better sampling Ask me for details
16
Examples from March 12, 2007
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
Conclusions and Future Directions

• Correction of SCIAMACHY L1b difficulties
• Improved OMI stripe correction
• Improved vertical column determination
• Validation to separate temporal from instrument
  differences

21
The End!