WP 3: Absorbing Aerosol Index (AAI) WP 10: Level-1 validation

1
WP 3 Absorbing Aerosol Index (AAI)WP 10
Level-1 validation

• L.G. Tilstra1, I. Aben2, and P. Stammes1
• 1Royal Netherlands Meteorological Institute
  (KNMI)
• 2Netherlands Institute for Space Research (SRON)

SCIAvisie Meeting, SRON, Utrecht, 10-12-2009
2
WP3 scientific AAI (SC-AAI) and operational AAI
(L2-AAI)
Algorithm improvements
SQWG YR 1/2
SQWG YR 3/4

1. correction for calibration offset at t0
2. correction for the obstruction in the FOV for
   westernmost scan mirror positions (problem
   affects data until April 2003)
3. correction for (scan-angle dependent) instrument
   degradation (on top of the
   standard m-factor correction)
4. look-up tables (LUTs) calculated by RTM taking
   polarisation into account
5. completely new algorithm approach more accurate
   allows negative albedos
6. improved surface height calculation method
   database
7. proper flagging of (potential) sunglint
   situations
8. flagging of solar eclipse events
9. viewing and solar angles calculated w.r.t.
   sea-level (instead of w.r.t. 100 km)
10. ozone column dependency of AAI taken into account
11. LUTs calculated by RTM taking the atmospheres
    sphericity into account
12. O2-O2 absorption included in LUTs

3
L2-AAI improvements implementation,
verification, validation
Case 1 ozone column used L2-AAI level-2
ozone SC-AAI TOSOMI
L2-AAI
SC-AAI
Validation data set 180 orbits, 830.000
observations
4
L2-AAI improvements implementation,
verification, validation
Case 2 ozone column fixed to 250 DU
L2-AAI
SC-AAI
Almost identical results. The two outliers error
in footprint coordinates. Switch date 09
December 2009 (L1 v7.02 L2 v5.01)
5
Copenhagen side event SCIAMACHY AAI over the
period 20022009
Siberian forest fires in July 2006
Canadian and Alaskan forest fires June-July 2004
Taklamakan desert
Californian forest fires
Libian desert
Thar desert
Rice straw burning
Desert dust
Sahara
Saudi Arabian lowlands
Bodélé
Sahel biomass burning and desert dust storms
Indonesian forest fires
Amazonian rainforest biomass burning
biomass burning smoke
biomass burning smoke
Smoke and Dust
Smoke from forest fires
Desert dust and smoke particles are globally the
most dominant types of natural aerosols. This map
of events is based on absorbing aerosol index
information from 2002 to 2009 measured by
SCIAMACHY on ESA's Envisat satellite. The
detected aerosols originate mainly from desert
dust storms and biomass burning events, but also
forest fires, such as the ones in Siberia and
Canada, and occasional volcanic eruptions.
SCIAMACHY was built by Germany, The Netherlands
and Belgium.
weak events
strong events
more data and information can be found at
www.temis.nl
6
Plans for WP3

• Keep improving the scientific AAI product
  (SC-AAI)
• Continue validation of SC-AAI and L2-AAI
• Maintain SC-AAI data archive at the TEMIS website
• Further study the scan-angle dependent
  degradation for support of WP10

7
WP10 level-1 validation
Work in progress Validation of the new key data
that were developed by SRON by looking directly
at level-1 reflectances and Stokes fractions.
Plans for WP10

• Verify improvement in calibration brought about
  by the new SRON key data using the various tools
  we developed
• Monitor degradation and analyse quality of
  applied degradation correction (m-factors) in
  the UV using the Absorbing Aerosol Index (AAI)
• Monitor and validate polarisation product using
  special geometries
• Analyse the reflectance over specific stable
  Earth targets

8
Extra slides (R)
9
R1 The Global Dust Belt
10
R2 Introduction of the Absorbing Aerosol Index
(AAI) and the residue
The AAI represents the scene colour in the UV

• Definition of the residue
• where the surface albedo A for the simulations
  is such that
• (A is assumed to be wavelength independent
  A340 A380)
• no clouds, no aerosols r 0
• clouds, no absorbing aerosols r lt 0
• absorbing aerosols r gt 0
• B. Definition of the AAI
• AAI residue gt 0 (and the AAI is not defined
  where residue lt 0)

11
R3 Typical global aerosol distribution
The Global Dust Belt Desert Dust Aerosols
(DDA) (dust storms, all year)
Biomass Burning Aerosols (BBA) (dry season,
anthropogenic)
AAI from other UV satellite instruments TOMS,
GOME-1, GOME-2. Combined with SCIAMACHY there are
more than three decades (19782009) of AAI data
available for studies of trends in desert dust
and biomass burning aerosol.
12
(10) Taking ozone absorption into account in the
simulated reflectances / LUTs
? ozone column fixed to 334 DU
Effect of neglecting ozone on the AAI ?

• New LUTs containing ozone column dependence
• Changes to algorithm code
• Total (SCIAMACHY L2) ozone columns are required
  as input

13
(11) Pseudo-spherical treatment of the
atmospheres sphericity in the LUTs
LUTs of the reflectances at 340 and 380 nm were
originally calculated for plane parallel
atmospheres (using the RTM DAK). We proposed to
improve this. Impact on the AAI
LUTs recreated using DAK v3.1.1, which offers
pseudo-sphericity. Improvement is relatively
large for solar zenith angles above 75.
14
(12) Taking O2O2 absorption into account
Effect of including O2O2 absorption in the AAI
LUTs
O2O2 absorption bands at 360 and 380 nm

• non-negligible effect 0.10.5 index points for
  thick clouds
• offset of 0.1 index points for thin/no clouds
• same offset for positive residues

Effect cannot be neglected. Including O2O2
absorption in the LUTs improves matters at little
or no cost.
15
WP10 level-1 validation
Validation techniques for the reflectance

• comparison with radiative transfer model DAK
  (in the UV)
• comparison with other satellite instruments
  (GOME-1, MERIS, POLDER-2, )
• qualitative analysis of the spectra (spectral
  properties)

calibration offset
spectral features
16
Degradation monitoring using the scientific AAI
product global mean AAI
Without m-factors - increase of 4 index points
With m-factors