Aucun titre de diapositive

1
Vicarious calibration of MERIS level-1b
observations Early results obtained at the
Villefranche AERONET site
David Antoine, Malik Chami Laboratoire
dOcéanographie de Villefranche, CNRS-UPMC, BP 8
Quai de la Darse, 06238 Villefranche sur mer,
FRANCE
This work has been performed in the frame of the
BOUSSOLE project, which is funded / supported by
the following Agencies, programs and institutes
2
Vicarious radiometric calibration What is it ?

• In our Case
• Simulating the Top-of-the-atmosphere (TOA) total
  radiance (Lt), using radiative transfer
  computations that are fed by atmosphere optical
  properties collected in situ
• Comparing this synthetic radiometric target to
  what is measured at the TOA by the sensor
• Determining if vicarious calibration is needed

3
Vicarious radiometric calibration Why ?
The requirement for modern ocean colour sensors
is to derive the water-leaving radiance with a 5
accuracy in the blue for an oligotrophic ocean
(Gordon, 1997 Antoine and Morel, 1999). In other
words, this is corresponding to a 0.002 accuracy
in reflectance This is meaning something like a
1 accuracy on the top-of-the-atmosphere total
radiances
To meet this stringent requirement, several
complementary calibration paths (steps) are to be
followed

• Pre launch radiometric calibration and
  instrument characterisation
• Pre launch spectral calibration
• Post launch radiometric calibration and
  instrument characterisation (when possible)
• Post launch vicarious calibration

4
Vicarious calibration why the requirement is
so high ? (1/2)
All oceanic waters Are between these two lines
5
Vicarious calibration why the requirement is
so high ? (2/2)
Temporal changes were actually due to a slight
calibration drift at 670 nm Would be more or less
impossible to detect without vicarious
calibration procedures
CZCS data, L.O.V./RSMAS re-processing effort
6
Vicarious radiometric calibration Importance ?
Examples of the varying importance of the
vicarious calibration process

• POLDER-2 no on-board calibration capability ?
  Vicarious calibration is fundamental and is the
  only possibility.

• SeaWiFS capability for tracking time changes
  of the calibration (aiming at the moon), but no
  onboard devices for absolute cal. ? Vicarious
  calibration is definitely needed to adjust the
  prelaunch calibration coefficients (to which the
  time changes are applied)

• MERIS onboard capability for absolute
  calibration and for tracking time changes of the
  calibration (diffusers) ? This is providing a
  solid basis for onboard autonomous calibration,
  and vicarious calibration must be seen in that
  case as another, independent, element of the
  overall post-launch calibration process.

7
Vicarious radiometric calibration How ?
Reconstruction of the Top-of-the-atmosphere (TOA)
total radiance, to be compared to its observed
value
At sensor
Note what is described here is different from
the vicarious calibration of a sensor like SeaWiFS
8
Tentative error budget of the vicarious
calibration 1 near infrared domain (typically
at 865 nm no marine signal over Case 1 waters)
9
Tentative error budget of the vicarious
calibration 2 visible domain (e.g., 443 nm for
an oligotrophic ocean Lw is 15 of Lt)
10
The Villefranche AERONET site (1) location
11
The Villefranche AERONET site (2) How it looks
like
Our lab
CIMEL CE-318 sun photometer, part of the
AERONET (Holben et al., 1998)
Altitude 130 meters
12
The Villefranche AERONET site (3) AOT time
series (http//aeronet.gsfc.nasa.gov)
2003
2002
Operation are re-starting mid october 2003 with a
new sun photometer
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
Inverting the measurements of the sky radiance
and of the degree of polarisation

• Based on measurements in the principal plane
  (large scattering angles are reachable) the use
  of a RT code
• Using radiances measurements the closest to the
  satellite pass
• Includes all multiple scattering effects
• Retrieval of the best candidate aerosol model
  based on the sky radiances distribution and the
  spectral optical thickness (l 870, 670 440
  nm) and the degree of polarization (at 870 nm)
• Using the degree of polarization is a tentative
  for a more realistic retrieval of the aerosol
  models in terms of their index of refraction.
  Ambiguities on the aerosol models sometimes can
  be resolved thanks to the degree of polarization.
• For the moment trial and error method

19
Radiative transfer computations The code that
is used
OSOA Chami, Santer Dilligeard, Applied
Optics, 2001 (heritage OS code from Deuzé,
Herman Santer, JQRST, 1989, plus an ocean
compartment -- not used in the present
work) Main characteristics

• Successive orders of scattering code
• Includes polarisation
• Inputs aerosol IOPs (either Shettle Fenn,
  1979, or Jünge models)
• Outputs radiances and degree of polarisation
  at all angles
• Flat air-sea interface

20
Example of the retrieval the aerosols properties
Sky radiances (measured dots computed dashed
line)
Polarization rate (measured dots computed
dashed line)
Best fit is estimated on the basis of scattering
angles gt 90
21
MERIS Vicarious calibration Results using
Jünge aerosol models
UPD ()
22
MERIS Vicarious calibration Results using
Shettle Fenn (1979) aerosols
23
MERIS Vicarious calibration our plans for the
near future

• Completing the time series
• A new sun photometer is installed this week.
• All possible situations will be considered, in
  case MERIS data are available
• Collaboration with the MOBY project (in-water
  buoy radiometry measurement an AERONET site)
  will allow different types of atmospheres to be
  considered.

Extending the exercise to the visible wavelengths
Additional difficulty is to get simultaneous
in-water radiometry, so as to reconstruct the
total TOA reflectance as the sum of the
atmospheric signal (same technique than for the
near IR bands) and the marine signal multiplied
by a diffuse transmittance.
Using other RT codes in addition to the OSOA
(L.O.V. Monte Carlo, F.U.B. MOM, RTM ?) in order
to improve the uncertainty budget
Possibly deriving vicarious calibration
coefficients, and/or identifying/confirming
temporal trends in the calibration, in
collaboration with other activities of the
calibration team.
24
Acknowledgements
AERONET (Holben et al., 1998) CNES, for lending
of a sun photometer before we got the capability
to acquire our own instrument, and specifically
François CABOT, CNES, for installation of this
sun photometer Francis Louis, Alec Scott, for
bi-monthly servicing of the sun
photometer ACRI-st Brockmann consult for MERIS
data processing and distribution
25
End of presentation Thank you for your attention