Sky Dome Correction For SeaPRISM and TriOS Above
Water Radiometric Measurements in MERMAID
K. Barker (1), F. Zagolski (2), R. Santer (3), C. Kent (1), J.-P. Huot (4), K. Ruddick (5), G. Zibordi(6)
(1)ARGANS
Ltd, UK; (2)PARBLEU Technologies, Canada; (3)ADRINORD, France; (4)ESA/ESTEC, The Netherlands; (5)MUMM, Belgium; (6)JRC, Italy
INTRODUCTION
The water reflectance, ρw, derived from in-situ measurements represents a key element in the validation of MERIS (MEdium Resolution
Imaging Spectrometer) atmospheric corrections over ocean. The ESA/MERMAID (MERis MAtchup In situ Database) supports this validation
effort with concurrent MERIS/in-situ ρw matchups (> 30 sites). The water-leaving radiance, Lw, needs to be corrected for the Fresnel
reflection of the sky dome by the sea surface. In the standard protocol, the neglect of polarization induces an error on ρw.
An alternative correction, fully accounting for polarization processes (i.e., sea surface reflection and atmospheric scattering) has been
developed to be implemented into two new sky dome correction processors; the latters have been verified with the MERMAID in-situ
processing chain for the AERONET-OC SeaPRISM sensor at 8 centre wavelengths (412-1020 nm) and the MUMM* hyperspectral TriOS
instrument convoluted to the 13 MERIS band filters.
The impact of this new correction on the MERIS level-2 product over ocean has been evaluated by validating the ρw retrieval in MERMAID.
SKYDOME CORRECTION FOR SEAPRISM & TriOS
The impact of neglecting polarization the polarization is well
illustrated in Santer et al., 2012, and a simulator, namely,
POLREF (Zagolski et al., 2012), has then been developed to
produce this new Fresnel reflection coefficient.
0.40
The recommendation to ESA:
→ To include the polarization in
the correction of above water
measurements for the skydome
reflection.
0.25
0.35
Rayleigh
Rayleigh
AOT560=0.3
AOT560=0.3
AOT560=0.6
AOT560=0.6
0.20
AOT560=0.9
AOT560=0.9
0.30
0.25
0.15
dρ sky [%]
dρ sky [%]
In-situ Lw measurements are collected under geometric
conditions favourable to the appearance of a significant
polarization effect. Neglecting the polarization may induce a
bias of several percents in relative error on the marine ρw,
depending on the illumination geometry, the state of sea
surface and the atmosphere.
Rayleigh scattering in the blue is a major source of impact,
due to its high polarization. In the near-infrared (NIR) region,
the polarization effect is less important because of reduced
light scattering. Aerosols may also impact the polarization
process.
0.20
0.15
0.10
0.10
0.05
0.05
0.00
0.00
0
10
20
30
40
50
60
70
0
SZA [deg.]
10
20
30
● Bias increasing with SZA
● Stronger bias in visible than in NIR, and
60
70
not strongly aerosol dependent.
RESULTS OF MERMAID IMPLEMENTATION
SeaPRISM
TriOS
0.038
IMPACT OF WIND SPEED ON Rpol
0.048
865 nm
0.046
0.036
Mobley
Rmobley is driven by SZA & ws.
Assumed to include Sun glint
Rmobley is overestimated
compared to Rpol (SeaPRISM
processor) and underestimated
compared to Rpol (TriOS processor)
0.044
0.042
Fresnel Reflection Coefficient
0.034
R pol
0.032
0.030
MERIS LUTs of upward/downward radiances pre-computed at bottom of atmosphere
(BOA) have been extracted from SEAPOL-DB (Zagolski et al., 2012) for the SeaPRISM
and TriOS/MUMM view geometries (VZA=40 deg., RAA=90 and 135 deg., respectively)
0.028
0.040
0.038
0.036
0.034
0.032
0.030
0.028
0.026
869 nm
0.026
Mobley
0.024
0.024
0.0
2.5
5.0
7.5
10.0
12.5
15.0
0
1
2
3
4
5
Wind-Speed [m/s]
SEAPRISM PROCESSOR
6
7
8
9
10
11
12
13
14
15
Wind-Speed [m/s]
0.032
IMPACT OF SZA ON Rpol
0.048
510 nm
0.046
0.030
All these quantities are extracted from MERIS level-2 product
0.042
0.026
0.040
0.038
Rpol
0.024
0.022
Comparisons have been made
between the standard method
(Mobley, 1999) and this new
correction with in-situ data
acquired over AAOT (Venice)
with SeaPRISM, and over the
Northern sea with TriOS.
0.036
0.034
0.032
0.020
0.030
0.018
0.028
0.016
0.026
0.024
0.014
20
25
30
35
40
45
50
55
60
65
70
15
75
20
25
30
35
40
45
50
55
60
65
70
75
SZA [deg.]
SZA [deg.]
3.6E-03
Δρw: STANDARD (POLREF) vs. NEW
1.5E-03
3.2E-03
1.3E-03
2.8E-03
1.1E-03
2.4E-03
9.0E-04
dρ w (442.5 nm )
dρ w (412 nm )
2.0E-03
1.6E-03
1.2E-03
8.0E-04
7.0E-04
5.0E-04
3.0E-04
4.0E-04
1.0E-04
0.0E+00
-1.0E-04
-4.0E-04
-3.0E-04
-8.0E-04
-1.2E-03
-5.0E-04
20
25
30
35
40
45
50
55
60
65
70
15
20
25
SZA [deg.]
35
40
45
50
55
60
65
0.05
Standard Protocol
Standard Protocol
SeaPRISM Processor
TriOS Processor
0.04
ρw(412): NEW IN-SITU vs MERIS
As expected, the new sky dome
correction slightly improve the
water reflectances when they
are compared with the MERIS
matchups.
0.04
ρ w, In-Situ (442.5 nm)
ρ w, In-Situ (490 nm)
30
SeaPRISM: Rpol < Rmobley Larger
values of new in-situ ρw are
expected. Effect increases with
SZA in the SeaPRISM geometry.
TriOS: Rpol > Rmobley Smaller values
of new in-situ ρw are expected.
SZA [deg.]
0.05
Then, for each pixel within a
MERIS-RR (5 x 5) pixels window,
the mean Rpol (λSeaPRISM) and
Rpol (λTriOS) are output by the
processors to be used for the
skydome correction of above
water radiometric measurements.
Increasing SZA decreases the
effect of roughness on the
Fresnel reflection; therefore at
the particular geometry of the
SeaPRISM (RAA=90 deg.) Rpol
decreases as SZA increases.
0.044
● the ECMWF wind-speed
● the atmospheric model
observed with MERIS
(2 bracketing models)
● the aerosol optical thickness
at 865 nm,
● the Angstroem exponent
● the aerosol mixing rate
865 nm
0.028
R pol (441 nm )
The processors derive, for each
SeaPRISM/TriOS-MUMM
matchup data sequence, Rpol
from:
MERMAID Team and Service (ESA, ACRI-ST, ARGANS) for provision of L2 extraction
data files for analysis.
50
Example of bias on the reflected sky radiance by a wind-roughened black sea
surface (5 m/s) induced by a standard protocol ignoring polarization.
Computations have been completed at 442.5 nm (left side) and 753.75 nm (right
side) for 4 atmospheres: AOT550=0, 0.3, 0.6 and 0.9 (α=-1.5).
Objectives:
1. Produce new Fresnel reflection coefficients (Rpol) accounting for polarization
(surface reflection + atmospheric scattering) at the 8 (SeaPRISM) and 13 (TriOS) λ
2. Compute new water-leaving radiances at the 8 (SeaPRISM) and 13 (TriOS) λ,
corrected for: (i) BOA reflected Sun glint radiance and (ii) Sky dome reflection
ACKNOWLEDGEMENTS
40
SZA [deg.]
SKYDOME CORRECTION PROCESSORS FOR MERMAID
TRIOS PROCESSOR
Two skydome correction processors
have been developed to perform
operational sky dome correction in
MERMAID for SeaPRISM and TriOS/
MUMM in-situ Lw, accounting for
the polarized nature of:
(i) the incident atmospheric light
(ii) the Fresnel reflection of a
wind-roughened sea surface.
0.03
0.02
0.03
0.02
0.01
0.01
0.00
0.00
0.00
0.00
0.01
0.02
0.03
0.04
0.05
ρ w,MERIS (490 nm )
0.01
0.02
0.03
0.04
0.05
ρ w,MERIS (442.5 nm )
CONCLUSIONS
 This work is essential for the MERIS CalVal activities and for quality assurance of
both in-situ data and the MERIS derived products over ocean.
 Rpol for AAOT has been validated using the POLREF simulator, and analyses show
that the sky dome correction tools implemented in MERMAID processing could be
operationally applied to other in-situ databases.
 Reprocessed MERIS matchups for SeaPRISM and TriOS/MUMM will be soon
available.
REFERENCES
CONTACT DETAILS
Mobley, C.D., 1999. "Estimation of the remote-sensing reflectance from above-surface measurements", Applied Optics, 38: 7442–7455..
Santer, R., and F. Zagolski, 2012. "Correction of the water-leaving radiance for the Fresnel reflection of the sky dome accounting for the polarization – ATBD",
ADRINORD Report-D01: ' SEAPOL – Sea Polarization’, Lille, France, January 2012: 35 p.
Zagolski, F., and R. Santer, 2012. "POLREF – Simulator of Fresnel reflection coefficient of a wind-roughened black sea surface", ADRINORD Report-D02: '
SEAPOL – Sea Polarization’, Lille, France, March 2012: 94 p
kbarker@argans.co.uk
Francis_Zagolski@yahoo.ca
Richard.Santer@yahoo.fr
3rd MERIS/(A)ATSR & OLCI/SLSTR Preparatory workshop. ESRIN.
15-19 October 2012