Since 18 December 2004, the PARASOL satellite is a member of the so-called A-train atmospheric orbital observatory, flying together with Aqua, Aura, CALIPSO, CLOUDSAT, and OCO satellites. These satellites combine for the first time a full suite of instruments for observing aerosols and clouds, using passive radiometer complementarily with active lidar and radar sounders. The PARASOL payload is extensively derived from the instrument developed for the POLDER programs that performs measurements of bidirectionality and polarization for a very wide field-of-view and for a visible∕near-infrared spectral range. An overview of the results obtained during the commissioning phase and the reevaluation after one year in orbit is presented. In-flight calibration methods are briefly described, and radiometric and geometric performances are both evaluated. All algorithms are based on a panel of methods using mainly natural targets previously developed for POLDER missions and adapted or redeveloped in the PARASOL context. Regarding performances, all mission requirements are met except for band 443 (not recommended for use). After one year in orbit, a perfect geometrical stability was found while a slight decrease of the radiometric sensitivity was observed and corrected through an innovative multitemporal algorithm based on observations of bright and scattered convective clouds. The scientific exploitation of PARASOL has now begun, particularly by coupling these specific observations with other A-train sensor measurements.
J. Martijn Smit, Jeroen H. H. Rietjens, Gerard van Harten, Antonio Di Noia, Wouter Laauwen, Brian E. Rheingans, David J. Diner, Brian Cairns, Andrzej Wasilewski, Kirk D. Knobelspiesse, Richard Ferrare, and Otto P. Hasekamp Appl. Opt. 58(21) 5695-5719 (2019)
You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
You do not have subscription access to this journal. Figure files are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
You do not have subscription access to this journal. Article tables are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
You do not have subscription access to this journal. Equations are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
Main Characteristics for PARASOL Satellite and Payload
Launch date
18th December 2004
Platform
Myriad
Altitude
705 km
Local time
13h30
Mass
120 kg
Size
0.6 × 0.8 × 0.8 m
Instrument
POLDER
Spectral bands
9
Polarized bands
3
Spectral range
443–1020 nm
Detector
CCD Matrix 242 × 274
Swath
1600 km cross-track
2100 km along-track
Resolution
6.18 km (level-1 grid)
Field-of-view
±57°
Table 2
PARASOL Spectral Bands Including Central Wavelength, Bandwidth, Ability to Measure Polarization, and Saturation Level in Reflectance Unit
Spectral Band
443
490
565
670
763
765
865
910
1020
Central wavelength (nm)
443.5
490.9
563.8
669.9
762.9
762.7
863.7
907.1
1019.6
Bandwidth (nm)
13.4
16.3
15.4
15.1
10.9
38.1
33.7
21.1
17.1
Polarization
—
yes
—
yes
—
—
yes
—
—
Saturation level (reflectance)
1.28
0.99
1.06
1.01
1.04
0.96
1.00
0.98
1.70
Table 3
Summary of the Calibration Results Ak for Rayleigh, Sunglint, Clouds, and Deserts Methodsa
Spectral Band
443
490
565
670
763
765
865
910
1020
Preflight
1.000
0.959
0.964
0.947
0.950
0.930
0.926
0.926
0.914
Rayleigh
0.899
1.002
0.999
1.001
Sunglint
0.960
0.994
0.997
1.008
1.005
1.005
1.008
1.004
Clouds
1.047
0.999
1.006
0.995
0.984
0.984
0.989
Deserts
0.970
0.998
1.026
0.996
1.018
0.999
%∕year
−4.05
−2.98
−1.70
−0.99
0.00
0.00
−0.40
−0.40
−1.50
When the computed radiance is in accordance with the measured radiance, . The interband calibration results (sunglint and clouds methods) are reported after a normalization to . The preflight calibration values are also reported. The temporal drift of the radiometric sensitivity is calculated in % per year.
Table 4
Summary of Methods Used to Calibration all PARASOL Spectral Bands for Both Absolute and Temporal Aspectsa
Calibration
Method
Spectral Band
490
565
670
763
765
865
910
1020
Absolute
Rayleigh (ray)
ray
ray
ray
Interband
Sunglint (sun)
ref.
ref.
ref.
sun
sun
sun
865
sun
Multitemporal
Clouds (cld)
cld
cld
cld
765
ref.
cld
865
cld
Methods are absolute calibration over Rayleigh scattering (ray), interband calibration over sunglint (sun), and multitemporal calibration over convective clouds (cld). “ref” means that spectral bands are used as reference (temporal or absolute) to calibrate other spectral bands. When 765 or 865 is reported, this band is used a reference for inter-calibration (for 763 or 910 respectively).
Table 5
In-Flight Noise Estimation Using Dark Current Acquisitions, Dark Ocean Homogeneous Areas, and Analysis of the U Stokes Parameter in the Principal Plane, Compared to Specified Values and Preflight Previsions for Corresponding Spectral Bands and Normalized Radiance Levels
Spectral Band
490
565
670
765
865
1020
Dark current
I = 0
NedI
0.00009
0.00010
0.00009
0.00010
0.00011
0.00018
Preflight
0.00023
0.00023
0.00023
0.00023
0.00023
0.00023
Spec.
0.00050
0.00050
0.00050
0.00050
0.00050
0.00050
Dark ocean
I ≤ 0.1
NedI
0.00047
0.00029
0.00042
Preflight
0.00025
0.00024
0.00033
Spec.
0.00050
0.00050
0.00050
Ocean
Ip < 0.1
NedIp
0.00085
0.00085
0.00042
Preflight
0.00052
0.00051
0.00050
Spec.
0.00100
0.00100
0.00100
Clouds
Ip ∼ 0.1
NedIp
0.00255
0.00325
0.00269
Preflight
0.00113
0.00267
0.00085
Spec.
0.00100
0.00100
0.00100
Table 6
In-Flight Geometrical Calibration of the Alignment Matrix (Roll, Pitch, and Yaw) of the Instrument in Microradiansa
Measure (in microradian)
Roll
Pitch
Yaw
Initial calibration
−1176.3
−1023.1
−151.6
Iteration 1
−429.9
−957.7
−481.9
Stdev
117.2
39.3
63.8
Iter1—initial
746.4
65.4
−330.3
Iteration 2
−317.2
−955.1
−521.7
Stdev
103.2
42.1
48.9
Iter2—iter1
112.7
2.6
−39.8
Mean and standard deviation values found for the two iterative steps using space triangulation algorithm are compared to the initial preflight calibration.
Table 7
Multipolarization Registration Performance for 490, 670, and 865 Polarized Tripletsa
Measure (in pixel)
Npts
490
670
865
Mean
Stdev
Mean
Stdev
Mean
Stdev
Range [0°–35°]
225959
0.0312
0.0042
0.0455
0.0010
0.0204
0.0031
Range [35°–45°]
223490
0.0263
0.0045
0.0458
0.0014
0.0263
0.0099
Range [45°–50°]
95421
0.0317
0.0070
0.0496
0.0042
0.0357
0.0130
Performance at 50°
0.0347
0.0530
0.0431
Specification
0.0500
0.0500
0.0500
Mean, standard deviation and number of measurements are reported for 3 angular ranges, for the 50° viewing angle, and confronted to the specification expressed in pixel.
Table 8
Multi-Spectral Registration Performance Estimated for 8 Couples of Neighbor Spectral Bandsa
Bands
March 2005
March 2006
Pixel
Npts
Stdev
Pixel
Npts
Stdev
443–490
0.0778
40686
0.0158
0.0746
68847
0.0195
490–565
0.0895
33672
0.0147
0.0478
61191
0.0075
565–670
0.0924
31639
0.0134
0.0907
52080
0.0170
670–763
0.0607
34984
0.0122
0.0693
43635
0.0145
763–765
0.0342
50774
0.0046
0.0325
82513
0.0045
765–865
0.0520
47497
0.0155
0.0511
78461
0.0182
865–910
0.0627
44134
0.0099
0.0618
74917
0.0128
910–1020
0.0794
38646
0.0198
0.0797
66373
0.0246
Specification
0.1000
0.1000
Mean standard deviation, and number of measurements are reported for the commissioning phase in March 2005 and for the annual verification in March 2006. Values are confronted to the specification in pixel.
Table 9
Multiangular Registration Performance Estimated for 7 Turn Numbers of the Filter Wheela
Number of Directions
March 2005
March 2006
Pixel
Npts
Stdev
Pixel
Npts
Stdev
3
0.0836
60077
0.071
0.0851
31484
0.067
5
0.0909
38999
0.074
0.0917
19374
0.067
8
0.0981
20839
0.074
0.1074
9568
0.073
11
0.1219
1932
0.082
0.1255
2468
0.081
12
0.1339
1096
0.085
0.1318
1388
0.087
13
0.1628
326
0.089
0.1529
419
0.098
14
0.1370
3
0.056
0.2530
5
0.104
Specification
0.1000
0.1000
Mean standard deviation, and number of measurements are reported for the commissioning phase in March 2005 and for the annual verification in March 2006. Values are confronted to the specification in pixel. Results for 14 viewing directions are considered as nonsignificant.
Table 10
Absolute Location Accuracy Estimation Using Vegetation SPOT5 as Referencea
Npts
Mean
Stdev
Min
Max
X (in pixel)
96
−0,193
0,242
0,683
−0,610
Y (in pixel)
96
0,097
0,256
0,718
−0,413
Norm (in pixel)
96
0,366
0,191
0,832
0,040
Norm (km)
96
2,26
1,18
5,14
0,24
Specification (km)
4,00
Objective (km)
2,00
Mean, standard deviation, minimum and maximum values, and number of measurements are reported for X axis (along track), Y axis (cross track), and norm of the deviation in pixel and in km, and finally confronted to specifications.
Tables (10)
Table 1
Main Characteristics for PARASOL Satellite and Payload
Launch date
18th December 2004
Platform
Myriad
Altitude
705 km
Local time
13h30
Mass
120 kg
Size
0.6 × 0.8 × 0.8 m
Instrument
POLDER
Spectral bands
9
Polarized bands
3
Spectral range
443–1020 nm
Detector
CCD Matrix 242 × 274
Swath
1600 km cross-track
2100 km along-track
Resolution
6.18 km (level-1 grid)
Field-of-view
±57°
Table 2
PARASOL Spectral Bands Including Central Wavelength, Bandwidth, Ability to Measure Polarization, and Saturation Level in Reflectance Unit
Spectral Band
443
490
565
670
763
765
865
910
1020
Central wavelength (nm)
443.5
490.9
563.8
669.9
762.9
762.7
863.7
907.1
1019.6
Bandwidth (nm)
13.4
16.3
15.4
15.1
10.9
38.1
33.7
21.1
17.1
Polarization
—
yes
—
yes
—
—
yes
—
—
Saturation level (reflectance)
1.28
0.99
1.06
1.01
1.04
0.96
1.00
0.98
1.70
Table 3
Summary of the Calibration Results Ak for Rayleigh, Sunglint, Clouds, and Deserts Methodsa
Spectral Band
443
490
565
670
763
765
865
910
1020
Preflight
1.000
0.959
0.964
0.947
0.950
0.930
0.926
0.926
0.914
Rayleigh
0.899
1.002
0.999
1.001
Sunglint
0.960
0.994
0.997
1.008
1.005
1.005
1.008
1.004
Clouds
1.047
0.999
1.006
0.995
0.984
0.984
0.989
Deserts
0.970
0.998
1.026
0.996
1.018
0.999
%∕year
−4.05
−2.98
−1.70
−0.99
0.00
0.00
−0.40
−0.40
−1.50
When the computed radiance is in accordance with the measured radiance, . The interband calibration results (sunglint and clouds methods) are reported after a normalization to . The preflight calibration values are also reported. The temporal drift of the radiometric sensitivity is calculated in % per year.
Table 4
Summary of Methods Used to Calibration all PARASOL Spectral Bands for Both Absolute and Temporal Aspectsa
Calibration
Method
Spectral Band
490
565
670
763
765
865
910
1020
Absolute
Rayleigh (ray)
ray
ray
ray
Interband
Sunglint (sun)
ref.
ref.
ref.
sun
sun
sun
865
sun
Multitemporal
Clouds (cld)
cld
cld
cld
765
ref.
cld
865
cld
Methods are absolute calibration over Rayleigh scattering (ray), interband calibration over sunglint (sun), and multitemporal calibration over convective clouds (cld). “ref” means that spectral bands are used as reference (temporal or absolute) to calibrate other spectral bands. When 765 or 865 is reported, this band is used a reference for inter-calibration (for 763 or 910 respectively).
Table 5
In-Flight Noise Estimation Using Dark Current Acquisitions, Dark Ocean Homogeneous Areas, and Analysis of the U Stokes Parameter in the Principal Plane, Compared to Specified Values and Preflight Previsions for Corresponding Spectral Bands and Normalized Radiance Levels
Spectral Band
490
565
670
765
865
1020
Dark current
I = 0
NedI
0.00009
0.00010
0.00009
0.00010
0.00011
0.00018
Preflight
0.00023
0.00023
0.00023
0.00023
0.00023
0.00023
Spec.
0.00050
0.00050
0.00050
0.00050
0.00050
0.00050
Dark ocean
I ≤ 0.1
NedI
0.00047
0.00029
0.00042
Preflight
0.00025
0.00024
0.00033
Spec.
0.00050
0.00050
0.00050
Ocean
Ip < 0.1
NedIp
0.00085
0.00085
0.00042
Preflight
0.00052
0.00051
0.00050
Spec.
0.00100
0.00100
0.00100
Clouds
Ip ∼ 0.1
NedIp
0.00255
0.00325
0.00269
Preflight
0.00113
0.00267
0.00085
Spec.
0.00100
0.00100
0.00100
Table 6
In-Flight Geometrical Calibration of the Alignment Matrix (Roll, Pitch, and Yaw) of the Instrument in Microradiansa
Measure (in microradian)
Roll
Pitch
Yaw
Initial calibration
−1176.3
−1023.1
−151.6
Iteration 1
−429.9
−957.7
−481.9
Stdev
117.2
39.3
63.8
Iter1—initial
746.4
65.4
−330.3
Iteration 2
−317.2
−955.1
−521.7
Stdev
103.2
42.1
48.9
Iter2—iter1
112.7
2.6
−39.8
Mean and standard deviation values found for the two iterative steps using space triangulation algorithm are compared to the initial preflight calibration.
Table 7
Multipolarization Registration Performance for 490, 670, and 865 Polarized Tripletsa
Measure (in pixel)
Npts
490
670
865
Mean
Stdev
Mean
Stdev
Mean
Stdev
Range [0°–35°]
225959
0.0312
0.0042
0.0455
0.0010
0.0204
0.0031
Range [35°–45°]
223490
0.0263
0.0045
0.0458
0.0014
0.0263
0.0099
Range [45°–50°]
95421
0.0317
0.0070
0.0496
0.0042
0.0357
0.0130
Performance at 50°
0.0347
0.0530
0.0431
Specification
0.0500
0.0500
0.0500
Mean, standard deviation and number of measurements are reported for 3 angular ranges, for the 50° viewing angle, and confronted to the specification expressed in pixel.
Table 8
Multi-Spectral Registration Performance Estimated for 8 Couples of Neighbor Spectral Bandsa
Bands
March 2005
March 2006
Pixel
Npts
Stdev
Pixel
Npts
Stdev
443–490
0.0778
40686
0.0158
0.0746
68847
0.0195
490–565
0.0895
33672
0.0147
0.0478
61191
0.0075
565–670
0.0924
31639
0.0134
0.0907
52080
0.0170
670–763
0.0607
34984
0.0122
0.0693
43635
0.0145
763–765
0.0342
50774
0.0046
0.0325
82513
0.0045
765–865
0.0520
47497
0.0155
0.0511
78461
0.0182
865–910
0.0627
44134
0.0099
0.0618
74917
0.0128
910–1020
0.0794
38646
0.0198
0.0797
66373
0.0246
Specification
0.1000
0.1000
Mean standard deviation, and number of measurements are reported for the commissioning phase in March 2005 and for the annual verification in March 2006. Values are confronted to the specification in pixel.
Table 9
Multiangular Registration Performance Estimated for 7 Turn Numbers of the Filter Wheela
Number of Directions
March 2005
March 2006
Pixel
Npts
Stdev
Pixel
Npts
Stdev
3
0.0836
60077
0.071
0.0851
31484
0.067
5
0.0909
38999
0.074
0.0917
19374
0.067
8
0.0981
20839
0.074
0.1074
9568
0.073
11
0.1219
1932
0.082
0.1255
2468
0.081
12
0.1339
1096
0.085
0.1318
1388
0.087
13
0.1628
326
0.089
0.1529
419
0.098
14
0.1370
3
0.056
0.2530
5
0.104
Specification
0.1000
0.1000
Mean standard deviation, and number of measurements are reported for the commissioning phase in March 2005 and for the annual verification in March 2006. Values are confronted to the specification in pixel. Results for 14 viewing directions are considered as nonsignificant.
Table 10
Absolute Location Accuracy Estimation Using Vegetation SPOT5 as Referencea
Npts
Mean
Stdev
Min
Max
X (in pixel)
96
−0,193
0,242
0,683
−0,610
Y (in pixel)
96
0,097
0,256
0,718
−0,413
Norm (in pixel)
96
0,366
0,191
0,832
0,040
Norm (km)
96
2,26
1,18
5,14
0,24
Specification (km)
4,00
Objective (km)
2,00
Mean, standard deviation, minimum and maximum values, and number of measurements are reported for X axis (along track), Y axis (cross track), and norm of the deviation in pixel and in km, and finally confronted to specifications.