J. P. Doyle (j.doyle@ic.ac.uk) is with the Department of Computational Physics and Geophysics, T. H. Huxley School, Imperial College of Science, Technology, and Medicine, University of London, London SW7 2BP, UK.
G. Zibordi (giuseppe.zibordi@jrc.it) is with the Marine Environment Unit, Space Applications Institute, Joint Research Centre of the European Commission, I-21020 Ispra (Va), Italy.
John P. Doyle and Giuseppe Zibordi, "Optical propagation within a three-dimensional shadowed atmosphere–ocean field: application to large deployment structures," Appl. Opt. 41, 4283-4306 (2002)
Estimation of optical shadowing effects that occur on in situ submerged radiance and irradiance measurements conducted in the proximity of a large and complex three-dimensional deployment structure is addressed by use of Monte Carlo simulations. We have applied backward Monte Carlo techniques and variance reduction schemes in three-dimensional radiative transfer computations of in-water light field perturbations by taking into account relevant geometric, environmental, and optical parameters that describe a realistic atmosphere-ocean system. Significant parameters, determined by a sensitivity analysis study, have then been systematically varied for the computation of an extensive set of correction factors, included in look-up tables designed for operational removal of tower-shading uncertainties, which typically induce an ∼1–10% decrease in absolute radiometric data values near a specific oceanographic tower located in the northern Adriatic Sea. In principle, the proposed correction methodology can be transferred to other deployment systems, instrument casings, and measurement sites if a comprehensive description is provided for the system parameters and their variability.
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.
Comparison of the Gordon and the pho-tran MC Estimatesa
Radiometric Quantity (at 0-)
g = 0
g = 0.750
g = 0.875
g = 0.950
KA
KA′
Gordon
pho-tran
Gordon
pho-tran
Gordon
pho-tran
Gordon
pho-tran
Gordon
pho-tran
Gordon
pho-tran
Lu
0.08122
0.08119
0.02019
0.02021
0.00825
0.00855
0.00256
0.00251
0.01113
0.01069
0.01045
0.01022
Lu:σ
0.00132
0.00014
0.00100
0.00015
0.00115
0.00018
0.00197
0.00022
0.00348
0.00060
0.00061
0.00014
Lu:SRE
0.01630
0.00180
0.04950
0.00710
0.13940
0.02160
0.76950
0.08840
0.31260
0.05570
0.05830
0.01360
(Lu - L̃u)
0.01016
0.01013
0.00249
0.00257
0.00096
0.00103
0.00029
0.00030
0.00157
0.00102
0.00102
0.00098
(Lu - L̃u):σ
0.00049
0.00006
0.00044
0.00006
0.00059
0.00007
0.00068
0.00008
0.00187
0.00010
0.00026
0.00006
(Lu - L̃u):SRE
0.04820
0.00600
0.17670
0.00240
0.61450
0.07120
2.34480
0.26670
1.19100
0.09800
0.25490
0.02330
(Lu - L̃u)/Lu in %
12.50
12.48
12.30
12.72
11.60
12.10
11.30
12.03
14.10
9.53
9.70
9.59
(Eu - Ẽu)/Eu in %
21.30
21.27
14.50
13.74
12.20
12.55
14.20
14.94
14.70
14.90
14.20
14.14
The estimates are of absolute values (means, standard deviations σ, and SRE) of unperturbed and ship-shadow perturbed radiometric quantities. Results are given as a function of the in-water Henyey-Greenstein phase function (variable asymmetry parameters g) and for two separate phase functions (KA and KA′).
Table 2
Comparison of the Gordon and the pho-tran Backward MC Simulationsa
MC Simulation
ϕ0 =
∊Ed (5 m)
∊Ed (40 m)
0°
45°
90°
0°
45°
90°
I Gordon
1.27
1.43
2.36
3.98
5.80
15.88
I pho-tran
1.33
1.52
2.47
3.83
5.76
15.69
III Gordon
0.43
0.60
1.57
4.01
5.71
15.71
III pho-tran
0.40
0.58
1.55
4.05
5.62
16.04
Included are percentage relative error values for ship-shadowed Ed at 5- and 40-m depth and variable solar azimuth. Two simulation techniques were adopted: I, ship’s hull in the water; III ship’s hull above the water and no surface interactions allowed.
Table 3
Comparison of the Gordon and the pho-tran Backward MC Simulationsa
MC Simulation
ϕ0 =
∊Eu (0- m)
∊Lu (0- m)
0°
45°
90°
0°
45°
90°
I Gordon
7.74
8.13
8.95
3.33
4.30
8.96
I pho-tran
8.40
8.78
9.92
3.61
4.57
9.97
III Gordon
7.81
8.22
9.01
3.17
3.98
8.83
III pho-tran
7.90
8.17
9.16
3.22
3.95
8.90
Included are percentage relative error values for ship-shadowed Eu and Lu at 0--m depth and at variable solar azimuths. The same simulation techniques as in Table
2 were adopted.
Table 4
Standard Sun-Atmosphere-Ocean-AAOT-Detector Reference Systema
The values were allowed to vary within the simplified atmosphere-ocean system and were used to compute indexed AAOT subsurface tower-shading correction factors ηℜ̃(λ)[i,j,k,l,m]
included in the look-up tables.
The →∞ represents the overcast situation and is taken as the limiting parameter value for high (≫1.00) τaer values.
Their spectrally fixed values were allowed to covary with λ within the simplified atmosphere-ocean system and were used to compute AAOT tower-shading correction factors included in the look-up tables.
Table 9
Average and Standard Deviation of Subsurface Tower-Shadowing Percentage Relative Errorsa
Parameter
Unit
412 nm
443 nm
490 nm
510 nm
555 nm
665 nm
∊Ẽd
—
4.37 [1.42]
3.49 [1.19]
2.43 [0.86]
2.06 [0.73]
1.40 [0.50]
0.98 [0.22]
∊Ẽu
—
4.58 [1.05]
4.32 [1.03]
4.14 [1.03]
3.76 [1.07]
3.20 [1.13]
0.71 [0.34]
∊L̃u
—
4.28 [1.31]
3.67 [1.04]
3.19 [1.04]
2.87 [0.87]
2.35 [0.77]
0.55 [0.30]
ahyd
m-1
0.217 [0.084]
0.100 [0.066]
0.098 [0.036]
0.067 [0.026]
0.037 [0.014]
0.032 [0.028]
w0
—
0.795 [0.069]
0.823 [0.066]
0.858 [0.055]
0.856 [0.060]
0.847 [0.068]
0.508 [0.162]
Ir
—
0.844 [0.369]
0.671 [0.307]
0.499 [0.254]
0.438 [0.247]
0.336 [0.186]
0.223 [0.148]
∊ℜ values for Ẽd, Ẽu, and L̃u AAOT data, and the average and standard deviation of the relevant parameters used to retrieve the subsurface ηℜ̃ values from the look-up tables. Measurements were taken with solar zenith θ0 varying between 22° and 77° (48° average), and compass solar azimuth ϕ0′ varying between 87° and 245° (163° average). The numbers within brackets represent the standard deviation.
Tables (9)
Table 1
Comparison of the Gordon and the pho-tran MC Estimatesa
Radiometric Quantity (at 0-)
g = 0
g = 0.750
g = 0.875
g = 0.950
KA
KA′
Gordon
pho-tran
Gordon
pho-tran
Gordon
pho-tran
Gordon
pho-tran
Gordon
pho-tran
Gordon
pho-tran
Lu
0.08122
0.08119
0.02019
0.02021
0.00825
0.00855
0.00256
0.00251
0.01113
0.01069
0.01045
0.01022
Lu:σ
0.00132
0.00014
0.00100
0.00015
0.00115
0.00018
0.00197
0.00022
0.00348
0.00060
0.00061
0.00014
Lu:SRE
0.01630
0.00180
0.04950
0.00710
0.13940
0.02160
0.76950
0.08840
0.31260
0.05570
0.05830
0.01360
(Lu - L̃u)
0.01016
0.01013
0.00249
0.00257
0.00096
0.00103
0.00029
0.00030
0.00157
0.00102
0.00102
0.00098
(Lu - L̃u):σ
0.00049
0.00006
0.00044
0.00006
0.00059
0.00007
0.00068
0.00008
0.00187
0.00010
0.00026
0.00006
(Lu - L̃u):SRE
0.04820
0.00600
0.17670
0.00240
0.61450
0.07120
2.34480
0.26670
1.19100
0.09800
0.25490
0.02330
(Lu - L̃u)/Lu in %
12.50
12.48
12.30
12.72
11.60
12.10
11.30
12.03
14.10
9.53
9.70
9.59
(Eu - Ẽu)/Eu in %
21.30
21.27
14.50
13.74
12.20
12.55
14.20
14.94
14.70
14.90
14.20
14.14
The estimates are of absolute values (means, standard deviations σ, and SRE) of unperturbed and ship-shadow perturbed radiometric quantities. Results are given as a function of the in-water Henyey-Greenstein phase function (variable asymmetry parameters g) and for two separate phase functions (KA and KA′).
Table 2
Comparison of the Gordon and the pho-tran Backward MC Simulationsa
MC Simulation
ϕ0 =
∊Ed (5 m)
∊Ed (40 m)
0°
45°
90°
0°
45°
90°
I Gordon
1.27
1.43
2.36
3.98
5.80
15.88
I pho-tran
1.33
1.52
2.47
3.83
5.76
15.69
III Gordon
0.43
0.60
1.57
4.01
5.71
15.71
III pho-tran
0.40
0.58
1.55
4.05
5.62
16.04
Included are percentage relative error values for ship-shadowed Ed at 5- and 40-m depth and variable solar azimuth. Two simulation techniques were adopted: I, ship’s hull in the water; III ship’s hull above the water and no surface interactions allowed.
Table 3
Comparison of the Gordon and the pho-tran Backward MC Simulationsa
MC Simulation
ϕ0 =
∊Eu (0- m)
∊Lu (0- m)
0°
45°
90°
0°
45°
90°
I Gordon
7.74
8.13
8.95
3.33
4.30
8.96
I pho-tran
8.40
8.78
9.92
3.61
4.57
9.97
III Gordon
7.81
8.22
9.01
3.17
3.98
8.83
III pho-tran
7.90
8.17
9.16
3.22
3.95
8.90
Included are percentage relative error values for ship-shadowed Eu and Lu at 0--m depth and at variable solar azimuths. The same simulation techniques as in Table
2 were adopted.
Table 4
Standard Sun-Atmosphere-Ocean-AAOT-Detector Reference Systema
The values were allowed to vary within the simplified atmosphere-ocean system and were used to compute indexed AAOT subsurface tower-shading correction factors ηℜ̃(λ)[i,j,k,l,m]
included in the look-up tables.
The →∞ represents the overcast situation and is taken as the limiting parameter value for high (≫1.00) τaer values.
Their spectrally fixed values were allowed to covary with λ within the simplified atmosphere-ocean system and were used to compute AAOT tower-shading correction factors included in the look-up tables.
Table 9
Average and Standard Deviation of Subsurface Tower-Shadowing Percentage Relative Errorsa
Parameter
Unit
412 nm
443 nm
490 nm
510 nm
555 nm
665 nm
∊Ẽd
—
4.37 [1.42]
3.49 [1.19]
2.43 [0.86]
2.06 [0.73]
1.40 [0.50]
0.98 [0.22]
∊Ẽu
—
4.58 [1.05]
4.32 [1.03]
4.14 [1.03]
3.76 [1.07]
3.20 [1.13]
0.71 [0.34]
∊L̃u
—
4.28 [1.31]
3.67 [1.04]
3.19 [1.04]
2.87 [0.87]
2.35 [0.77]
0.55 [0.30]
ahyd
m-1
0.217 [0.084]
0.100 [0.066]
0.098 [0.036]
0.067 [0.026]
0.037 [0.014]
0.032 [0.028]
w0
—
0.795 [0.069]
0.823 [0.066]
0.858 [0.055]
0.856 [0.060]
0.847 [0.068]
0.508 [0.162]
Ir
—
0.844 [0.369]
0.671 [0.307]
0.499 [0.254]
0.438 [0.247]
0.336 [0.186]
0.223 [0.148]
∊ℜ values for Ẽd, Ẽu, and L̃u AAOT data, and the average and standard deviation of the relevant parameters used to retrieve the subsurface ηℜ̃ values from the look-up tables. Measurements were taken with solar zenith θ0 varying between 22° and 77° (48° average), and compass solar azimuth ϕ0′ varying between 87° and 245° (163° average). The numbers within brackets represent the standard deviation.