Experimental validation of the MODTRAN 5.3 sea surface radiance model

Denis Dion, Vincent Ross* andDaniel St-Germain

33rd Review of Atmospheric Transmission Models Meeting

June 2011

* With AEREX Avionic Inc

1 of 13

Contents

• Introduction

• The sea surface BRDF model

• Complementary models

• The MIRAMER campaign

• Experimental and modeling uncertainties

• Experimental Results

• Conclusions

2 of 13

Introduction

Complexity of sea surface radiance– Background

• Thermal emissions• Direct solar reflections• Indirect solar reflections

– Foreground• Contributes to atmospheric radiance

Radiative coupling

3 of 13

Introduction

A sea surface BRDF for greater radiance accuracy

4 of 13

Situation with MODTRAN

• No full BRDF coupling up to MODTRAN 4• No sea surface BRDF up to MODTRAN 5 v2• Basic aerosol models, limited user inputs• Shortcomings in refracted path at horizon ranges in

MODTRAN 4

• MODTRAN 5 v3– Fully coupled analytical sea surface BRDF– SAP (Spectral Aerosol Profile) input– Refracted path input

5 of 13

The sea surface BRDF

( ) ( )[ ]3

, ( ) ( , ) ( , ),

4 ( ) 1 ( ) ( ) cos coss r r r

s rn n r r s r s

r p W Hf

z v vζπ

θ θ=

⋅ + Λ + Λψ ψ ζ ζ Ψ ζ Ψ

ψ ψU U

Ross, V., D. Dion, and G. Potvin, “Detailed analytical approach to the Gaussian surface bidirectional reflectance distribution function specular component applied to the sea surface,” J. Opt. Soc. Am. A Opt. Image Sci., 22, 2442– 2453 (2005).

Fresnel reflectance

Probability of specular reflection

Wave facet angle weighing

Wave facet hiding

Bistatic wave shadowing

6 of 13

Coupling to MODTRAN

• BRDF is coded in FORTRAN in MODTRAN 5

• Fourier moments are computed– Input in DISORT

• DISORT multiple scattering uses BRDF as a lower boundary condition

7 of 13

Other modeling considerations

• Marine aerosols are computed using MEDEX– Well suited for the Mediterranean– Input in MODTRAN using the SAP input

• MBL (marine boundary layer) thermodynamic profiles are computed using the DRDC modules– Monin-Obukhov similarity theory

• Refracted optical paths are used as input

• Sea surface statistical properties: Elfouhaily et al.– Fetch, atmospheric stability

8 of 13

The MIRAMER campaign (May 2008)• Cedip Jade (Flir ATS) cameras on board

the Atalante ship– 3.4 – 5.5 mm (3.93 – 4.14 mm filter for glint)– 8.19 – 8.96 mm

• Environmental characterization– Radiosondes (2-3 day)– Local meteorological measurements

• Air and sea temperature• Wind speed/direction• Relative humidity

– Visibility meter (aerosols)– Aeronet station nearby (Toulon)– Solar pyranometer (solar irradiance)

9 of 13

• Experimental– Image calibration and limited dynamic range

• Can reach 20% but probably lower (4-5%)– Horizontal variations (temperature, etc.) not measured

• Temperature +/- 1o

– Wind gusts– Bulk vs. skin temperature

• +/- 1o

• Modeling– Slope statistics values

• 80% between models– Aerosol modeling– Multiple reflections– Cirrus clouds

Experimental and modeling uncertainties

10 of 13

Results• Non glint

– Midwave

– Longwave

-7 -6 -5 -4 -3 -2 -1 02.15

2.2

2.25

2.3

Rad

ianc

e (W

/m2 /s

ter)

ATAL 95 (1159) BII

-7 -6 -5 -4 -3 -2 -1 0-2

-1

0

1

Diff

eren

ce (%

)

Elevation (degrees)

12.5

13

13.5

14

14.5

15

App

aren

t tem

pera

ture

(o C)

Measurment

Simulation (-0.227 W/m2/ster)

σerr = 0.56%

-7 -6 -5 -4 -3 -2 -1 02.3

2.35

2.4

2.45

Rad

ianc

e (W

/m2 /s

ter)

ATAL 109 (1267) BII

-7 -6 -5 -4 -3 -2 -1 0-2

0

2

Diff

eren

ce (%

)

Elevation (degrees)

14.5

15

15.5

16

16.5

17

App

aren

t tem

pera

ture

(o C)

Measurment

Simulation (-0.099 W/m2/ster)

σerr = 0.76%

-7 -6 -5 -4 -3 -2 -1 04

4.1

4.2

4.3

4.4

4.5

4.6

4.7

4.8

Rad

ianc

e (W

/m2 /s

ter)

ATAL 89 (1116) BIII

-7 -6 -5 -4 -3 -2 -1 0-2

0

2

4

Diff

eren

ce (%

)

Elevation (degrees)

-10

-5

0

5

App

aren

t tem

pera

ture

(o C)

Measurment

Simulation (-1.39 W/m2/ster)

σerr = 1.76%

-7 -6 -5 -4 -3 -2 -1 05.5

5.6

5.7

5.8

5.9

6

6.1

Rad

ianc

e (W

/m2 /s

ter)

ATAL 107 (1255) BIII

-7 -6 -5 -4 -3 -2 -1 0-2

0

2

Diff

eren

ce (%

)

Elevation (degrees)

4

6

8

10

12

14

16

18

20

App

aren

t tem

pera

ture

(o C)

Measurment

Simulation (-0.133 W/m2/ster)

σerr = 1.14%

-7 -6 -5 -4 -3 -2 -1 05.9

6

6.1

6.2

6.3

6.4

Rad

ianc

e (W

/m2 /s

ter)

ATAL 109 (1267) BIII

-7 -6 -5 -4 -3 -2 -1 0-3-2-101

Diff

eren

ce (%

)Elevation (degrees)

10

12

14

16

18

20

App

aren

t tem

pera

ture

(o C)

Measurment

Simulation (+0.265 W/m2/ster)

σerr = 1.42%

-7 -6 -5 -4 -3 -2 -1 0

2.32

2.34

2.36

2.38

2.4

2.42

2.44

Rad

ianc

e (W

/m2 /s

ter)

ATAL 107 (1255) BII

-7 -6 -5 -4 -3 -2 -1 0-1

-0.5

0

0.5

Diff

eren

ce (%

)

Elevation (degrees)

14.5

15

15.5

16

16.5

17

App

aren

t tem

pera

ture

(o C)

Measurment

Simulation (-0.113 W/m2/ster)

σerr = 0.19%

11 of 13

Results

• Glint

Relative azimuth (degrees)E

leva

tion

(deg

rees

)

ATAL 145 (1416) BII reference image

-4 -2 0 2 4 6

-7

-6

-5

-4

-3

-2

-1

0

Rad

ianc

e (W

/m2 /s

ter)

0

0.5

1

1.5

2

2.5

3

3.5

4A

B

B

A

-7 -6 -5 -4 -3 -2 -1 00

0.5

1

1.5

2ATAL 145 (1416) BII vertical

Rad

ianc

e (W

/m2

/ste

r)

-7 -6 -5 -4 -3 -2 -1 0-100

0

100

200

Elevation (degrees)

Diff

eren

ce (%

)

Measurment

Simulation (-0.05 W/m 2 /ster)

-4 -3 -2 -1 0 1 2 3 4 5 60

0.5

1

1.5ATAL 145 (1416) BII horizontal

Rad

ianc

e (W

/m2

/ste

r)

-4 -3 -2 -1 0 1 2 3 4 5 6

-20

0

20

Relative Azimuth (degrees)

Diff

eren

ce (%

)

Measurment

Simulation (-0.05 W/m 2/ster)

(Note: Cirrus cloud modeled using Aeronet AOD data)

12 of 13

Results

• Glint

Relative Azimuth (degrees)E

leva

tion

(deg

rees

)

ATAL 64 (902) BII reference image

-6 -4 -2 0 2

-8

-7

-6

-5

-4

-3

-2

-1

Rad

ianc

e (W

/m2 /s

ter)

5

10

15

20AA

B

B

-9 -8 -7 -6 -5 -4 -3 -2 -1 01

2

3

4

5

6

7ATAL 64 (902) BII vertical

Elevation (degrees)

Rad

ianc

e (W

/m2 /s

ter)

MeasurementSimulation (+0.2 W/m2/ster)Horizon correction

-8 -6 -4 -2 0 2 40

1

2

3

4

5

6ATAL 64 (902) BII horizontal

Relative Azimuth (degrees)

Rad

ianc

e (W

/m2 /s

ter)

MeasurementSimulation (+0.2 W/m2/ster)Horizon correction

Ross, V., Dion, D., "Sea surface slope statistics derived from Sun glint radiance measurements and their apparent dependence on sensor elevation, J. Geophys. Res., 112, C09015, (2007)

13 of 13

Conclusions

• A radiatively coupled sea surface BRDF is important in maritime environment radiative transfer

• MODTRAN 5 v3 will introduce a coupled sea surface BRDF

• Analysis of MIRAMER radiometric images supports the validation of the MODTRAN sea radiance implementation– Simulations and measurements agree well within experimental

uncertainties