SIMULATION OF ALBEDO AT A LANDSCAPE SCALE WITH THE D.A.R.T. MODEL

SIMULATION OF ALBEDO AT A LANDSCAPE SCALE WITH THE D.A.R.T. MODEL

AN EFFICIENT TOOL FOR EVALUATING COARSE SCALE SATELLITE PRODUCTS?

Sylvie DUTHOIT*, Valérie DEMAREZ*Jean-Philippe GASTELLU-ETCHEGORRY*, Emmanuel MARTIN*

Jean-Louis ROUJEAN**

* CESBIO, ** Météo-France

Sylvie DUTHOIT Workshop on Albedo Products, April 2005

• Field measurements over an agricultural area

• High resolution imagery (SPOT HRV data)

• DART tool simulations of remote sensing images

Measurements over a long time period with good spatial sampling

+ Simulations of remote sensing images with DART

Spatial and temporal dynamic of the BRDF and albedo

Methodology

• The DART model

• The study area

• Input DART parameters• Land cover map• LAI • Soils optical properties• Leaves optical properties

• Landscape simulations• Results• Conclusions

• Main prospects


Outline

The D.A.R.T modelDiscrete Anisotropic Radiative Transfer

(Gastellu et al., 1996; 2005)

Principe

• Ray tracing method

• Discrete ordinate technique

(space and directions)

• Matrix of parallelepipedic cells

• With/without atmosphere, topography

Products

• Reflectance and temperature products (directional images)

• Radiative budget products (incident, intercepted, absorbed, scattered radiation for each scene level)

• Atmospheric products


THE DART model

Opaque cellsSoil, water surfaces, urban elements, trunks

surface scattering

Plane opaque surfaces: parametric reflectance models


DART images, nadir view (green, red & NIR spectral bands)

Turbid cellsVegetation, air volume scattering

LADLAI

THE DART model

“DART - Landscape” release

Optical and structural parameters

Leaves 1: LAI1, LAD1, 1, 1



soil

Land cover map

Plots mosaïc

Each plot is homogeneous (randomly distributed leaves)


THE DART model

An agricultural study area: Lamasquère


The study area

6x6 km and 3x3 km

•For the study area, DART need:

• Land cover map of the area

• Mean PAI of each stand

• Informations about canopy structure

• Leaves optical properties

• Soils optical properties


The study area

Land cover map


SPOT 2004 images

Land cover map(Gouaux, Ducrot, CESBIO)

DART input parameter

Input DART parameters

Wheat

Corn

Soya

Sunflower

Destructive PAI measurements



0

2

4

6

8

80 130 180 230Julian day

PAI

Corn Sunflower Soybean Wheat grassland

Seasonal evolution of the PAI

Grassland = SMOSRex measurements (KerrY., CESBIO)



y = 0,0017e8,9153x

R2 = 0,8774

0

1

2

3

4

5

6

7

8

0,5 0,6 0,7 0,8 0,9 1

MSAVI

PA

I

Spatialisation of the LAI

Vegetation index/LAI relation

SPOT images- corrected from atmospherical effects (SMAC model).- inter-calibrated(G. Dedieu, CESBIO)


7

0

PAI


2 major types of soils on the area (INRA database)

Soils optical properties

« Terreforts »« Boulbènes»


Measurements on plowed, sowed, dry and wet bare soils


Soils BRDF measurements

Spectroradiometer ASD

Directional measurements in several view angles


Hapke model to simulate the bi-directionnal reflectance

> parameters in the DART model



Leaves optical properties

0

10

20

30

40

50

60

70

0,4 0,5 0,6 0,7 0,8 0,9 1

Wave lenght

Re

fle

cta

nc

e (

%)

grass decidious species

ASTER database


April 21, May 16 and July 17

• Illuminations configurations for the time (date and hour) of the images

• 3 spectral bands (0.5-0.59, 0.61-0.68 and 0.79-0.89 m)

• diffuse radiation 20%

LANDSCAPE SIMULATIONSComparisons with SPOT reflectances


0

0,1

0,2

0,3

0,4

110 130 150 170 190 210Julian dayGreen Red NIR

SPOT reflectance

Landscape simulations

FIRST RESULTS


NIR (0,79-0,89 m)

0,26

0,28

0,3

0,32

0,34

0,36

0,38

100 120 140 160 180 200 2200

5

10

15

20

25

30

35

1 8

15

22

29

36

43

50

57

64

71

78

85

92

99

10

6

11

3

12

0

da

ily p

rec

ipit

ati

on

s (

mm

)

APRIL MAY JUNE JULY

Comparisons SPOT/DART valuesDry soils


Green( 0,5-0,59 m)

0,03

0,06

0,09

0,12

100 120 140 160 180 200 220

Red (0,61-0,68 m)

0,06

0,09

0,12

0,15

100 120 140 160 180 200 220

Julian day

Ref

lect

ance


Comparisons SPOT/DART values Dry soils and wet soils

Dry soils

Wet soils


Green (0.5-0.59)

0,04

0,06

0,08

0,1

0,12

100 120 140 160 180 200 220

Julian day

Re

fle

cta

nc

e

NIR (0.79-0.89)

0,2

0,24

0,28

0,32

0,36

0,4

100 120 140 160 180 200 220

Red (0.61-0.68)

0,04

0,06

0,08

0,1

0,12

0,14

100 120 140 160 180 200 220

CONCLUSIONS


BUT the difference between simulations and SPOT reflectance values are not negligible:

1- soils optical properties for wet soils (rain, irrigation in july)measurements in progress on the area

2- leaves optical properties : different for each crop species AND seasonnal evolution PROSPECT simulations

Results are encouragingThe seasonnal evolution is well reproduced, in general

PROSPECTS…

• Stand scale validation - Directionnal reflectance measurements over crops canopies

(ASD instrument) - Albedo measurements (K&Z net radiometer) over 2 crops stands

• Landscape scale simulations - Comparison with SPOT 3x3 km and 6x6 km. - Comparison with MODIS products: Black-sky albedo (16 days), Nadir- adjusted reflectance.


• DART: professional release (CNES, 2005), freely available for scientists

- Follow up of the work with new simulations for 2005

- Comparison with SEVIRI products - Evaluation in the framework of the SAF-Land project.


THANK YOU !


Green

0

0,03

0,06

0,09

0,12

100 120 140 160 180 200 220Red wave band

0

0,03

0,06

0,09

0,12

0,15

100 120 140 160 180 200 220

Julian day

NIR

0,28

0,3

0,32

0,34

0,36

0,38

100 120 140 160 180 200 220

MODIS Black-sky albedo

Documents

SIMULATION OF ALBEDO AT A LANDSCAPE SCALE WITH THE D.A.R.T. MODEL