Atmospheric correction over coastal and in land waters

Atmospheric correction over coastal and in land waters: Doing better with MERIS? Richard Santer Université du Littoral Côte d’Opale – 32, avenue Foch – 62930 Wimereux – FRANCE ( santer@mren 2. univ-littoral. fr ) Frascati, September 2005 1

Outlines • Improving flags using the oxygen band: – Cirrus clouds – Sunglint – Stratospheric aerosols • Improving analysis when flagged • Improving formulation of the signal for coastal and inland waters – Adjacency effects – Alternative AC 2

Cirrus clouds 3

Cirrus clouds over land Classification MERIS level 2: yellow (standard) and black ( using pressure product) 4

CIRRUS MERIS image of the ratio between the TOA radiance at 761 nm and 753 nm North Sea on the 31/03/2004 5

Sunglint 6

GLITTER MERIS image of the ratio between the TOA radiance at 761 nm and 753 nm Lampedusa on the 16/07/2003 7

Dust over England 8

the Montserrat Volcano eruption 9

Objective 1 • Use 761/765 for identification of specific events • Illustrate the influence on products • Correlate discrepancies on 761/765 and bias on products 10

Outlines • Improving flags using the oxygen band: – Cirrus clouds – Sunglint – Stratospheric aerosols • Improving analysis when flagged • Improving formulation of the signal for coastal and inland waters – Adjacency effects – Alternative AC 11

Modelise the TOA signal over ocean in the O 2 band 5 S formulation in the reference band at 753 nm Introduction of specific O 2 transmittances in the O 2 band Modelisation of these terms in the primary scattering approximation 12

Outlines • Improving flags using the oxygen band: – Cirrus clouds – Sunglint – Stratospheric aerosols • Improving analysis when flagged • Improving formulation of the signal for coastal and inland waters – Adjacency effects – Alternative AC 13

Objective 2 • Use 761/765 product • Illustrate the influence on level 2 products • Correlate discrepancies on 761/765 product and bias on level 2 products • Correct level 2 product 14

Formulation of the signal L toa = ( L atm + T * Lg + t * (Lw + Lwc ) )Tg 15

The atmospheric path radiance ( a + ar) versus as at 865 nm (full line) and 443 nm (dashed lines) for 3 albedos : 0. 6, 0. 8 et 1. 0 left: continental, right: maritime. 16 o=60°, =90 (Gordon, 1997).

Atmospheric transmittance 17

Validity of the transmittance computation Rayleigh transmittances following Gordon and 6 S 18

Objective 3 • Evaluate bias on the level 2 product due to inaccurate formulation of the signal • Remediate 19

AC over inland water How are the ocean AC for inland water? What alternative do we have? IST-2000 -28187 Satellite-based Information System on Coastal Areas and Lakes 20

AOT over water and over land Constance lake 21

AOT over water and over land Balaton lake 22

AOT over water and over land Constance lake Balaton lake 23

How compare the GW and ML for the water products Balaton IST-2000 -28187 Satellite-based Information System on Coastal Areas and Lakes Constance 24

Water or land? Good spatial homogeneity over land. Over water, higher values correspond to a non null reflectance in the NIR. Higher dispersion over water reflects the adjacency effect. IST-2000 -28187 Satellite-based Information System on Coastal Areas and Lakes Paris, 13 December 2004 25

Objective 4 • Aerosol remote sense over land can be alternatively used for AC over inland water • Implement • Validate 26

Adjacency effect • Light reflected by neighbourg pixels and scattered by the atmosphere towards the sensor – increases with contrast – increases with scattering Land mask for the Fresnel reflection IST-2000 -28187 Satellite-based Information System on Coastal Areas and Lakes 27

MERIS, North of Germany, August, 13, 2003 28

AOT_865 29

Objective 5 • quantify the adjacency effects – BRADEX 30

Bradex 31