Update on validation of MERIS r w ’s at the BOUSSOLE site

2nd MERIS/(A)ATSR User Workshop22nd to 26th September 2008 ESA/ESRIN Frascati (Rome) Italy

Update on validation of MERIS w’s at the BOUSSOLE site

David Antoine1, Constant Mazeran2, Christophe Lerebourg2, Ludovic Bourg2 , Marc Bouvet3, Fabrizio D'Ortenzio1, Samantha Lavender4,

Kathryn Barker4, Bernard Gentili1, Vincenzo Vellucci1

1: Laboratoire d’Océanographie de Villefranche, Villefranche sur mer, France2: ACRI-st, Sophia Antipolis, France3: ESA-ESTEC, Noordwijk, The Netherlands4: ARGANS Ltd, Plymouth, UK


A reminder about BOUSSOLE

- Scientific objective: IOPs et AOPs documentation and understanding (bio-optics research), short-time changes...

- Operational objective: vicarious radiometric calibration of ocean color satellite observations, and validation of the Level-2 geophysical products derived from these observations (e.g., chlorophyll, reflectances...).

Motivation, objectives: establishing a long-term time series of optical properties (IOPs and AOPs), with two parallel objectives:

Strategy: combination of 3 elements:

- A Deep-sea mooring, for continuous collection of data at the surface

- Monthly cruises for the buoy servicing and the collection of data complementary to the buoy data

- An AERONET coastal station, to provide the necessary aerosol parameters


The BOUSSOLE sitein the Ligurian Sea (northwestern Mediterranean)


Waters at BOUSSOLE are permanently Case 1 waters

Buoy data

Morel & Maritorena, 2001Upper limit for Case 1 waters

Irradiance reflectance at 560 nm as a function of Chl


The range of bio-optical properties at BOUSSOLE is representative of global Case 1 waters

Thick line : BOUSSOLEThin line and thin dotted line : SeaWiFS global or Med Sea onlyGray : NOMAD


The BOUSSOLE buoy, “transparent-to-swell” taut mooring


Buoy data are in agreement with more“classical data”, i.e., profiling radiometer’s data

Slope is 1r2 is 0.98

Bias is 8 10-4

reflectance from a free-falling radiometer (Satlantic SPMR)

refl

ecta

nce

fro

m t

he

bu

oy


Status of buoy deployments- 5 years of deployment, with a quasi permanent data collection

(September 2003 September 2008)

- Over these 5 years (1825 days), we had 1560 days with a buoy at sea. This is a 85% “deployment success”, i.e., nearly operational

- Over these 1560 days, 1300 were with data acquisition. This is a 83% “acquisition success”, which is also very satisfactory

- Bi-monthly servicing since July of 2007

- No data from January 15 to June 15 2008 (5 months)

- Next rotation is scheduled Jan 2009

- Evolution towards hyper-spectral measurements has started in June 2008 (350800 nm, res 2 nm). Nominal set of 7- multispectral data is however preserved for continuity


Matchup proceduresEssentially follows: Bailey S.W., and P.J. Werdell, 2006: A multi-sensor approach for the on-orbit validation of ocean color satellite data products. Remote Sens. Environ., 102, 12-23. - 5x5 pixel box

- Flags: glint, clouds, haze, s<70°, v<60°, clear sky

-Spatial homogeneity is checked

- MERIS: level-2 RR, last reprocessing 109 matchups

-SeaWiFS: level-2 Merged Local Area Coverage (MLAC; until Dec. 2004) or Global Area Coverage (GAC; 2005-2006) data from

reprocessing 5 (completed March 18, 2005) 193 matchups

- MODIS-A: level-2 GAC data from reprocessing 1 (completed in February 2005) 185 matchups


4-year time seriesof w

MERIS

SeaWiFS

MODIS-A

Field data (10am – 2pm, clear sky)


Full matchup set, using data from Sept 2003 to Dec 2007


N Number of points

r Mean ratio, defined as

Ni

1i i

i

y

x

N

1

RPD Average relative (signed) percent difference (%), defined as

Ni

1i i

ii

y

)yx(

N

1100

|RPD| Average relative (unsigned) percent difference (%), defined as

Ni

1i i

ii

y

yx

N

1100

r2, m and y Coefficient of determination, slope and intercept of the linear fit

RMSE Root mean square error (or root means square of differences), defined as

Ni

1i

2ii )yx(

N

1

Note the Chl data are log-transformed before computing r2, m, y and RMSE.

Statistics


Matchup statistics, using data from Sept 2003 to Dec 2007

MERIS

N r RPD |RPD| r2 m y rmse412 39 1,55 55,2 59,4 0,16 0,77 0,00985 0,01032443 98 1,28 27,8 32,5 0,36 0,74 0,00681 0,00529490 104 1,14 14,0 18,7 0,45 0,78 0,00433 0,00289510 103 1,19 19,2 23,3 0,14 0,51 0,00615 0,00260560 98 1,18 18,3 23,5 0,08 0,41 0,00378 0,00144665 65 2,32 132,2 133,1 0,00 -0,08 0,00114 0,00069681 72 1,94 94,5 95,9 0,20 0,69 0,00069 0,00063

All bands 579 0,79 1,12 0,00078

SeaWiFS

N r RPD |RPD| r2 m y rmse412 61 0,94 -5,7 20,6 0,40 0,70 0,00377 0,00449443 191 1,02 1,9 16,2 0,47 0,71 0,00458 0,00310490 193 0,96 -3,5 13,2 0,34 0,56 0,00565 0,00244510 193 0,93 -6,9 14,0 0,11 0,35 0,00611 0,00194555 108 0,91 -9,0 16,9 0,09 0,41 0,00271 0,00122670 179 0,95 -5,2 53,8 0,00 0,11 0,00040 0,00034

All bands 925 0 0 0 0,87 0,92 0,00030 0,00000

MODIS-A

N r RPD |RPD| r2 m y rmse412 59 0,93 -7,1 24,1 0,53 0,78 0,00218 0,00437443 180 0,98 -2,0 16,7 0,56 0,77 0,00299 0,00305488 185 0,93 -7,2 14,1 0,45 0,62 0,00425 0,00255551 182 0,91 -9,3 15,9 0,22 0,59 0,00165 0,00105667 183 0,85 -15,3 44,1 0,03 0,22 0,00029 0,00028

All bands 789 0 0 0 0,90 0,91 0,00018 0


Conclusions

- The requirements in terms of accuracy of the atmospheric correction are only met at 443 and 490 nm by the SeaWiFS and MODIS-A products.

- The MERIS products don’t meet the requirements.

- The water-leaving radiance reflectances provided by the three sensors at 412 nm are severely affected by atmospheric correction errors.

- The uncertainty is significantly reduced for the “blue-to-green” reflectance ratio.

- These results and the matchup statistics are in agreement with the results obtained by two other similar efforts carried out at a coastal site [Zibordi et al., 2006] and globally [Bailey and Werdell, 2006].

Vicarious calibration should be included in the MERIS L2 processing, as it is done for other OC sensors presently in orbit

See next presentations


Further information

Antoine, D. M. Chami, H. Claustre, F. D’Ortenzio, A. Morel, G. Bécu, B. Gentili, F. Louis, J. Ras, E. Roussier, A.J. Scott, D. Tailliez, S. B. Hooker, P. Guevel, J.-F. Desté, C. Dempsey and D. Adams. 2006, BOUSSOLE : a joint CNRS-INSU, ESA, CNES and NASA Ocean Color Calibration And Validation Activity. NASA Technical memorandum N° 2006–214147, 61 pp.

Antoine, D., P. Guevel, J.-F. Desté, G. Bécu, F. Louis, A.J. Scott and P. Bardey, 2008, The « BOUSSOLE » buoy – A new transparent-to-swell taut mooring dedicated to marine optics : design, tests and performance at sea, Journal of Atmospheric and Oceanic Technology, 25, 968-989.

Antoine, D., F. D’Ortenzio, S.B. Hooker, G. Bécu, B. Gentili, D. Tailliez and A. Scott, 2008, Assessment of uncertainty in the ocean reflectance determined by three satellite ocean color sensors (MERIS, SeaWiFS, MODIS) at an offshore site in the Mediterranean Sea (BOUSSOLE project), Journal of Geophysical Research, 113, C07013, doi:10.1029/2007JC004472.

Bailey S., S.B. Hooker, D. Antoine, J.P. Werdell and B. Franz, Vicarious calibration of ocean color satellite remote sensors: An analysis of sources and assumptions, Applied Optics, 47(12), 2035-2045.


Thank you for your attention

www.obs-vlfr.fr/Boussole

Update on validation of MERIS r w ’s at the BOUSSOLE site

Documents