Quantification of mineral particles from remote sensing.Using of spectroradiometric measurements and WASI

simulations

Results obtained by V. Lafon, C. Giry, N. Bonneton, D. Doxaran, D. Bru C. Petus, M. Schmeltz and J-M Froidefond

1) Spectroradiometric measurements

2) Examples of SPMC quantification in the Bay of Biscay,the French Guiana and the Congo coastal waters

3) Inversion of spectra from the WASI code (P. Gege)

Spectroradiometric measurements

Radiance measurements with Trios sensors located

above + 2cm and below -2cm (Lu). ± 1cm

Irradiance measurement with a Triossensor (Ed). 350nm – 950nm

Calibrations in air before or after thesurvey at “Trios”

Froidefond and Ouillon, 2005

),0(),0(*98.0)( durs ELR

Remote sensing reflectance

Example of reflectance spectra (Rrs (l)

Water radiance (Lu)

Irradiance (Ed)

Remote sensing reflectance Rrs(l)

Rrs(sr-1) = Lu/Ed

Shadow effects relatively low

Water radiance just below the sea surface

Turbid waters (> 30mg/L)

Very high shaddow effects

Radiance measured above the sea surface

Backscattered light attenuated by theshadow of the sensor

Clear waters

Water radiance just above the sea surface

Identification and quantification of suspended particles

Gironde area

Arcachon areaGironde area : quantification ofmineral particles (PNEC)

Adour area :OOSEA programRemote sensing monitoring inAquitaine – Euskadi(AZTI, LASAGEC, EPOC)

Adour areaArcachon area(CNES, Kalideos-Littoral)Intertidal and subtidal mapping

French Guiana (IRD, PNEC

Optic-Congo (SHOM)

Bay of Biscay (SHOM, INSU, Region)Optic-Med (SHOM)

Bissecotte (IRD)

About 400 spectra recorded during differentoceanographic surveys in case 2 waters. At eachstation, hydrologic data (SPM, CDOM, Chlorophyll-aor fluorescence, CDOM

Reflectance Rrs OPTIC-CONGO (805)

0

0.002

0.004

0.006

0.008

0.01

0.012

0.014

0.016

0.018

0.02

350 400 450 500 550 600 650 700 750 800 850 900 950

nm

Rrs

(sr-

1)

Reflectance B66

0

0.002

0.004

0.006

0.008

0.01

0.012

0.014

0.016

0.018

0.02

350 400 450 500 550 600 650 700 750 800 850 900 950

nm

Rrs

(sr-

1)

Blue waters. Very low concentrations in SPM Beige waters: mineral suspended particles +…

Dark brown waters: CDOM +…Green waters: phytoplancton +…

Pic de fluorescence

Reflectance B26E

0

0.002

0.004

0.006

0.008

0.01

0.012

0.014

0.016

0.018

0.02

350 400 450 500 550 600 650 700 750 800 850 900 950

nm

Rrs

(sr-

1)

Différent types of reflectance spectra

(Jourdin et al., 2006)

1) Gironde plume.Relationship betweensuspended particulatematter concentrations andMODIS reflectances inbands 1 and 2 (250mresolution)

Between 0 and 50 mg/L

SPMC(mg/L) = 1142.4*Rrs(B1)

SPMC(mg/L) = 1007.3*Rrs(B1)

2) Adour plume

(Petus et al., 2010)

3) French Guiana. ELISA-7 survey 2004, 4–7June

Modis/Aqua, 2004, June 5

SPMC(mg/L) = 1377.7*Rrs(B1)

(Froidefond et al., accepted)

Amazonturbid plume

SPM concentrations in the Bay of Arcachon from SPOT data

Arcachon

Gabon and Congo coastal waters (Optic-Congo survey, SHOM, Schmeltz et al., 2009)

SPMC = 1260.7*Rrs(B1)

No relationship

(Bru, 2010)

Summary of the measurements

SPMC between 1 and 30 mg/L

B1 (Modis Band 1. 620nm – 670nm)

Gironde (turbid plume): SPMC(mg/L) = 1142.4*Rrs(B1)Adour R. (Petus et al. 2010): SPMC(mg/L) = 1007.3*Rrs(B1)French Guiana (turbid waters): SPMC(mg/L) = 1377.6*Rrs(B1)Arcachon bay (Bru, 2010): SPMC(mg/L) = 1260.7*Rrs(B1)Congo coast : No relationshipIrish Sea (Binding et al., 2005): SPMC(mg/L) = 516.3*Rrs(665nm) + 1.13Mississipi plume (Miller et al): SPMC(mg/L) = 1140.3*Rrs(B1) – 1.9

Different empirical relationships, explained by the optical properties of thewater components:Various clay minerals (illite, kaolinite, chlorite, smectite…), quartz, micas…Various granulometric size and flocsConcentration and composition of organic particles and CDOM.

Comparison with a spectra simulation code (WASI)

P. Gege, 2004

Initial values

Fit parameters(output data)

WASI, Water color simulator (P. Gege, 2004)

Chlorophyllconcentration

SPM concentration

Exponent of yellowsubst. absorption

Yellow subst. (CDOM)concentration

IOP (a, bb, Kd)

Original spectrum

Fitted spectrum

Model and options:Rrs-(l) = frs*[bb(l) / (a(l) + bb(l))] and Rrs+(l) = xi*Rrs-(l)/(1-sigma-*R)+Rsurf

with xi=1/nw2, sigma- = 0.54, nw = 0.33

frs = p1*(1+p2*x+p3*x2+p4*x3*[1+p5/cos(sun_w)]*[1+p7/cos(view_w)] (Albert and Mobley, 2003)Ancillary parameters: bottom depth, zenith angle

Example with MERIS data recorded during the Batel-1 survey

MERIS June 5, 2007,10h40 – 10h42 TU

Station B5,June 5, 2007, 9h51 TUSPMC = 0.9mg/L,Chlor-a C. = 0.5mg/m3

Adour River

Rrs reconstruction MERIS L2 and inversion from WASI

Results with in situ SPMC and CDOM, but not with Chlorophyll a

(Schmeltz et al., 2010)

Inversion of MERIS –L2 spectrum (B5 station) from WASI

Constituents and Z90 comparison for station B5

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Chlor (µg/l) SPM(mg/l) CDOM(1/m) Z90max(m/10)at 530nm

in_situmeasurements

Concentrations fromL2 products

WASI derived values fromin-situspectrum

WASI derived values fromMERISL2spectrumreconstructed

BLACK: In-situ measurements

GREEN: Concentration from MERIS-L2 products

RED: WASI derived values from in-situ spectrum

BLUE: WASI derived values from MERIS reconstructed spectrum

Schmeltz et al., 2010

1) Improvement of the reflectance measurements with theradiance sensor juste below the water surface

2) Empirical algorithms are similar if the optical propertiesare similar, but… the organic matter can change theserelationships.

The WASI code (P. Gege) allows to test differenthypothesis and to inverse the reflectance spectra.

Conclusion

Thank you