Optical variability and optical anomalies in Mediterranean waters

Optical variability and optical « anomalies » in Mediterranean waters André Morel, David Antoine and Hervé Claustre Laboratoire d’Océanographie de Villefranche (CNRS and Univ. P. M. Curie)

Is desert dust making oligotrophic waters greener? (H. Claustre et al. , GRL, 2002) Field data (PROSOPE Cruise), versus Remote Sensing OC algorithms (OC 4 V 4 and OC 4 Me) Blue-to-green ratio: Anomalously low In Med-Sea PROSOPE cruise

Depth of the euphotic layer, as expected from Surface [Chl] Hyperspectral Model (MM 01) Sun at 0° or 75° -Uniform profiles (solid lines) or -profiles including a DCM (dashed lines) Recent data Med-Sea data (anomalously low Zeu) (Morel & Gentili, 2004)

Mediterranean Sea : A semi-enclosed basin, Rather well ventilated, however (water residence time rather short ~ 70 y) Arid climate, and reduced run off limited continental shelf A priori, a Case 1 water domain, with varying trophic status: Predominantly oligotrophic, sporadically and periodically mesotrophic (Blooms) SO, why « anomalies » ?

Anomalies, or nuances, detectable against « a standard » for Case 1 waters Therefore, Defining a « standard » is a prerequisite, Possibility: Consider the average empirical relationships established between some optical properties (IOP and AOP) and (Chl), used as an index of the bio-optical state.

« Standard » for Case 1 waters? Historical empirical relationships provide such average laws (+ SD), (generally non-linear laws of (Chl)) For instance: IOP ap(λ, [Chl]) = A(λ) [Chl] ^B(λ) Bricaud et al, bp(λ, [Chl]) = Bo(λ) [Chl] ^ β(λ) Gordon-Morel, Loisel-Morel cp(λ, [Chl]) = Co(λ) [Chl] ^ γ (λ) Voss AOP Kd(λ, [Chl]) = Kw (λ) + χ(λ) [Chl] ^e(λ) Morel-Maritorena. Rrs(λ 1) / Rrs(λ 2) = Pol ([Chl]) O’Reilly et al.

COMPATIBILITY BETWEEN THESE EMPi. RICAL LAWS ? has to be checked before ascertaining “Standards”, and being able to identify “nuances” 1) Closure (IOP) ? ? ap (λ, [Chl]) (Bricaud et al, ) + bp(λ, [Chl]) = cp(λ, [Chl]) (Gordon-Loisel-Morel) (Voss) 2) Coherency (IOP/AOP) ? Possible Inversion Kd (λ, [Chl]) → atot (λ, [Chl]), ? atot (λ, [Chl]) > ap (λ, [Chl]) (with atot = ap + ay + aw) if yes then : ay (λ, [Chl]) = ?

From Bricaud et al. , 1998: ap(λ) as f(Chl) Average law and confidence interval (example for 440 nm) Data Med-Sea

Loisel-Morel, L&O, 1998 slope 0. 766

Closure: c = a + b - Medit. Sea (above average relationship

(Med-Sea)

INVERSION (AOP -> IOP) Kd = 1. 0395 (μd) -1 (a + bb) R = f’ [bb/ (a + bb)] a(λ) = 0. 962 Kd (λ) µd(θs, λ, Chl ) [ 1 - R(λ ) / f’(θs, λ, Chl )] bb(λ) = 0. 962 Kd (λ) μd (θs, λ, Chl ) [ R(λ ) / f’(θs, λ, Chl )] Look up Tables for μd and f’ (Morel-Gentili, JGR 2004)

Compatibility between ap and Kd Inversion (Kd → atot) through: atot(λ, [Chl]) = 0. 962 Kd (λ, [Chl], θs) μd(λ, [Chl], θs) { 1 - R(λ, [Chl], θs)/f’(λ, [Chl], θs)} R(λ, [Chl], θs), f’(λ, [Chl], θs), and μd(λ, [Chl], θs) in LUTs (from RTE computations) Then: atot = ap + ay + aw Is atot coherent with ap ? (i. e. , atot > ap )

Examples of Kd(λ) ↔ (Chl) empirical relationships NOMAD Data Black curves: Morel-Maritorena Statistical relationships

Example of Inversion Kd(412) atot (412) Then Decomposition atot = (ap + aw) + ay

ay(412) from the previous figure + ay Data (Pacific) also obtained by inversion using Kd and R Conclusion; Pacific waters close to standard Case 1 water regarding ay

ay(412) from the previous figure + Data (Med. Sea) also obtained by inversion Moroccan upwelling Conclusion: Med-Sea waters above standard regarding ay

CONCLUSION • Empirical relationships are compatible, can be used to define « standard » , Chl-dependent, Case 1 waters • With respect to this standard, Med-Sea waters exhibit notable, and identified nuances (thence anomalous reflectances already detected) ( Note: particulate absorption, as usual ) 1) excess of Yellow Substance (seasonal? regional? Blooms in the Northern part? bacterial activity? ) 2) slight excess of particle scattering (Saharan dust? Coccoliths? ) 3) Bio-geo-chemical reasons are not yet elucidated