This article was downloaded by: [University of Tasmania]On: 15 November 2014, At: 15:12Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number:1072954 Registered office: Mortimer House, 37-41 Mortimer Street,London W1T 3JH, UK

International Journal ofRemote SensingPublication details, including instructions forauthors and subscription information:http://www.tandfonline.com/loi/tres20

MERIS potential for oceancolour studies in the openoceanAnnick Bricaud , Andre Morel & Vittorio BaralePublished online: 25 Nov 2010.

To cite this article: Annick Bricaud , Andre Morel & Vittorio Barale (1999) MERISpotential for ocean colour studies in the open ocean, International Journal ofRemote Sensing, 20:9, 1757-1769, DOI: 10.1080/014311699212461

To link to this article: http://dx.doi.org/10.1080/014311699212461

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of allthe information (the “Content”) contained in the publications on ourplatform. However, Taylor & Francis, our agents, and our licensorsmake no representations or warranties whatsoever as to the accuracy,completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views ofthe authors, and are not the views of or endorsed by Taylor & Francis.The accuracy of the Content should not be relied upon and should beindependently verified with primary sources of information. Taylor andFrancis shall not be liable for any losses, actions, claims, proceedings,demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, inrelation to or arising out of the use of the Content.

This article may be used for research, teaching, and private studypurposes. Any substantial or systematic reproduction, redistribution,

reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of accessand use can be found at http://www.tandfonline.com/page/terms-and-conditions

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

int. j. remote sensing, 1999, vol. 20, no. 9, 1757± 1769

MERIS potential for ocean colour studies in the open ocean

ANNICK BRICAUD² and ANDRE MORELLaboratoire de Physique et Chimie Marines, Universite Pierre et Marie Curieand CNRS, BP 8, 06230 Villefranche-sur-Mer, France

VITTORIO BARALESpace Applications Institute, JRC EC, 21020 Ispra (VA), Italy

(Received 22 April 1996; in ® nal form 2 April 1997 )

Abstract. The interest of space observations of ocean colour for determiningvariations in phytoplankton distribution and for deriving primary production(via models) has been largely demonstrated by the Coastal Zone Color Scanner(CZCS) which operated from 1978 to 1986. The capabilities of this pioneer sensor,however, were limited both in spectral resolution and radiometric accuracy. Thenext generation of ocean colour sensors will bene® t from major improvements.The Medium Resolution Imaging Spectrometer (MERIS), planned by theEuropean Space Agency (ESA) for the Envisat platform, has been designed tomeasure radiances in 15 visible and infrared channels. Three infrared channelswill allow aerosol characterization, and therefore accurate atmospheric correc-tions, to be performed for each pixel. For the retrieval of marine parameters, ninechannels between 410 and 705nm will be available (as opposed to only four withthe CZCS). In coastal waters this should, in principle, allow a separate quanti® ca-tion of di� erent substances (phytoplankton, mineral particles, yellow substance)to be performed. In open ocean waters optically dominated by phytoplanktonand their associate detrital matter, the basic information (i.e. the concentrationof phytoplanktonic pigments) will be retrieved with improved accuracy due tothe increased radiometric performances of MERIS. The adoption of multi-wave-length algorithms could also lead to additional information concerning auxiliarypigments and taxonomic groups. Finally, MERIS will be one of the ® rst sensorsto allow measurements of Sun-induced chlorophyll a in vivo ¯ uorescence, whichcould provide a complementary approach for the assessment of phytoplanktonabundance. The development of these next-generation algorithms, however,requires a number of fundamental studies.

1. Introduction

The capacity of satellites to provide synoptic information on phytoplanktondistribution in the upper layer of the ocean for various applications has been largelydemonstrated by the Coastal Zone Color Scanner (CZCS), launched by NationalAeronautics and Space Administration (NASA) on the satellite Nimbus 7 and whichoperated from late 1978 to early 1986 (see, for example, the review by Abbott and

² e-mail: annick@ccrv.obs-vlfr. fr

International Journal of Remote SensingISSN 0143-1161 print/ISSN 1366-5901 online Ñ 1999 Taylor & Francis Ltd

http://www.tandf.co.uk/JNLS/res.htmhttp://www.taylorandfrancis.com/JNLS/res.htm

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

A. Bricaud et al.1758

Chelton (1991)) . The ® rst application of ocean colour data was to study thespatial/temporal evolution of phytoplankton distribution for ecological applicationsor to trace water circulation. Since then, ocean colour data analysis has also beenfocused on attempting to derive more elaborate information from these near-surfacepigment ® elds. International programs such as the International Geosphere±Biosphere Program (IGBP) and the Joint Global Ocean Flux Study (JGOFS) haveemphasized the importance of deriving near-surface pigment ® elds from ocean colourdata, in particular (i) for initiating and validating numerical models of ecosystemsand biogeochemical processes, (ii) for calculating primary production at regional toglobal scales, and (iii) when used in synergy with other remote sensing techniques(altimetry, scatterometry, etc.), for forcing circulation ® elds in upper ocean numericalmodels of biogeochemical processes (Yoder 1996). Furthermore, with the availabilityof the global CZCS archive (McClain et al. 1991), the observation scales could beextended from the mesoscale to the basin- or world-scale (Banse and English 1994,Sathyendranath et al. 1995, Antoine et al. 1996).

In recent years much emphasis has been placed on the role of the oceanic`biological pump’ in atmospheric CO2 variations and the global carbon cycle.Evaluation and modelling of this biological pump requires knowledge of the spatialand temporal evolutions of marine primary production both at global and at regionalscales. Large uncertainties remain, however, in the magnitude of primary productionat these di� erent scales. For instance, until recently, the commonly proposed valuesat the global scale, based on compilations of in situ carbon ® xation, were in therange 20± 50GtC per year (see, for example, Berger (1989)) . These estimates wererecently re® ned using satellite ocean colour data combined with primary productionmodels to allow maps of surface pigment concentration to be converted into mapsof carbon ® xation rate (e.g. Platt and Sathyendranath 1988, Morel 1991). `Worldmaps’ of primary production were thus proposed, as derived from the CZCS archive(Antoine et al. 1996). Using CZCS data, regional maps or estimates of primaryproduction were also derived for speci® c areas, such as the Mauritanian upwellingarea (Bricaud et al. 1987), the North Atlantic (Sathyendranath et al. 1991, 1995)and the Western (Morel and Andre 1991) and Eastern (Antoine et al. 1995)Mediterranean Basins.

Analysis of CZCS images has also demonstrated that ocean colour is a powerfultool to characterize dynamical features (eddies, meanders, plumes, etc.) and to studythe variability of ocean circulation patterns, at mesoscale or large-scale (see, forexample, Barale and Trees (1987); see also the reviews by Abbott and Chelton (1991)and by various authors cited in the work of Barale and Schlittenhardt (1993)) . The® rst studies were essentially descriptive, mostly aimed at describing dynamic struc-tures and their temporal evolution via phytoplankton distribution. Over the last tenyears, however, the use of ocean colour data has greatly evolved, in conjunction withthe development of coupled physical ± biological models of increasing complexity,generally aimed at describing the biological response of the ocean to physical forcingsand evaluating the associated biogeochemical ¯ uxes (e.g. Bissett et al. 1994, Gloveret al. 1994, Pribble et al. 1994). In parallel with such models, data assimilationtechniques have also been developed, with the aim of constraining these models toremain realistic (e.g. Prunet et al. 1996a, b). Finally, ocean colour data have alsobeen used to predict, via the pigment concentration, the vertical pro® le of the heatingrate within the upper layer of the ocean (Morel and Antoine 1994).

For these various applications the capabilities of the CZCS were unfortunately

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

MERIS 1759

limited, both in spectral coverage (only three visible channels could be used formarine information) and radiometric accuracy (with a signal coding over 8 bits).Furthermore, atmospheric corrections, which have a critical importance since typi-cally 90% of the signal reaching the sensor originates from the atmosphere, weremostly based on empirical (iterative) schemes (Bricaud and Morel 1987, Gordonet al. 1988), because no channels were available to estimate the contribution ofaerosol scattering and its spectral dependence independently. As a consequence,under typical conditions, the satellite-derived pigment concentration values (Chl )were within Ô 0.3 log10 Chl when using the 443 to 550nm re¯ ectance ratio (Gordonet al. 1983). The accuracy was further degraded at high Chl concentrations, becauseof the low sensitivity of the blue (443nm) channel and the constraint of resorting toa less e� cient re¯ ectance ratio (520 to 550nm). Further, the CZCS, initially scheduledfor a lifetime of only one year, faced severe sensitivity degradation problems, whichgreatly impaired the quality of the retrieved pigment values, particularly during itslast years of operation (1983± 1986). This required the development of vicarious’correction schemes (Evans and Gordon 1994). Finally, it must be remembered thatthe CZCS experiment was planned as a proof-of-concept operation and as such didnot allow the sensor to be operated in a continuous mode, so that global coverageof the oceans could not be ensured.

The concept and usefulness of such an experiment being amply proven, permanentacquisition of data along the daylight part of the orbit has been selected for presentand future ocean colour sensors. Relying on the experience drawn from the CZCS,this new generation of instruments (Modular Optoelectronic Scanner (MOS), OceanColor and Thermal Scanner (OCTS) and Polarization and Directionality of EarthRe¯ ectances (POLDER) in 1996, Sea-viewing Wide Field-of-view Sensor (SeaWiFS)in 1997, Moderate Resolution Imaging Spectrometer (MODIS) in 1999, MediumResolution Imaging Spectrometer (MERIS) and Global Imager (GLI) in 2000)bene® ts, or will bene® t, from major instrumental improvements such as an increasedradiometric accuracy and sensitivity, the availability of infrared channels to estimatethe aerosol thickness and its spectral dependence, and better spectral information inthe visible. Moreover the sensors to be deployed in the late 1990s (MERIS, MODISand GLI) will have two speci® c additional features. (i ) Being equipped with numer-ous, programmable, channels, they will provide the most detailed spectral information(for MERIS, radiances will be measured in 15 channels over the visible± infrareddomain (see Rast and Bezy, this issue). (ii) They will provide measurements of theSun-induced chlorophyll a in vivo ¯ uorescence band around 680nm, which willconstitute a complementary approach to the classical assessment of phytoplanktonabundance.

Together with the improvements speci® cally concerning marine information, thepossibility to perform a more accurate determination of aerosol properties (andconsequently a more accurate retrieval of water-leaving radiances) using a schemeadapted from that developed for SeaWiFS by Gordon and Wang (1994) will producemajor bene® ts (Antoine and Morel, this issue). In the following we discuss only thoseaspects concerning the interpretation of marine signals.

2. Ocean colour signal inversion

Once atmospheric corrections are performed, MERIS will provide water-leavingradiances in nine visible channels, provisionally at 412.5, 442.5, 490, 510, 560, 620,665, 681.25 and 705nm. This will yield considerably more detailed information than

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

A. Bricaud et al.1760

that provided by the CZCS (where marine radiances were available at only 443, 520,550 and 670nm), or even, for instance, SeaWiFS (with visible channels at 412, 443,490, 510, 555 and 670nm). It is expected that the main bene® t of this increasedspectral information will concern Case II waters (essentially coastal areas). In theseareas, di� erent types of substances (phytoplankton, mineral particles, yellow sub-stance) coexist without being correlated in most cases. The possibility of quantifyingthe concentrations of these substances has prompted several studies over recentyears, with various approaches leading to the development of inverse modellingtechniques (Doer� er and Fischer 1994, Schiller and Doer� er, this issue) or multiple-wavelength algorithms (Carder et al. 1991, Tassan 1994, Moore et al., this issue.

In the open ocean (essentially Case I’ waters), the problem is somewhat simpler,as the predominant role in determining the optical properties is played by phyto-plankton and its associated biogenous (dissolved and particulate) by-products.Because phytoplankton cells absorb primarily blue radiation (with a maximumaround 435± 445nm), and only weakly green radiation, the quantitative retrieval ofalgal biomass from CZCS data has been based mostly on the empirical (statistical)relationship linking the variations of the blue-to-green re¯ ectance ratio (e.g. R 443 /

R 560 ) to those of chlorophyll a concentration (see ® gure 1). It is worth noting thatthe blue absorption results not only from the chlorophylls (a , b , c) themselves, butalso from various carotenoids always associated with chlorophyll a . The commonmethod, however, is to quantify the algal biomass in terms of the chlorophyll a

concentration.Di� erent combinations of band ratios have been proposed to handle speci® c

environmental cases (e.g. Aiken et al. 1995), also using empirical approaches. SuchCZCS-type’ algorithms will be applicable to MERIS as well; the main improvement,in comparison with the CZCS, will lie therefore in the increased radiometric accuracy,

Figure 1. Variations of re¯ ectance ratios versus the chlorophyll a concentration, Chl

(in mg mÕ3 ), as derived from the model of Morel (1988). The steps on the curves

correspond to the pigment classes to be detected, i.e. ten classes per decade (adaptedfrom the LPCM/ACRI contract report to ESA (1991)) .

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

MERIS 1761

which will result in increased accuracy of pigment retrieval. Thus, the requirementsof 5Ö 10Õ

4 for the NEDR (noise equivalent spectral re¯ ectances at sea level) set forMERIS should allow the instrument to detect ten classes of pigment concentrationper decade (with a 26% increase from one class to another), except over restrictedportions of scenes (LPCM/ACRI 1991). As shown in ® gure 1, the R 443 /R 560 ratiospans the widest range and is therefore the most sensitive to variations in algalpigment concentration.

Independent of other sources of uncertainty (atmospheric corrections, radiometricnoise), the ability of CZCS-type (i.e. purely empirical ) algorithms to retrieve thewater pigment content is necessarily limited by the bio-optical variability of waters,even in the open ocean. Spectral variations of the di� use re¯ ectance at null depth,R (0Õ ), are actually ruled by the bulk absorption (a ) and backscattering (bb ) coe� -cients for the water body:

R (0Õ )= f bb /a (1)

where f is a factor dependent on illumination conditions (Kirk 1984, Gordon 1989)and water content and varies between 0.3 and 0.5 (Morel and Gentili 1991). Thea and bb coe� cients (inherent optical properties (IOPs) (Preisendorfer 1961)) arethe sums of the contributions of the diverse substances. For Case I waters these arepure seawater, living phytoplankton, non-algal particulate matter (NAP) (includingvarious detritus as well as heterotrophic bacteria) and endogenous yellow substance(coloured dissolved organic matter (DOM) resulting from biological activity):

a =aw+[Chl] a*u +[NAP] a*

NAP+[Y] aY (2)

bb = bbw+[Chl] b*b u +[N AP] b*

b NAP (3)

where the above substances are represented, respectively, by the subscripts w, w ,NAP and Y, the terms in brackets are their respective concentrations, and the starredcoe� cients are the substance-speci® c optical coe� cients, which are known to bevariable (the wavelength l is omitted in each term for the sake of simplicity).Therefore any progress with respect to empirical approaches depends on the develop-ment of analytical’ or semi-analytical’ algorithms, i.e. based on bio-optical modelswhich relate the bb /a ratio to the substance concentrations and take into accountvariations in the IOPs.

Semi-analytical algorithms generally refer to simpli® ed bio-optical models (i.e.simpli® ed versions of equations (2) and (3)), implicitly admitting natural covariationbetween the concentrations of some substances (e.g. Gordon et al. 1988, Morel 1988,Carder et al. 1991, Lee et al. 1994). Such models result in nonlinear relationshipsbetween re¯ ectance ratios and chlorophyll a concentration, which are more represent-ative of reality than simple regression lines. These relationships can be ® tted topolynomials for practical use (® gure 2). Note that the existing models di� er notablywhen the entire range of concentrations is considered, as a result of di� eringparameterizations of the variations in absorption coe� cient (or alternatively,di� use attenuation coe� cient) and backscattering coe� cient with chlorophyll a

concentration. This emphasizes the need for further fundamental work at sea and inthe laboratory for validating such semi-analytical algorithms.

Entirely analytical algorithms, in which the improved spectral resolution providedby MERIS could ® nd optical use, still remain to be developed. One of the pointsto be studied is whether detailed spectral information would indeed allow some

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

A. Bricaud et al.1762

Figure 2. Variations of blue-to-green ratio, R 443 /R560, versus chlorophyll a concentration,Chl (in mg mÕ

3 ), for two di� erent bio-optical models: Morel 1988 (M88) and Gordonet al. 1988 (G88), ® tted to polynomial functions of order three. Note that recentmodels account in particular for the new ® ndings concerning absorption coe� cientsof pure seawater (Pope and Fry 1997).

phytoplanktonic groups to be distinguished. The inherent limitation of CZCS-type(or even semi-analytical) algorithms is that they can only retrieve a single parameter(chlorophyll a concentration), while phytoplanktonic absorption in the various chan-nels (and especially at 443nm) originates from di� erent groups of pigments, namelychlorophylls a , b, c and the various carotenoids (upper panel of ® gure 3). Thesegroups of pigments, in highly variable concentrations according to species and physio-logical or photoadaptive state of cells (see example in ® gure 4), are (along with thepackaging e� ect) the main source of variability in the absorption properties of naturalphytoplanktonic populations (see lower panel of ® gure 3 and, for example, Mitchelland Kiefer (1988), Hoep� ner and Sathyendranath (1992), Sosik and Mitchell (1995),Bricaud et al. (1995) and Cleveland (1995)) . For instance, dino¯ agellates contain acarotenoid (peridinin) absorbing mainly in the domain 490± 530nm and responsiblefor the reddish discolouration of waters, known as red tide’ (e.g. Gieskes and Kraay1986). Red tides (which occur mostly in coastal areas, but sometimes also in Case Iwaters) can also be produced by cryptophytes, which contain phycobilin pigments,absorbing in the domain 500± 600nm (i.e. the domain of supposed minimum absorp-tion’ for phytoplankton). Absorption in the green channels is also strongly increasedin the presence of the phycobiliprotein-containing cyanobacteria, which may resultin a conspicuous depression on downwelling irradiance spectra (® gure 5). Finally,

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

MERIS 1763

Figure 3. Upper panel. In vivo pigment-speci® c absorption coe� cients for various phyto-planktonic pigments. These spectra were obtained by scaling relative absorptionspectra in solvents obtained by HPLC (H. Claustre, personal communication) toweight-speci® c absorption coe� cients at given wavelengths (Goericke and Repeta1993), and by shifting the absorption maxima to their known in vivo positions. Lowerpanel. Examples of chlorophyll a-speci® c absorption spectra measured in the upperlayer of the tropical North Atlantic, in three sites di� ering by their trophic status. Thedi� erences between these spectra originate from di� erences in both pigment composi-tion and packaging e� ect (the latter is ruled by variations in average cell size andpigment content per cell (Morel and Bricaud 1981)) .

coccolithophorids are also well known for their particular optical properties, notbecause of speci® c pigmentation but because they produce detached calcite plateswhich increase the scattering coe� cient of the water body and therefore its re¯ ectanceto levels as high as 15%, as observed by Holligan et al. (1983). Algorithms simplybased on re¯ ectance ratios become inoperative in such cases.

Speci® c re¯ ectance or water-leaving radiance models have already been developed

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

A. Bricaud et al.1764

Figure 4. Divinyl chlorophyll a-speci® c absorption spectra of two Prochlorococcus cultures,acclimated at two di� erent growth irradiances. The lower absorption coe� cientsobserved at low irradiance result from two combined photoadaptive e� ects: a lowerzeaxanthin-to-divinyl chl a ratio, and a higher content of divinyl chlorophyll a per cell.

for dino¯ agellate blooms (Carder and Steward 1985, Balch et al. 1989) and cocco-lithophore blooms (Balch et al. 1989, Ackleson et al. 1994). Optimization techniqueshave been proposed to retrieve in-water parameters from model inversions, usingthe whole spectral information instead of simple re¯ ectance ratios. Applying anothertechnique (factor analysis) to a set of water-leaving radiance spectra measured incoastal waters from the New York Bight region, Sathyendranath et al. (1994) showedthat with the SeaWiFS channels, it could be possible to distinguish between phyco-erythrin and chlorophyll pigments. Therefore, the availability of still more numerousvisible channels on MERIS suggests the potential for the identi® cation of speci® calgal groups. Analysing particulate absorption spectra from open ocean waters,Garver et al. (1994) found nevertheless that more than 99% of the variance in theirdataset was associated with chlorophyll, and emphasized the di� culty of developingrobust algorithms for species identi® cation. It is expected that such identi® cationwill in any case require a conspicuous pigment signature from a particular group(see Morel 1997).

3. Sun-induced chlorophyll a ¯ uorescence

The presence of the Sun-induced chlorophyll a ¯ uorescence band in the water-leaving radiance spectra of phytoplankton-rich waters led Neville and Gower (1977),more than twenty years ago, to propose a quanti® cation of the pigment concentrationvia the height of this band with respect to a baseline’ . The main advantage of thismethod is its speci® city, because the ¯ uorescence band, contrary to the spectralvariations of re¯ ectance, is not a� ected by the presence of particulate substancesother than phytoplankton. Therefore such a method could prove to be a valuabletool, especially in Case II waters. Even in oceanic Case I waters with moderate orhigh phytoplankton concentrations, it may represent a useful alternative to the useof CZCS-type or advanced re¯ ectance ratio algorithms. However, limitations inapplications of this technique do exist, and result mainly from absorption by water

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

MERIS 1765

Figure 5. Upper panel. Example of absorption spectrum measured o� Mauritania(EUMELI 4 cruise, June 1992), in waters dominated by the cyanobacteriumSynechococcus. The arrow indicates the absorption band of phycoerythrin character-istic of this group. Lower panel. Downwelling irradiance spectra measured at the samesite. Note the marked depression around 550nm, induced by the presence of phyco-erythrin (see also re¯ ectance spectra in Morel 1997).

itself, from the perturbation of the ¯ uorescence signal by the absorption band ofatmospheric oxygen, and from the variability of the ¯ uorescence emission per unitof chlorophyll a concentration (Babin et al. 1996). A description of the capabilitiesand limitations of this approach is provided by Gower et al. (this issue) .

4. Conclusions

MERIS will acquire data in a continuous way from pole to pole (actually whenthe zenith Sun angle is less than 80ß ), and will thus ful® l the requirements of a globaland permanent coverage of the world’s oceans, as recommended by JGOFS (Yoder1996). The previously developed ocean colour techniques (inherited from CZCS,SeaWiFS and OCTS) will be operated on a routine basis. Thanks to an internationalcoordination resulting in the adoption of a series of common wavelengths, thecompatibility of algorithms will be ensured, leading to comparable basic products.

The MERIS sensor will also provide the potential for a more accurate retrieval

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

A. Bricaud et al.1766

of optically active water constituents, because of (i ) increased radiometric per-formances, (ii) improved atmospheric corrections, and (iii) more detailed spectralinformation on marine water-leaving radiances (including ¯ uorescence) . In conjunc-tion with these increased performances, some phenomena previously neglected withthe CZCS will become of critical importance, and will have to be taken into accountfor an optimized retrieval of pigment ® elds. The phenomena to be accurately modelledconcern both atmospheric and marine optics. Preserving the sensitivity of the instru-ment, i.e. propagating radiometric accuracy from satellite level to the determinationof water-leaving radiances, requires a very precise atmospheric correction. A schemefor a second generation’ correction procedure, much more complex than that usedwith CZCS, has been developed for SeaWiFS (Gordon and Wang 1994) andimproved for MERIS. It accounts for a variety of aerosols and for the e� ects of theinterface (sea surface roughness, white caps) in modifying radiative transfer in theatmosphere and then the atmospheric path radiance (Antoine and Morel, this issue).In the domain of marine optics, the non-isotropic character of the upward radiance® eld, previously disregarded, has been evidenced from Monte Carlo computations(Morel and Gentili 1993) and experimentally con® rmed (Morel et al. 1995); it willtherefore have to be accounted for. Because the anisotropy factor Q varies withpigment concentration in the water column, it will be necessary to incorporateiterative procedures in the data processing sequence in order to derive a preliminarypigment index before proceeding further (Morel and Gentili 1996). In addition,fundamental work in developing analytical bio-optical models able to account forthe variability in the relevant speci® c coe� cients and their relationship to environ-mental factors is still needed. Establishing advanced’ algorithms applicable toMERIS spectral data and deriving speci® c products implies, as a prerequisite, suchfundamental research.

Acknowledgments

Continuous ® nancial support from the European Spatial Agency for the devel-opment of atmospheric and bio-optical algorithms for MERIS is gratefullyacknowledged.

References

Abbott, M. R., and Chelton, D. B., 1991, Advances in passive remote sensing of the ocean.Reviews of Geophysics, Supplement, 29, 571± 589.

Ackleson, S. G ., Balch, W . M ., and Holligan, P. M., 1994, Response of water-leavingradiance to particulate calcite and chlorophyll a concentrations: a model for Gulf ofMaine coccolithophore blooms. Journal of Geophysical Research, 99, 7483± 7499.

Aiken, J., Moore, G . F., Clark, D . K ., and Trees, C. C., 1995, The SeaWiFS CZCS-typepigment algorithm. SeaWiFS Technical Report Series, NASA Technical Memorandum,vol. 29, edited by S. B. Hooker and E. R. Firestone.

Antoine, D., Andre’, J. M., and Morel, A., 1996, Oceanic primary production: II. Estimationat a global scale from satellite (Coastal Zone Color Scanner) chlorophyll. GlobalBiogeochemical Cycles, 10, 57± 69.

Antoine, D ., and Morel, A., 1999, A multiple scattering algorithm for atmospheric correctionof remotely-sensed ocean colour (MERIS instrument): principle and implementationfor atmospheres carrying various aerosols including absorbing ones. InternationalJournal of Remote Sensing, 1999.

Antoine, D ., Morel, A., and Andre’, J. M., 1995, Algal pigment distribution and primaryproduction in the Eastern Mediterranean as derived from Coastal Zone Color Scannerobservations. Journal of Geophysical Research, 100, 16193± 16209.

Babin, M ., Morel, A., and Gentili, B., 1996, Remote sensing of sea surface sun-induced

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

MERIS 1767

chlorophyll ¯ uorescence: consequences of natural variations in the optical character-istics of phytoplankton and the quantum yield of chlorophyll a ¯ uorescence.International Journal of Remote Sensing, 17, 2417± 2448.

Balch, W . M ., Eppley, R. W ., Abbott, M . R., and Reid, F. M. H., 1989, Bias in satellite-derived pigment measurements due to coccolithophores and dino¯ agellates. Journalof Plankton Research, 11, 575± 581.

Banse, K., and English, D. C., 1994, Seasonality of coastal zone color scanner phytoplanktonpigment in the o� shore oceans. Journal of Geophysical Research, 99, 7323± 7345.

Barale, V., and Schlittenhardt, P. M. (editors), 1993, Ocean Colour: T heory and Applicationsin a Decade of CZCS Experience (Dordrecht: Kluwer).

Barale, V., and Trees, C. C., 1987, Spatial variability of the ocean colour ® eld in CZCSimagery. Advances in Space Research, 7, (2)95± (2)100.

Berger, W. H., 1989, Global maps of ocean productivity. In Productivity of the ocean: presentand past, edited by W. H. Berger, V. S. Smetacek and G. Wefer (New York: Wiley),pp. 429± 455.

Bissett, W . P., Meyers, M . B., Walsh, J. J., and M uller-Karger, F. E., 1994, The e� ectsof temporal variability of mixed layer depth on primary productivity around Bermuda.Journal of Geophysical Research, 99, 7539± 7553.

Bricaud, A., Babin, M ., Morel, A., and Claustre, H., 1995, Variability in the chlorophyll-speci® c absorption coe� cients of natural phytoplankton: analysis and parameteri-zation. Journal of Geophysical Research, 100, 13 321± 13 332.

Bricaud, A., and Morel, A., 1987, Atmospheric corrections and interpretation of marineradiances in CZCS imagery: use of a re¯ ectance model. Oceanologica Acta, SP, 7, 33± 50.

Bricaud, A., Morel, A., and Andre’, J. M., 1987, Spatial/temporal variability of algal biomassand potential productivity in the Mauritanian upwelling zone, as estimated fromCZCS data. Advances in Space Research, 7, (2), 53± 62.

Carder, K . L., Hawes, S. K ., Baker, K . A., Smith, R. C., Steward, R. G ., and M itchell,B. G., 1991, Re¯ ectance model for quantifying chlorophyll a in the presence of produc-tivity degradation products. Journal of Geophysical Research, 96, 20599± 20611.

Carder, K. L., and Steward, R. G., 1985, A remote-sensing re¯ ectance model of a red-tidedino¯ agellate o� west Florida. L imnology and Oceanography, 30, 286± 298.

Cleveland, J. S., 1995, Regional models for phytoplankton absorption as a function ofchlorophyll a concentration. Journal of Geophysical Research, 100, 13333± 13344.

Doerffer, R., and Fischer, J., 1994, Concentrations of chlorophyll, suspended matter, andgelbsto� in Case II waters derived from satellite coastal zone color scanner data withinverse modeling methods. Journal of Geophysical Research, 99, 7457± 7466.

Evans, R. H., and Gordon, H. R., 1994, Coastal Zone Color Scanner system calibration’: aretrospective examination. Journal of Geophysical Research, 99, 7293± 7307.

Garver, S. A., Siegel, D . A., and M itchell, B. G., 1994, Variability in near-surface particulateabsorption spectra: what can a satellite ocean color imager see? L imnology andOceanography, 39, 1349± 1367.

G ieskes, W . W . C., and Kraay, G. W., 1986, Analysis of phytoplankton pigments by HPLC,before, during and after mass occurrence of the micro¯ agellate Corymbellus aureusduring the spring bloom in the open northern Sea in 1983. Marine Biology, 92, 45± 52.

Glover, D . M ., Wroblewski, J. S., and McClain, C. R., 1994, Dynamics of the transitionzone in coastal zone color scanner-sensed ocean color in the North Paci® c duringoceanographic spring. Journal of Geophysical Research, 99, 7501± 7511.

Goericke, R., and Repeta, D. J., 1993, Chlorophylls a and b and divinyl-chlorophylls a andb in the open subtropical North Atlantic. Marine Ecology Progress Series, 101, 307± 313.

Gordon, H. R., 1989, Dependence of the di� use re¯ ectance of natural waters on the Sunangle. L imnology and Oceanography, 34, 1484± 1489.

Gordon, H . R., Brown, O. B., Evans, R. H ., Brown, J. W ., Smith, R. C., Baker, K. S.,and Clark, D. K., 1988, A semianalytic radiance model of ocean color. Journal ofGeophysical Research, 93, 10909± 10924.

Gordon, H . R., Clark, D . K ., Brown, J. W ., Brown, O. B., Evans, R. H ., and Broenkow,W. W., 1983, Phytoplankton pigment concentrations in the Middle Atlantic Bight:comparison between ship determinations and Coastal Zone Color Scanner estimates.Applied Optics, 22, 20± 36.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

A. Bricaud et al.1768

Gordon, H. R., and Wang, M., 1994, Retrieval of water-leaving radiance and aerosol opticalthickness over the ocean with SeaWiFS: a preliminary algorithm. Applied Optics,33, 443± 452.

Gower, J. F. R., Doerffer, R., and Borstad, G., 1999, Interpretation of the 685nm peak inwater-leaving radiance spectra in terms of ¯ uorescence, absorption and scattering, andits observation by MERIS. International Journal of Remote Sensing, 1999.

Hoepffner, N., and Sathyendranath , S., 1992, Bio-optical characteristics of coastal waters:Absorption spectra of phytoplankton and pigment distribution in the Western NorthAtlantic. L imnology and Oceanography, 37, 1660± 1679.

Holligan, P. M ., Viollier, M ., Harbour, D . S., Camus, P., and Champagne-Philippe, M.,1983, Satellite and ship studies of coccolithophore production along a continentalshelf edge. Nature , 304, 339± 342.

K irk, J. T. O., 1984, Dependence of relationship between inherent and apparent opticalproperties of water on solar altitude. L imnology and Oceanography, 29, 350± 356.

Lee, Z., Carder, K . L., Hawes, S. K ., Steward, R. G ., Peacock, T. G ., and Davis, C. O.,1994, Model for the interpretation of hyperspectral remote-sensing re¯ ectance. AppliedOptics, 33, 5721± 5732.

LPCM/ACRI, 1991, Quanti® cation of the ocean backscattered radiance. Characteristics ofthe signal at the top of the atmosphere. ESA Contract no 9111/90/NL/PR(SC), editedby A. Morel and B. Gentili.

McClain, C. R., Feldman, G . C., and Esaias, W. E., 1991, A review of the Nimbus 7-CoastalZone Color Scanner data set. In Global Change Atlas, edited by C. Parkinson, J. Fosterand R. Gurney (Cambridge: Cambridge University Press).

M itchell, B. G., and K iefer, D. A., 1988, Variability in pigment speci® c particulate ¯ uores-cence and absorption spectra in the northeastern Paci® c Ocean. Deep-Sea Research,35, 665± 689.

Moore, G . F., Aiken, J., Lavender, J., and Botrell, H., 1999, The atmospheric correctionof water colour and the quantitative retrieval of biogeochemical variables in Case IIwaters: application to MERIS. International Journal of Remote Sensing, 1999.

Morel, A., 1988, Optical modeling of the upper ocean in relation to its biogenous mattercontent (Case I waters). Journal of Geophysical Research, 93, 10749± 10768.

Morel, A., 1991, Light and marine photosynthesis: a spectral model with geochemical andclimatological implications. Progress in Oceanography, 26, 263± 306.

Morel, A., 1997, Consequences of a synechococcus bloom upon the optical properties ofoceanic (Case 1) waters. L imnology and Oceanography, 42, 1746 ± 1754.

Morel, A., and Andre’, J. M., 1991, Pigment distribution and primary production in thewestern Mediterranean as derived and modeled from Coastal Zone Color Scannerobservations. Journal of Geophysical Research, 96, 12685± 12698.

Morel, A., and Antoine, D., 1994, Heating rate within the upper ocean in relation to its bio-optical state. Journal of Physical Oceanography, 24, 1652± 1665.

Morel, A., and Bricaud, A., 1981, Theoretical results concerning light absorption in a discretemedium, and application to speci® c absorption of phytoplankton. Deep-Sea Research,28, 1375± 1393.

Morel, A., and Gentili, B., 1991, Di� use re¯ ectance of oceanic waters: its dependence onSun angle as in¯ uenced by the molecular scattering contribution. Applied Optics, 30,4427± 4438.

Morel, A., and Gentili, B., 1993, Di� use re¯ ectance of oceanic waters (2): bi-directionalaspects. Applied Optics, 32, 6864± 6879.

Morel, A., and Gentili, B., 1996, Di� use re¯ ectance of oceanic waters (3): implication ofbidirectionality for the remote sensing problem. Applied Optics, 35, 4850± 4862.

Morel, A., Voss, K . J., and Gentili, B., 1995, Bidirectional re¯ ectance of oceanic waters: acomparison of modeled and measured upward radiance ® elds. Journal of GeophysicalResearch, 100, 13143± 13150.

Neville, R. A., and Gower, J. F. R., 1977, Passive remote sensing of phytoplankton via chla ¯ uorescence. Journal of Geophysical Research, 82, 3487± 3493.

Platt, T., and Sathyendranath , S., 1988, Oceanic primary production: estimation by remotesensing at local and regional scales. Science, 241, 1613± 1620.

Pope, R. M., and Fry, E. S., 1997, Absorption spectrum (380± 700nm) of pure water, II,Integrating cavity measurements. Applied Optics, 36, 8710± 8723.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014

MERIS 1769

Preisendorfer, R. W., 1961, Application of radiative transfer theory to light measurementsin the sea. In: Radiant Energy in the Sea. Vol. 10, International Union of Geodesy andGeophysics, Monography, 11± 29.

Pribble, J. R., Walsh, J. J., D ieterle, D . A., and M uller-Karger, F. E., 1994, A numericalanalysis of shipboard and coastal zone color scanner time series of new productionwithin Gulf Stream cyclonic eddies in the South Atlantic Bight. Journal of GeophysicalResearch, 99, 7513± 7538.

Prunet, P., M inster, J. F., Ruiz-Pino, D ., and Dadou, I., 1996a, Assimilation of surfacedata in a one-dimensional physical-biogeochemical model of the surface ocean. (1)Method and results with chlorophyll data. Global Biogeochemical Cycles, 10, 111± 138.

Prunet, P., M inster, J. F., Echevin, V., and Dadou, I., 1996b, Assimilation of surface datain a one-dimensional physical-biogeochemical model of the surface ocean. (2) Adjustinga simple trophic model to chlorophyll, temperature, nitrate and pCO2 data. GlobalBiogeochemical Cycles, 10, 139± 158.

Rast, M., and Bezy, J. L., 1999, The MERIS instrument. International Journal of RemoteSensing, 1999.

Sathyendranath, S., Hoge, F. E., Platt, T., and Swift, R. N., 1994, Detection of phytoplank-ton pigments from ocean color: improved algorithms. Applied Optics, 33, 1081± 1089.

Sathyendranath, S., Longhurst, A. R., Caverhill, C. M ., and Platt, T., 1995, Regionallyand seasonally di� erentiated primary production in the North Atlantic. Deep-SeaResearch, 42, 1773± 1802.

Sathyendranath, S., Platt, T., Horne, A. P. W ., Harrison, W . G ., Ulloa, O.,Outerbridge, R., and Hoepffner, N., 1991, Estimation of new production in theocean by compound remote sensing. Nature , 353, 129± 133.

Schiller, H., and Doerffer, R., 1999, Neural network for emulation of an inverse model-operational derivation of Case II water properties from MERIS data. InternationalJournal of Remote Sensing, 1999.

Sosik, H. M., and M itchell, B. G., 1995, Light absorption by phytoplankton, photosyntheticpigments, and detritus in the California Current System, Deep-Sea Research, 42,1717± 1748.

Tassan, S., 1994, Local algorithms using SeaWiFS data for the retrieval of phytoplankton,pigments, suspended matter, and yellow substance in coastal waters. Applied Optics,33, 2369± 2378.

Yoder, J. A. (editor), 1996, Remote sensing in the JGOFS Program. Report No. 20, JGOFS,Bergen, Norway.

Dow

nloa

ded

by [

Uni

vers

ity o

f T

asm

ania

] at

15:

12 1

5 N

ovem

ber

2014