use of in vivo chlorophyll a fluorescence to quantify ... · chl a per square metre and enables the...

MARINE ECOLOGY PROGRESS SERIESMar Ecol Prog Ser

Vol. 218: 45–61, 2001 Published August 20

INTRODUCTION

The efficiency of utilisation of sunlight by photosyn-thetic communities is determined by the ratio of photo-

synthetically converted to available light energy (Mo-rel 1978, Dubinsky et al. 1984). Because photosyntheticproduction results from the sequential absorption of lightand conversion of the absorbed light energy, the com-munity-level photosynthetic efficiency is determinedboth by the efficiency of light absorption (the ratio ofabsorbed to available light) and by the efficiency ofconversion of absorbed light (the ratio of converted to

© Inter-Research 2001

*Present address: Departamento de Biologia, Universidadede Aveiro, Campus de Santiago 3810-193 Aveiro, Portugal.E-mail: [email protected]

Use of in vivo chlorophyll a fluorescence to quantify short-term variations in the productive

biomass of intertidal microphytobenthos

João Serôdio1, 2,*, Jorge Marques da Silva2, 3, Fernando Catarino1, 2

1Instituto de Oceanografia, 2Departamento de Biologia Vegetal and 3Centro de Engenharia Biológica, Faculdade de Ciências da Universidade de Lisboa, Campo Grande, 1749–016 Lisboa, Portugal

ABSTRACT: This study investigates the ability to use in vivo chlorophyll a fluorescence to quantifythe productive biomass of undisturbed microphytobenthic communities, defined as the photosyn-thetic biomass present in the photic zone of the sediment and actually contributing to measurablephotosynthesis. The purposes of defining and quantifying productive biomass are (1) to evaluate theeffect of the migratory rhythms on the variability of microphytobenthic photosynthesis and (2) tocharacterise the community photophysiological response independently of the migratory stage,through the estimation of biomass-specific, community-level photosynthetic rates. The possibility ofusing chl a fluorescence, as measured non-destructively at the sediment surface, to trace variationsin the productive biomass of microphytobenthos was confirmed by testing (1) the variability of therelationship between fluorescence emission and chl a concentration under varying temperature andirradiance levels and (2) the effects of natural variability in the vertical profile of chl a within thephotic zone of the sediment on the depth-integrated fluorescence signal, F, measurable at the sur-face. Dark-level fluorescence, F0, was found to allow for tracing variations in chl a concentrationunder the range of temperature and irradiance variability found in situ. Depth-integration of fluores-cence emission resulted in only a fraction of total productive biomass being detectable at the surface.However, this fraction was found to be sufficiently constant to allow for the use of F0 to proportionallyfollow variations in community productive biomass. On average, measurements of productive bio-mass on natural samples overestimated by a factor of 1.30 during low tide and underestimated by afactor of 0.66 during high tide, mostly due to variations in the chl a profile in the photic zone associ-ated with vertical migrations. The method was applied to intertidal microphytobenthic communitiesof the Tagus estuary, Portugal, by non-destructively measuring F0 (using pulse amplitude modulatedfluorometry) and photosynthesis (using oxygen microelectrodes) on the same samples under in situconditions. The results showed that a significant proportion of the hourly and fortnightly variability inthe community photosynthetic light response was explained solely by variations in F0 associated withmovements of microalgae, identifying migratory rhythms as the main cause for short-term variabilityin intertidal benthic primary productivity.

KEY WORDS: Microphytobenthos · Photosynthesis · Migratory rhythms · Chlorophyll a fluorescence ·Light absorption

Resale or republication not permitted without written consent of the publisher

Mar Ecol Prog Ser 218: 45–61, 2001

absorbed light). Intertidal microphytobenthos living onsoft-sediment habitats is composed of motile microalgaeexhibiting rhythmic vertical migrations across the photiczone of the sediment in close synchronisation with tidaland day/night cycles (Round & Palmer 1966, Paterson1986). The vertical movement of the cells results inrapid and substantial variations of the fraction of theincident light on the surface that is intercepted by themicroalgal light-harvesting complexes and absorbedfor photosynthesis. Variability in the community pri-mary production may thus originate not only fromchanges in the quantum efficiency of photosynthesis,determined by the photophysiological characteristicsof the microalgal population exposed to light, but alsofrom variations in light absorption efficiency.

The quantification of the fraction of incident lightthat is absorbed for photosynthesis by the whole micro-phytobenthic community is necessary to distinguishthe contribution of migratory rhythms and photophysi-ological processes for the short-term variability oftenobserved in the community photosynthesis. By elimi-nating the variability associated with light absorption,this also allows characterisation of the photophysiolog-ical response of the benthic microalgae independent ofvariations in cell concentration in the photic zone. Thishas been approached by the calculation of areal bio-mass-specific primary production rates, through thenormalisation to the chlorophyll (chl) a content of theupper millimetres of sediment (Pinckney & Zingmark1991, Blanchard & Montagna 1992, Brotas & Catarino1995). However, in spite of its widespread utilisation,this practice has some important shortcomings. First,because the sectioning depth is much larger than thephotic zone of the sediment, most of the quantified bio-mass is certainly not contributing to the measured pro-duction (Grant 1986, Pinckney & Zingmark 1993). Thismakes it impossible to distinguish changes in the bio-mass present in the photic zone from physiological pro-cesses such as photoadaptation or photoinhibition. Sec-ond, biomass-specific photophysiological parameterscalculated in this way are not directly comparable withbiomass-specific values obtained either for resuspendedmicrophytobenthos or phytoplankton, for which it maybe assumed that all quantified chl a contributes equallyto measured photosynthesis. Third, it is a destructivemethod, which prevents repeated measurements onthe same sample over time and limits the precision andthe spatial and temporal resolution of the experiments.

This study addresses the definition and measure-ment of the productive biomass of microphytobenthicassemblages, a quantity defined on the basis of thephotosynthetic biomass present in the photic zone ofthe sediment, and allowing the normalisation of pho-tosynthetic rates to the chl a actually contributing tomeasured photosynthesis. The study examines the

possibility of utilising in vivo chl a fluorescence, asmeasured non-destructively near the sediment surfaceusing a pulse amplitude modulation (PAM; Schreiberet al. 1986) fluorometer to (1) monitor variations in theproductive biomass of undisturbed microphytobenthicassemblages and (2) estimate biomass-specific com-munity-level photosynthetic rates and photosynthesis-irradiance (P vs E) curve parameters. Non-intrusivemeasurement of fluorescence intensity at the sedimentsurface was previously seen to allow the tracing ofmigratory rhythms, through the detection of variationsin the chl a content of the upper layers of the sediment(Serôdio et al. 1997), but the relation between the mea-sured fluorescence signal and productive biomass wasnot determined.

The method was applied in the quantification of theeffects of changes in productive biomass in the photo-synthetic light response of microphytobenthos commu-nities of the Tagus estuary, Portugal, under in situ con-ditions, over hourly (migratory rhythms), fortnightlyand seasonal time scales. The basis for the method isfirst established by considering the theoretical rela-tionship between photosynthesis and productivebiomass of microphytobenthic assemblages, and thedepth-integrated emission of chl a fluorescence, mea-surable at the sediment surface.

THEORY

Community-level photosynthesis and productivebiomass. The rate of photosynthesis taking place in aunit volume of sediment at depth z is determined bythe light available at that depth, the fraction of thislight which is absorbed by the energy-absorbing pig-ments of the light-harvesting complexes of the photo-systems, and the efficiency of photosynthetic conver-sion of absorbed light (Kiefer et al. 1989, Kolber &Falkowski 1993):

P (z) = E0(z)a*PAR(z)c (z)φP (z) (1)

where E0 is the incident scalar irradiance, integratedover the spectrum of photosynthetically active radia-tion (PAR; 400 to 700 nm), a*PAR is the optical spectrallyaveraged absorption cross-section of energy-absorbingpigments normalised to chl a [assuming E0a*PAR =∫PARE0(λ)a*(λ)dλ], c is the chl a concentration, and φP isthe quantum yield of photosynthesis (see Table 1 fornotation); the product of a*PAR and c defines the effi-ciency of photosynthetic light absorption at depth z,determined by the ratio of available to absorbed irradi-ance. The photosynthetic rate of the microphytoben-thic community is given by integrating over depth therate of photosynthesis at each depth z between the sur-face and some depth z = zP, at which the rate of photo-

46

Serôdio et al.: Microphytobenthos productive biomass

synthesis equals LP, the minimum detectable valuewith the measuring method in use. Depth-integratedphotosynthesis may be expressed as a function of thedownwelling irradiance at the surface of the sediment,E, the parameter routinely measured when character-ising light availability for benthic communities (Grant1986, Pinckney & Zingmark 1993, Brotas & Catarino1995), by:

(2)

where RPAR is the irradiance reflectance (the ratio ofupward to downward irradiance), µPAR the averagecosine (the mean cosine of the average path directionof incident photons), and kP the attenuation coefficientfor downwelling photosynthetically active irradiancewithin the sediment (Kühl & Jørgensen 1994).

The community-level microphytobenthic photosyn-thesis is thus determined by the joint vertical distribu-tion of the fraction of light entering the sediment that is

absorbed for photosynthesis and of the quantum yieldof photosynthesis along the gradient of available light.Migratory rhythms affect community-level photosyn-thetic rates by altering the fraction of the light incidentat the surface that is absorbed for photosynthesisby the whole community, caused by changes in theamount of chl a in the photic zone of the sediment. It istherefore useful to define the productive biomass ofthe microphytobenthic assemblage as the total chl apresent in the photic zone, weighted at each depth byits contribution to total photosynthesis, given by thefraction of incident irradiance that is available for pho-tosynthesis (Fig. 1A):

(3)

Defined in these terms, cP has units of milligrams ofchl a per square metre and enables the normalisationof community-level photosynthetic rates (mmol O2 m–2

h–1) to yield areal biomass-specific rates, PB (mmol O2

mg–1 chl a h–1), thus eliminating the effect of changes

c c z

R zz

zz

k zP

0

PAR

PAR

PPe d= ( ) − ( )

µ ( )∫ −1

P E P z z

Ea z c z P zR z

zz

z

zk z

( ) = ( )

= ( ) ( ) ( ) − ( )µ ( )

∫∫ −

0

0PAR

PAR

PAR

P

PP

d

e d * φ 1

47

Table 1. Notation used in the text

a*PAR, a*650 Optical spectral (400 to 700 nm, 650 nm) absorption cross-section normalised to chl a (m2 mg–1 chl a)α, αB Initial slope and biomass-specific initial slope of the depth-integrated P vs E curve

[mmol O2 m–2 h–1(µmol m–2 s–1)–1, mmol O2 mg–1 chl a h–1(µmol m–2 s–1)–1]c Chl a concentration (mg chl a m–3)cP, cF Productive biomass and biomass detected by depth-integrated fluorescence (mg chl a m–2)chl ChlorophyllE, E650 Downwelling spectral irradiance (400 to 700 nm, 650 nm) at the sediment surface (W m–2 nm–1 or µmol m–2 s–1)E0 Scalar spectrally averaged (400 to 700 nm) irradiance (W m–2 nm–1 or µmol m–2 s–1)ε Community-level efficiency of photosynthetic light absorptionF Depth-integrated upwelling fluorescence intensity detected at the surface (V)F(z) Upwelling fluorescence intensity emitted at depth z (W m–2 or µmol m–2 s–1)F0, Fm Dark-level and maximum fluorescence intensity emitted by a dark-adapted sample (V) Fs, Fm’ Steady-state and maximum fluorescence intensity emitted by a light-adapted sample (V)f Fraction of productive biomass detected by depth-integrated fluorescence φF Quantum yield of fluorescence [quanta emitted (quanta absorbed)–1]φF0, φFm Dark-level and maximum quantum yield of chl a fluorescence of a dark-adapted sample

[quanta emitted (quanta absorbed)–1]φP Quantum yield of photosynthesis [mol O2 (mol quanta absorbed)–1]G Conversion factor between detected fluorescent irradiance and fluorometer output (V W–1 m2)k650, k710, kF Attenuation coefficients for downwelling exciting irradiance and upwelling fluorescent irradiance;

kF = k650 + k710 (mm–1)kP Attenuation coefficient for downwelling photosynthetically active irradiance (mm–1)LF, LP Detection limits for fluorescence intensity (V) and photosynthetic rate (mmol O2 dm–3 h–1)m Convexity of the P vs E curve µPAR, µ650 Average cosine (mean cosine of the average path direction of incident photons)P, PB Depth-integrated and depth-integrated biomass-specific rate of gross photosynthesis

(mmol O2 m–2 h–1, mmol O2 mg–1 chl a h–1)P(z) Gross photosynthetic rate at depth z (µmol O2 dm–3 h–1)Pm,Pm

B Depth-integrated and depth-integrated biomass-specific maximum rate of gross photosynthesis (mmol O2 m–2 h–1, mmol O2 mg–1 chl a h–1)

PAR Photosynthetically active radiation, 400 to 700 nm (µmol m–2 s–1)RPAR, R650 Irradiance reflectance (ratio of upward to downward irradiance)z Depth below the sediment surface (mm)zF Depth at which fluorescence intensity measured at the surface equals LF (mm)zP Depth at which gross rate of photosynthesis equals LP (mm)


in biomass on community photosynthesis. To normaliseto cp has the advantage of not attributing the sameweight to all chl a present in some section of the sedi-ment, in spite of the certainty that not all the quantifiedchl a equally contributes to measured photosynthesis(Grant 1986, Pinckney & Zingmark 1993).

Depth-integrated fluorescence and tracing of pro-ductive biomass. As with photosynthesis, the emissionof chl a fluorescence may be described by the productof the rate of light absorption and the quantum effi-ciency of conversion of absorbed light into fluores-cence (Falkowski & Kiefer 1985, Büchel & Wilhelm1993). The fluorescence emitted at any depth z, F(z), isgiven by the product of the exciting irradiance appliedby the fluorometer (in the case of the used PAM fluo-rometer, a narrow-band red light beam of measuringlight peaking at λ = 650 nm) available at that depth, thefraction of this irradiance that is absorbed by the PSIIlight-harvesting complex, and the quantum yield offluorescence, integrated over the range of detectablewavelengths (with the PAM fluorometer, λ > 710 nm):

(4)

where k650 is the attenuation coefficient for down-welling exciting irradiance. The fluorescence signalmeasured non-invasively at the surface of the sedi-ment represents the integration over depth of the up-welling fluorescence emitted at each depth z that isdetectable at the surface:

(5)

where zF is the depth at which the fluorescencedetected at the surface equals LF, the minimumdetectable intensity, k710 is the attenuation coefficient

for the upwelling fluorescent irradiance, and G is aconstant representing the conversion between fluores-cence and instrument output (Fig. 1B) (for simplicity,the attenuation of the exciting light and of the emittedfluorescence in the air, and the reflexion losses at thesediment-air or sediment-water interface are assumedto be constants and are not parameterised).

The rationale for the use of chl a fluorescence totrace variations in productive biomass is based on thefact that any changes in chl a in the photic zone caus-ing variations in cp will concurrently cause proporti-onal variations in the intensity of fluorescence emis-sion, F (Kiefer et al. 1989, Chamberlin et al. 1990). Theapplicability of the method depends on a number offactors that potentially affect the proportionality be-tween fluorescence emission and chl a concentration,namely the photophysiological relationship betweenthe 2 variables and their relative vertical distributionwithin the photic zone of the sediment.

Fluorescence versus chl a concentration: Regardingthe relationship between fluorescence emission andchl a concentration at each depth z (i.e. not depth-inte-grated), the variability in E650, a*650 and φF (Eq. 4) overthe range of natural environmental conditions andmigratory stages has to be sufficiently low to not con-found the effects of variations in chl a on F (k650, R650

and µ650 are determined by the optical characteristicsof the sediment). φF, being directly affected by thefunctioning of the photosynthetic apparatus, is themain cause for large changes in the intensity of fluo-rescence on short time scales (Falkowski & Kiefer1985, Krause & Weis 1991). As the intertidal environ-ment is characterised by high variability in light andtemperature levels (Serôdio & Catarino 1999), bothfactors that primarily affect φF, large variability in fluo-rescence emission due to variations in φF are to beexpected. φF varies between 2 extreme values, theminimum- or dark-level quantum yield of fluores-cence, φF0, and the maximum quantum yield of fluo-rescence, φFm (corresponding to fluorescence intensi-ties F0 and Fm, Fs denoting the intermediate emission ofa light-adapted sample).

Depth-integrated fluorescence: Even under the con-ditions for the existence of a linear relationship be-tween fluorescence emission and chl a concentration ateach depth z, depth-integration may cause F and cP tovary disproportionately. First, due to the attenuation ofupwelling fluorescence, F may detect only a fraction oftotal cP, meaning that deeper zones of the sedimentmay contain chl a and be exposed to irradiance levelshigh enough to drive photosynthesis while not con-tributing to the fluorescence signal measured at thesurface. The upwelling attenuation of fluorescencecauses the proportion between F(z) and c(z) to varywith depth, and the relationship between depth-inte-

F F z zG

zk z= ( )∫ −

0

F710e d

F z E a z c z F z

R zz

k z( ) = ( ) ( ) ( )− ( )

µ ( )−

650 6501

* φ 650

650e 650

48

Fig. 1. (A) Productive biomass, cP, and (B) depth-integratedfluorescence, F (shaded areas), defined by detection limitsLP and LF, which determine zP and zF (Eqs 7 & 8). For clarity,curves correspond to homogeneous chl a profiles, and para-

meters R and µ are not included in equations


grated F and cP to become affected by the shape of thechl a profile over the depth interval defined by zP.Therefore, a condition under which F remains propor-tional to cP is when the chl a profile over the photiczone does not change over the migratory cycles, that is,the migratory movements affect mostly the total chl acontent of the photic zone but not significantly its ver-tical distribution. Second, the relation between F andcP may further vary with E and c due to uncorrelatedvariations in zP (expected to increase with E) andzF (dependent on the non-linear response of φF tochanges in E). As can be deduced for the simplest caseof a uniform chl a profile in the photic zone, variationsin the fraction of total productive biomass detected bydepth-integrated fluorescence, f, given by the ratiobetween cF, the biomass detected by F, and cP (consid-ering that R650/RPAR and µ650/µPAR remain relativelyinvariant and close to 1, and defining kF = k650 + k710),

(6)

may occur through changes in zP and zF, as they aredependent on E and c:

(7)

(8)

However, the effects of varying E or c on f variabilitydepend on the detection limits LF and LP: if these aresufficiently low, then e–kPzP ≈ 0 and e–kFzF ≈ 0 (Eq. 6),and f becomes independent of the chl a profile and ofE, ensuring that F proportionally follows the variationsin cP.

The ability of depth-integrated chl a fluorescence totrace variations in the productive biomass of microphy-tobenthic communities was approached by testing theconditions for the verification of a linear relationshipbetween F and cP: (1) the invariability of the relation-ship between fluorescence emission and chl a concen-tration under varying temperature and light levels, and(2) the invariability of f over variations in E or c, andover variations in the depth profile of chl a in naturalsamples under the range of in situ conditions.

METHODS

Measurement of chl a fluorescence and photosyn-thesis. Chl a fluorescence was measured using a PAMfluorometer (PAM 101 chlorophyll fluorometer, HeinzWalz, Germany) and recorded on a strip chart recorder.Fluorescence emission was induced by weak modu-

lated red light (ca 0.1 µmol m–2 s–1, 650 nm), emitted ata frequency of 1.6 kHz when measuring F0, or 100 kHzwhen measuring other parameters. F0 was measuredapplying the highest possible intensity of exciting beam,not inducing any significant variable fluorescence, andmaximum instrument damping to maximise the signalto noise ratio. F0 readings were considered after signalstabilisation following sample darkening. Actinic andsaturating light was provided by an FL-101 fibre illumi-nator (Schott KL1500 halogen lamp, Schott, Germany).

Photosynthesis was measured using Clark-type oxy-gen microelectrodes (5 to 20 µm tip 737-GC, DiamondGeneral, USA) according to Revsbech & Jørgensen(1983). The microelectrode was positioned verticallywithin the sediment using a micromanipulator (MM33,Diamond General) and was connected to a picoamme-ter (Model 8100 Electrometer, Keithley Instruments,USA). Illumination was provided by the same lightsource and fiberoptics used for measuring chl a fluo-rescence. Irradiance incident at the surface of thesediment was measured by positioning a PAR quantumsensor (LI-192SB, Li-Cor, USA) at the level of sedimentsurface.

F versus c: temperature and light effects. The vari-ability of the relationship between fluorescence inten-sity and chl a concentration was tested by measuringfluorescence emission (F0, Fm, Fs, and Fm’, the maxi-mum fluorescence level emitted from a light-adaptedsample) under the range of in situ temperature andirradiance variability, on species representative of themain groups of microalgae occurring in the estuarinemicrophytobenthos. Although the F0 emission per unitchl a is known to be virtually invariable under a widerange of temperature and light intensity conditions,both in higher plants (Schreiber & Bilger 1987, Krause& Weis 1991) and in microalgae (Ting & Owens 1994,Serôdio et al. 1997), its measurement requires the dark-adaptation of the sample, which may interfere with themicroalgae’s behaviour and lead to artifactual migra-tory responses. Another potential advantage of the useof parameters other than F0 is the enhanced signal tonoise ratio obtained at the higher frequency of measur-ing light used under actinic or saturating light.

The effects of temperature and irradiance on theparameters F0, Fm, Fs, and Fm’ were studied in the dia-tom Phaeodactylum tricornutum (Böhlin) (culture col-lection of the IPIMAR -Instituto Português de Investi-gação Marítima, Lisbon, Portugal), the cyanobacteriumSpirulina maxima (Setch. et Gard.) (UTEX LB2342),and the euglenophyte Euglena granulata (Strain 921,culture collection of the Department of Botany, Uni-versity of Coimbra, Coimbra, Portugal). These weregrown in semi-continuous batch cultures at 20°Cunder 50 µmol m–2 s–1 in a 14 h light :10 h dark cycleas described by Serôdio et al. (1997). Fluorescence of

z

kL

E a c FR

FF

F

650

650

650=

µ

1 1

650

1

–ln

*– –

φ

zk

LEa c P

RP

P

P

PAR

PAR

PAR=

µ

1 11

–ln

*–

–

φ

fcc

c z

c z

kk

z k z

z k z

k z

k z= = =

∫∫

−

−

F

P

0

0

P

F

–

–

FF

PP

F F

P P

e d

e d

1– e1– e

49


microalgal cultures was measured in a Hansatech DW2oxygen and fluorescence chamber (Hansatech, UK),coupled to a magnetic stirrer to ensure homogenisationof the algal suspension. Temperature of the samplewas controlled by a waterbath (Haake NK22, Haake,Germany). Cells were harvested by centrifugation(1000 × g, 5 min) during the exponential growth phase,at room temperature, and were resuspended in freshgrowth medium supplemented with NaHCO3 (10 mMfinal concentration).

Temperature effects on fluorescence emission weretested by measuring each fluorescence parameterat 7 different temperatures, from 5 to 35°C, under200 µmol m–2 s–1. Measurements were made after sta-bilisation of F0 at each temperature (minimum 20 min).Three replicate samples were used at each tempera-ture. Temperature was measured in the water circulat-ing around the sample with a temperature sensor (BT1,Delta-T Devices, UK) inserted in the Hansatech DW2chamber through one of its lateral ports, and connec-ted to a data logger (Delta-T Logger, Delta-T Devices).The values obtained for Fm, Fs and Fm’ were normalisedto F0 measured at growth temperature (20°C).

Light effects on chl a fluorescence parameters weretested by measuring the intensity of each type of fluo-rescence emission under 6 different irradiance levels,ranging from 200 to 2200 µmol m–2 s–1, at 20°C. Threereplicate samples were measured at each irradiance.To simulate the application of the technique on micro-phytobenthos samples under in situ conditions, in whichthe sample darkening must be minimal to avoid poten-tial effects on the migratory rhythm, F0 and Fm weremeasured after a 5 min period of darkness followinglight exposure. To eliminate the variability caused bydifferences in the chl a content among samples, Fs, Fm’and Fm were normalised to the value of F0 measured onthe same sample prior to light exposure.

f variability with E and c. The variability in f result-ing from changes in E and c (not considering changesin the vertical structure of the photic zone) was evalu-ated by calculating f for the range of variation in Eand c found under natural conditions, by measuringthe terms of Eq. (6), which can be rewritten as:

(9)

where P (0) and F (0) are given by considering z = 0 inEqs (2) & (4), respectively. The terms P (0), kP, F (0) andkF were experimentally determined on vertically homo-geneous sediment deposits, prepared by adding micro-algae to sediment from which chl a was previouslyremoved (Serôdio et al. 1997). Sediment without chl awas obtained through sterilisation by heat (120°C, 48 h)of natural samples, collected in tidal flats of the Tagusestuary (see below), and the absence of chl a was con-

firmed spectrophotometrically. Because of the largequantity of microalgae required for all the experiments,the collection of cells from natural samples was im-practicable and microalgae grown in culture were used.The diatom Phaeodactylum tricornutum was selectedfor having dimensions similar to most benthic diatomsand for being non-motile, thus preventing variations inthe chl a content in the photic zone during the mea-surements. Cells were harvested by filtering throughGF/C Whatman filters, gently scraped from the filter,and mixed with sediment and growth medium (naturalseawater enriched with f/2 nutrients) to obtain a finalwater content of ca 50% (w/w), which was similar tonatural samples and prevented the formation of a het-erogeneous vertical distribution of chl a during set-tling. The final cell concentration was set to correspondto the chl a concentration within the range of variationobserved with natural samples.

P (0) and kP were estimated from vertical profiles ofphotosynthesis within the uniform deposits of sedi-ment and microalgae, by fitting to profiles of P(z)/E theequation:

(10)

Photosynthesis was measured at depth intervals of40 µm, under surface irradiance levels ranging from30 up to 2300 µmol m–2 s–1. Three independent mea-surements were made at each depth. kP was estimatedusing the log-linear part of the profiles of photosynthe-sis corresponding to light-limitation, to optimise the fitof Eq. (10).

F (0) and kF were estimated by considering the varia-tion with depth of F0 emitted from a vertically homo-geneous microphytobenthic assemblage, given, fromEqs (4) & (5), by:

(11)

F0 was measured in sediment deposits with differentdepth but with the same chl a concentration, and F (0)and kF were estimated through fitting of Eq. (11) usinga quasi-Newton iterative estimation procedure. Thesediment and microalgae were deposited in a settlingchamber, and chl a fluorescence was measured byplacing the settling chamber, with transparent bottom,directly on top of the PAM fiberoptics. The depth ofeach deposit was estimated from its weight, applying apreviously established calibration curve relating depositthickness and weight. The thickness of the depositsused for constructing the calibration curve was deter-mined by reading on the micromanipulator scale thedifference between the vertical position of the tip of amicroelectrode touching the bottom of the empty set-tling chamber and touching the surface of the sedi-

F z z F

kk z

z k z

( ) = ( )−∫ e d1– e

710F

0

–

F0

ln ln –

P zE

PE

k z( )

= ( )

0P

f

kk

LF

LP

=( )

( )

−P

F

F P10

10

1

– –

50


ment deposit. The position of the electrode tip wasdetermined by observation through a binocular dis-secting microscope. The contribution of backgroundfluorescence emitted by the sediment particles wasestimated by measuring F0 on sediment without chl aand was subtracted from the F0 readings. LF and LP

were estimated by determining the minimum detectablechl a fluorescence intensity and rate of photosynthesisunder the applied instrumentation settings. f was cal-culated for E values varying up to 2000 µmol m–2 s–1,and for c values varying from 25 up to 150 g chl a m–3,values obtained from F0 data measured in natural sam-ples in the Tagus estuary.

f variability in situ. The variability in f in undis-turbed samples under natural conditions was charac-terised by estimating its value for the range of verticalchl a profiles in the photic zone occurring under in situconditions. Considering their importance in determin-ing the value of f, chl a profiles were tested for linear-ity, a condition that ensures proportionality between Fand cP.

Vertical chl a profiles in the photic zone: Under irra-diance levels correspondent to the linear part of the Pversus E curve, the quantum yield of photosynthesis ismaximum and constant (φP = φPm), and, at any depth z,P(z) is expected to be proportional to c(z) (Eq. 1). In thecase of a homogeneous photic zone (depth constantsa*, c, R and µ), P(z) is expected to vary exponentiallywith z, and the uniformity of the vertical distribution ofchl a in the photic zone may be assessed in naturalsamples by testing the log-linearity of the vertical pro-files of P(z)/E (Eq. 10). Vertical profiles of light-limitedphotosynthesis measured for the construction of com-munity P versus E curves on natural samples under insitu conditions (see below) were used to study theshape and variability of the chl a profile over migratorycycles. Eq. (10) was fitted by linear regression to verti-cal profiles of mean P(z) values calculated by averag-ing the rates measured under different limiting irradi-ance levels at each depth.

f estimation: Using the values estimated for kP andkF, f was calculated by applying Eq. (6) and the ratio-nale behind Eq. (10) to vertical profiles of light-limitedphotosynthesis:

(12)

Effects of varying productive biomass on photosyn-thetic light response. The method was applied to char-acterise the variability of cP in situ, and to assess thequantitative effects of short-term changes in produc-tive biomass on the community photosynthetic lightresponse. Depth-integrated fluorescence and commu-

nity-level P versus E curve parameters α [initial slope,mmol O2 m–2 h–1 (µmol m–2 s–1)–1] and Pm (mmol O2 m–2

h–1, maximum photosynthesis) were measured on un-disturbed natural microphytobenthos samples underthe range of conditions found in the estuarine intertidalenvironment, over hourly, fortnightly and seasonaltime scales.

The data thus obtained were also used to further testthe applicability of the method over the range of in situconditions, based on the following rationale: underlight limitation, on the linear part of the P versus Ecurve, φP is constant, and, in the presence of a verti-cally homogeneous photic zone, α may be expected tobe linearly related to cP (from Eq. 2):

(13)

Therefore, the verification of a linear relationship be-tween F and α in natural samples at different timesalong the migratory cycle represents a simultaneousvalidation of the conditions established for using F totrace variations in cP.

Sampling: Sediment samples were collected duringlow tide using plexiglass corers (1.9 cm internal diam-eter) on intertidal mudflats near Pancas salt marsh,located on the south margin of the Tagus estuary, Por-tugal. The sampling site has fine sediment (90% ofparticle sizes < 20 µm), colonised by microphyto-benthic communities typically dominated by diatoms(Brotas & Plante-Cuny 1998), the genera Navicula andGyrosigma being the most frequent at the samplingsite (Serôdio et al. 1997). Sediment cores were collec-ted on 5 to 6 d during each of 3 spring-neap tidalcycles, in July and November 1995, and March 1996.After collection, samples were taken to the laboratory(in Lisbon, about 20 km from the sampling sites) andimmediately placed outside, in an artificial tidal systemthat simulated the immersion during high tide periodsusing estuarine water collected on the day of sedimentsampling (Serôdio et al. 1997). Samples remained inthe tidal system overnight, and all measurements werecarried out in the laboratory, on the day after samplecollection. Samples were carefully moved to avoid dis-turbances in the vertical structure of the sediment.

F0 and P versus E curves: F0 and P versus E curveparameters were measured simultaneously on undis-turbed microphytobenthos assemblages at differenttimes along the migratory cycle, 3 to 5 times d–1,covering the whole range of diurnal low and high tidesituations over the spring-neap tidal cycle. A differentcore was used at each time of the day to avoid effectsof removal on the migratory rhythm. F0 was measuredby positioning the PAM fiberoptics (12 mm diameter)perpendicular to the sediment surface, at a fixed dis-tance of 1 mm, as described by Serôdio et al. (1997).

α φ= =P

Ea c P*PAR P m

fc z

c z

P zE

P zE

k z

zk z

z

k z

k zz

z

=( )

( )=

( )

( )

∑∑

∑

∑

e

e

e

e–

–

–

–F

P

F

P

51


The relative positions of the sample surface, the fiber-optics and the oxygen microelectrode were set with amicromanipulator to which the corer was attached.P versus E curves were constructed by exposing thesame sediment core to 8 different incident irradiances.Under each irradiance, photosynthesis was measuredat depth intervals of 50 µm and then integrated over alldepths. During high tides, photosynthesis was measuredmaintaining a ca 2 mm layer of water above the sedi-ment surface. Because the distribution of photosyntheticbiomass along light gradients may affect the shape ofthe P versus E curve independently of α and Pm (Tera-shima & Saeki 1985, Leverenz 1988, Henley 1993), Pversus E curves were fit to the model of Bannister(1979):

(14)

which includes a shape parameter m (1 ≤ m < ∞) thatrepresents the convexity of the curve. The model wasadjusted using a quasi-Newton estimation procedure,and only the curves for which the model explained>90% of total variance were considered for furtheranalysis. To maximise the temporal resolution of theshort-term variability associated with migratory rhythms,only 2 replicated P versus E curves, each typically tak-ing about 30 min to complete, were constructed at eachpoint in the migratory cycle. Chl a fluorescence wasmeasured immediately before and after the construc-tion of the P versus E curves, and the parameters esti-mated for each curve were compared with the adjoin-ing F0 measurement.

RESULTS

F versus c: temperature and light effects

F0 was the fluorescence parameter least sensitive tovariations in temperature or irradiance (Figs 2 & 3).With the exception of Spirulina maxima, F0 varied lesswith temperature than all the other fluorescence para-meters: 10.4% in Phaeodactylum tricornutum and10.5% in Euglena granulata. Most of the 50.2% varia-tion in F0 observed with S. maxima occurred below10°C; F0 varied by only 12.3% over the temperaturerange of 10 to 35°C. The chl a fluorescence parametersmeasured under actinic light, Fs and Fm’, were gener-ally much more affected by temperature than dark-adapted F0 or Fm, exhibiting complex and diverseresponse patterns in all species, particularly in P. tri-cornutum (Fig. 2). Exposure to high irradiance resultedin most cases in quenching of fluorescence emission,which was particularly evident under actinic light. Fs

and Fm’ were the parameters most affected by the vari-ation in incident irradiance in all species (Fig. 3). Thequenching of fluorescence intensity was also observedin dark-adapted F0 and Fm over the range of irradi-ances tested, resulting in a maximum decrease of45.9%, in P. tricornutum. Nevertheless, F0 was in allcases the least affected parameter by changes in inci-dent irradiance.

f variability with E and c

Both Eqs (10) & (11) described adequately theexperimental data used for estimating the parametersof Eq. (9), thus validating the assumptions made fortheir derivation, namely the vertical homogeneity ofthe sediment deposits and the invariance of kP, kF,R and µ with depth. The attenuation coefficients fordownwelling photosynthetically absorbed irradianceand for fluorescence intensity were estimated as kP =16.85 mm–1 [P (0)/E = 0.155 mmol O2 dm–3 h–1 (µmol

P EP E

P Em m m

( ) =( ) + ( )[ ]

α

α

m

m

1

52

Fig. 2. Variation of chl a fluorescence parameters F0 (D), Fm (J),Fs (S) and Fm’ (H) with temperature (actinic light: 200 µmolm–2 s–1). Points represent the mean of 3 independent mea-surements, normalised to the value of F0 measured at growthtemperature (20°C); error bars represent 1 standard deviation


m–2 s–1)–1; r = 0.934, n = 39] and kF =53.48 mm–1 [F (0) = 8.884 V; r =0.979, n = 7], respectively. The detec-tion limits for the measurement offluorescence intensity and rate ofphotosynthesis were determined tobe LP = 0.07 mmol O2 dm–3 h–1 andLF = 0.002 V (for the instrumentsettings used for measuring all thedata presented). By applying theseestimates to Eqs (7) & (8), zP wascalculated for the range of variationin E: zP increases non-linearly withsurface irradiance, exceeding zF formost of the range of incident irradi-ance (Fig. 4A); under 2000 µmol m–2

s–1, photosynthesis can be measureddown to a depth of zP = 0.5 mm,while the fluorescence measured at

the surface corresponds to a depth interval of zF =0.16 mm (zP/zF = 3.21). By applying Eq. (6), f wasfound to remain virtually constant under most ofthe range of E and c values typically found in situ,approaching the value of kP/kF = 0.32 (Fig. 4B).Largest variations in f take place when both E and cattain low values. For the extreme situation of c = 25 gchl a m–3 (typical value before sunrise), f varies <2%between 40 and 2000 µmol m–2 s–1.

f variability in situ

Vertical profiles of ln[P(z)/E ] were consistently dif-ferent between low and high tide situations. While alinear regression equation could be fitted to 77.8% ofthe profiles measured during low tide, this number wasreduced to 47.6% during high tides. Of the non-linearprofiles found during low tides, 83.3% were measured<30 min before high tide or <1 h before sunset, timesin which the downward migration of microalgae isexpected. Non-linearity of high tide profiles was oftendue to the presence of a production maximum clearlybelow the surface (Fig. 5A).

Changes in the chl a profile under in situ conditionswere found to cause more variability in f than E or c.f variability was lower during daylight exposure peri-ods (CV 17.2%) than during periods of inundation(CV 29.8%). The same pattern was also found for allsampling periods separately. During low tide, f vari-ability was higher near high tide or night; when ex-cluding from the analysis the first hour after or beforeinundation or night, variability (CV) decreased to15.4%, and to 11.4% when the first 2 h are excluded.

53

Fig. 3. Variation of chl a fluorescence parameters F0 (D), Fm (J),Fs (S) and Fm’ (H) with irradiance. F0 and Fm were measuredafter 5 min of dark adaptation following exposure to indicatedirradiance. Points represent the mean of 3 independentmeasurements, normalised to the value of F0 measured beforelight exposure; error bars represent 1 standard deviation

Fig. 4. Calculated variation of (A) depth interval contributing to depth-integrateddark-level chl a fluorescence detectable at the surface, zF, and for measurable pho-tosynthesis, zP, and (B) f, the fraction of productive biomass detected by depth-in-tegrated fluorescence, with surface irradiance, for different values of chl a concen-tration, homogeneously distributed over the photic zone (25, 75 and 150 g chl a m–3)


When considering single low tide periods, maximumvariation in f ranged from 103 to 155%, with a meanvalue of 135%. Intraday variability in f is illustrated inFig. 5B, which shows f remaining relatively constantduring low tide periods, and decreasing most markedlyduring inundation.

Effects of varying productive biomass onphotosynthetic light response

Both α and Pm varied significantly during the courseof daylight low tide periods, reaching maximum valuesduring the middle of exposure periods and minimumvalues when approaching high tide or night. F0 exhi-bited consistently similar hourly patterns of variation,and significant correlations were found both betweenF0 and α and between F0 and Pm for all spring-neaptidal cycles (Figs 6 & 7). As a consequence, highly sig-

nificant correlations were found between α and Pm inall periods (p < 0.001). The short-term variability incommunity-level P versus E curve parameters, paral-leled by F0, is illustrated in Fig. 5B to D. A clear differ-ence was found between low and high tide periods:for low tide, a significant correlation was also foundfor all sampling periods, with F0 explaining from57.5 to 81.8% of the variability observed in α, andfrom 69.4 to 95.2% of the variability in Pm; regardinghigh tide periods, a significant correlation was foundonly between F0 and α in July (p = 0.014, n = 8). Nosignificant correlation was found between F0 and m.No significant differences in the ratio F0/α (expectedto be proportional to f ) were found between linear andnon-linear profiles (2-sided t-test, p > 0.05). This wasverified when considering each sampling period sepa-rately, when pooling the data from the 3 samplingperiods, or when considering low and high tidesseparately.

54

Fig. 5. Example of the effects ofmigratory rhythms on short-termvariability of the photosyntheticlight response of intact microphy-tobenthos communities under insitu conditions during 1 d. Hourlyvariation of (A) vertical profiles ofP(z)/E, (B) F0 (line) and f (bars),(C) community-level P versus Ecurve, and (D) P versus E curveparameters α and Pm. Shaded per-iods represent high tide. Duringthe diurnal low tide, the verticalmigration of the microalgae causesthe variation of the total chl aconcentration in the photic zone,while not significantly affecting theuniformity of its vertical distri-bution (A). In (A), the probabilityassociated with the fitting of a lin-ear regression equation to the ver-tical profile of ln[P (z)/E] is indi-cated. Downward migration duringhigh tide causes a marked changein the shape of the vertical profileof P(z)/E, now deeper and maxi-mal below the surface; in theshort period between ebb andsunset microalgae migrate down-ward, and a uniform distributionis found only in 1 of the 2 profiles(A). The hourly variation in pro-ductive biomass is closely followedby F0 as f remains relatively con-stant, mainly during low tide (B),and leads to substantial changesin the community photosyntheticlight response P versus E curve(C,D). Both α and Pm showedhighly significant correlation with

F0 (p < 0.001)

f


DISCUSSION

Tracing changes in productive biomass

F versus c

Effects of light and temperature on fluorescenceemission per unit of chl a were minimal when fluores-cence was measured at its minimum or dark level.F0 was the least variable of the measured fluores-cence parameters under the tested range of variationof light and temperature, in spite of the considerablediversity in the response pattern of the differentspecies, being confirmed as the most adequate fluo-rescence parameter to trace variations in chl a con-centration. This variability in response patterns isassociated with fundamental differences in the com-position and organisation of the photosynthetic mem-branes and other characteristics affecting the func-

tioning of the photosynthetic apparatus, the discus-sion of which is beyond the scope of this study.F0 was also the least variable parameter when com-paring between species, which is important for main-taining a low variability in the quantum yield of fluo-rescence of the whole community, where short-termvariations in the taxonomic composition of the micro-phytobenthos exposed to measuring light may occurdue to differential cell migration. Fluorescence para-meters measured under actinic light, Fs and Fm’, werestrongly affected by changes in temperature and irra-diance and are thus clearly inadequate for tracingvariations of biomass in microphytobenthos samples,particularly when considering the high variability intemperature and light that characterises the estuarineintertidal environment (Serôdio & Catarino 1999). Thealternative approach of measuring Fm does not avoidan identically large variability with temperature andirradiance.

55

Fig. 6. Relationship between dark-level chl a fluorescence, F0,and the initial slope of the depth-integrated P versus E curve,α, in undisturbed microphytobenthos assemblages over 3spring-neap tidal cycles (July, November, March) (S: low tide;

D: high tide)

Fig. 7. Relationship between dark-level chl a fluorescence,F0, and depth-integrated maximum photosynthesis, Pm, inundisturbed microphytobenthos assemblages over 3 spring-neap tidal cycles (July, November, March) (S: low tide;

D: high tide)


Significant variations in F0 emission per unit chl amay still take place, caused either by changes in a*650

or in φF (Eq. 4). However, such variability is restrictedto extreme situations leading to the inactivation of func-tional reaction centres, such as extreme temperatures,photoinhibiting irradiances or anaerobiosis (Büchel &Wilhelm 1993, Ting & Owens 1993, 1994). In particular,changes in a*650 may occur as a result of relatively slowphotoadaptative processes affecting pigmentation andchloroplast morphology (e.g. 10 h for a 10% variationin a*PAR; Berner et al. 1989), or, on shorter time scales,as a result of non-photochemical quenching and theestablishment of a trans-thylakoidal proton gradient inthe dark caused by chlororespiration (Caron et al.1987, Ting & Owens 1993, Schreiber et al. 1995).Such processes lead to significant decreases in F0 perunit chl a in natural marine phytoplankton exposed tohigh irradiance levels, associated with decreases ina*PAR (Falkowski et al. 1994). However, it may beexpected that these processes will not confound thedetection of short-term variations in photosyntheticbiomass of intertidal microphytobenthos under naturalconditions, as the effects on F0 emission are muchsmaller than the typical variations in chl a concentra-tion in the photic zone caused by migratory rhythms(Serôdio et al. 1997).

However, uncorrelated changes in F0 and the chl aactually intervening in measurable photosynthesis,which constitutes the community productive biomass,may occur due to changes in a*PAR (Eq. 7), followingvariations in the spectrum of available solar light withtime of day, cloud cover, and water or sediment depth.Because PAM fluorometry is based on the inductionof fluorescence emission through the application of amonochromatic light, F0 emission may not reflect vari-ations in the spectral composition of the light in use bythe photosynthetic apparatus, which has been the basisfor the application of chl a fluorescence techniques tothe quantification of the photosynthetic light absorp-tion by marine phytoplankton (Sosik & Mitchell 1995)and the estimation of its biomass or primary production(Kiefer et al. 1989, Chamberlin et al. 1990). However,solar-induced (‘passive’) fluorescence, although reflect-ing the changes in the spectral characteristics of ab-sorbed light by resulting from the absorption of the fullsolar light spectrum, has the disadvantage of beingaffected by the large short-term variability in φF asso-ciated to the response of the photosynthetic apparatusto changes in temperature or incident irradiance.

Depth-integration: F versus cP

When considering the depth-integration of fluores-cence emission, the results indicate that only a fraction

of total cP is detectable by F0, as zF remains smallerthan zP under most of the range of E and c. The valueestimated for zF (0.160 mm) is quite similar to thatobtained by Kromkamp et al. (1998), 0.175 mm, in spiteof being determined by a different experimental ap-proach. The value estimated for kP, 16.85 mm–1, iswithin the range of previous estimates, both for similarsediments of the Tagus estuary (17.29 mm–1, Serôdio etal. 1997) and as for the muddy sediments of other estu-aries (32.6 mm–1, Colijn 1982; 11.3 to 12.7 mm–1, Baillie1987). The higher value of kF (kF/2 = 26.74 mm–1, theaverage value correspondent to only downwellingirradiance) relative to kP is to be expected from thehighest absorption in the red zone of the spectrum dueto the presence of chl a and other photosynthetic pig-ments.

However, the detectable fraction of cP can be ex-pected to be sufficiently constant to allow for the useof F0 to proportionally follow variations in cP. Thedetection limit for F0 was found to be sufficiently lowto ensure that the variability resulting solely fromchanges in E or c may be considered as negligible or tooccur during well-defined periods, when both E and care low. Because in situ, low values of E and c occursimultaneously only shortly after dawn and beforenightfall, variability in f may be expected to be mini-mal during most of daylight hours. Relevant variabilityin f may thus be assumed to result mostly from short-term changes in the chl a profile following migratorymovements of microalgae. However, the analysis ofthe P(z)/E profiles indicated that, at least during diur-nal low tide, the vertical distribution of chl a in thephotic zone appears not to depart significantly fromuniformity. Furthermore, the variability in f caused bychanges in the vertical chl a profile under in situ con-ditions, quantified by calculating f for a wide range ofchl a profiles, was found to be much smaller than thevariability caused by fluctuations in total chl a presentin the photic zone during single migratory cycles: insamples under in situ conditions, changes in F0 duringsingle low tide exposure periods typically range from400 to 900% (Serôdio et al. 1997), while the maximumdifference found between f values calculated for thesame day averaged 135%.

The finding of a relative uniformity in chl a profilesduring vertical migrations may be explained by thesmall depth interval of the photic zone (maximum500 µm, under 2000 µmol m–2 s–1) in comparison withthe depth interval where migrations cause variations inthe chl a concentration (down to 3 mm; Pinckney et al.1994). Because the maximum photic depth is about 10times the length of the smaller diatoms and 2 to 3 timesthe length of the larger cells, large variations in F0 orproduction may be expected to result more from cellsentering or leaving the photic zone than from changes

56


in their positioning within this zone. Moreover, becausethe fluorescence signal measured by the PAM fiberop-tics integrates over an area of ca 1.13 cm2 (104 timeslarger than the cross-section of an hypothetical spheri-cal cell of 50 µm in diameter), it may be expected toattenuate the effect of local extreme profiles and ap-proximate a uniform chl a profile.

While it may be assumed that f remains relativelyconstant during low tide, the results indicate thatimportant variations can occur near the transition toperiods of high tide or night. Such changes are pro-bably associated with the downward movement ofmicroalgae during the periods shortly preceding andfollowing high tide or night, causing the F0 signal tocorrespond to a smaller fraction of cP. In fact, f wasfound to be, on average, higher during low tide (f =0.41) than during high tide (f = 0.21). However, thevariability in the chl a profile appears not to be suffi-ciently high to substantially affect the relationshipbetween F0 and cP, as no significant differences werefound between the value of the ratio F0/α calculated forlinear and non-linear profiles.

The significant correlation found between F0 and theP versus E curve parameter α over the range of in situvariability further confirms the ability to trace varia-tions in cP using F0. Moreover, other factors besides therelationship between F0 and cP are likely to confoundthe verification of a correlation between F0 and α. First,the large difference between the scales of spatial (hor-izontal) variability resolved by the fluorometer fiberop-tics and the microelectrode tip may have led to much ofthe observed data variability. Second, the variabilityin a*650/a*PAR or φP affecting the physiological lightresponse of the community may have induced varia-tions in α uncorrelated to F0.

Normalisation of community-level photosyntheticrates

One of the main interests in defining and quantifyingthe productive biomass of microphytobenthic assem-blages lies in the possibility to normalise community-level photosynthetic rates, as a way to characterise thephotophysiological response of the community inde-pendently of migratory movements and other factorsaffecting the photosynthetic biomass in the photiczone. The common practice of estimating microphyto-benthic productive biomass through quantification ofthe chl a content of the top layers of sediment resultsfrom the direct application of the methodology devel-oped and routinely used for phytoplankton. However,in contrast with the water column, the quantification ofthe vertical distribution of the chl a within the photiczone of the sediment presents considerable difficulties,

and chl a is usually quantified for a single depth inter-val. Estimates of PB obtained in this way are largelyarbitrary, since they depend on the sectioning depthused for chl a extraction (MacIntyre et al. 1996).Moreover, because the depth interval used for chl aextraction, 2 mm (Rasmussen et al. 1983, Pinckney& Zingmark 1991), 5 mm (Sundbäck 1986, Blanchard& Montagna 1992) or 10 mm (Grant 1986, Brotas &Catarino 1995), is much larger than the photic zone,most of the quantified chl a originates from layersnever illuminated and not contributing to the mea-sured photosynthesis. As a consequence, biomass-spe-cific photosynthetic rates may be expected to belargely underestimated, which can be confirmed bythe fact that published values of αB estimated by thisprocedure, 1.9 to 3.1 × 10–3 (Pinckney & Zingmark1993) or 2.1 × 10–3 mg C mg–1 chl a h–1 (µmol m–2 s–1)–1

(Brotas & Catarino 1995), are typically 1 to 2 ordersof magnitude lower than the values estimated usingresuspended microphytobenthic cells, for which thequantified chl a approximates the amount actuallycontributing to the measured photosynthesis: 0.1 to1.6 × 10–1 (Blanchard & Montagna 1992), 1.3 to 2.8 × 10–2

(Rasmussen et al. 1983) and 3.5 to 8.0 × 10–2 mg C mg–1

chl a h–1 (µmol m–2 s–1)–1 (MacIntyre & Cullen 1995). One particularly appealing aspect of using active

chl a fluorometry to quantify microphytobenthic pro-ductive biomass lies in the fact that it uses a light beamas a measuring probe for the chl a present in the photiczone. Because the measuring light parallels the behav-iour of natural light within the sediment, the measuredfluorescent signal may be closely related to the inte-grated vertical distribution of chl a, weighted, at eachdepth, by its contribution to community photosynthe-sis. In principle, any chl a fluorometer applying a lowintensity measuring light beam conducted through afiberoptics bundle can be used to non-destructivelyquantify the productive biomass of intact microphyto-benthos samples. However, modulated fluorometersare particularly suitable for this purpose, as the inten-sity of the measuring light may be reduced to a levelthat is low enough to allow accurate measurement ofthe F0 level.

Considering the results obtained on the uniformity ofthe chl a vertical profile, on the range of variation of zP,and on the value of kP, the productive biomass of amicrophytobenthic assemblage may be approximatedby simplifying Eq. (3):

(15)

where c may be estimated from a calibration curveestablished between fluorescence intensity, F (V), andthe sediment chl a content, c (mg chl a m–3), of verti-cally homogeneous sediment deposits with varying

c

ck

ck

k zp

p

–

pe P P≈ ( ) ≈

57


microalgal content (Serôdio et al. 1997). The value ofkP may be estimated from vertical profiles of light-lim-ited photosynthesis, obtained in the measuring of P(Eq. 10). Alternatively, because kP is mainly deter-mined by sediment characteristics, and therefore rela-tively invariant in comparison to c, existing estimatesfor the same sediment type may be used.

The error associated with the quantification of cP

applying Eq. (15), resulting from assuming a homoge-nous chl a vertical distribution, can be assessed bycomparing the values of cP obtained by applyingEqs (3) & (15) to the range of chl a profiles found in situ[using, as for the calculation of f (Eq. 12), profiles ofP(z)/E as proxys for c(z), and, for Eq. 15, c = cFkF]. Onaverage, cP resulted in overestimation by a factor of1.30 during low tide, and underestimation by a factorof 0.66 during high tide. Even considering individualprofiles, cP estimates were always within the order ofmagnitude of expected values, with maximum errorattaining 78%. This level of error represents a signifi-cant improvement regarding the precision in the quan-tification of the biomass to be used in the normalisationof photosynthetic rates. The alternative way of calcu-lating areal chl a concentration from the chl a presentin the upper layers of the sediment yields PB estimatesthat are underestimated by at least 2 orders of magni-tude. This can be illustrated by a numerical exampleusing typical values for the variables of Eq. (3), for thecase of a uniform chl a profile (c = 105 mg chl a m–3,the order of magnitude of the values obtained by mea-suring F0, kP = 16.85 mm–1, and maximum zP = 0.50 mm,correspondent to an incident irradiance level of2000 µmol m–2 s–1): while the productive biomass iscP = 5.9 mg chl a m–2, areal chl a concentration calcu-lated from the top 2 mm section reaches 2.0 × 102 mgchl a m–2. This difference may increase by 1 order ofmagnitude in the case of a sectioning depth of 10 mmand lower surface irradiance levels.

Productive biomass and community-level efficiencyof light absorption

In the characterisation of the photophysiologicalresponse of photosynthetic organisms through the con-struction of P versus E curves, it is usually implicit thatthe measured irradiance, generally incident irradi-ance, either approximates (in the case of diluted phyto-plankton samples or microalgal cultures) or is propor-tional (in the case of leaves of macrophytes or higherplants) to the actual irradiance absorbed for photosyn-thesis. The P versus E curve parameters estimated inthis way can be assumed to reflect only variations inthe efficiency of conversion of absorbed light energy(quantum yield of photosynthesis), and be used to

characterise and compare the photophysiological re-sponse of identical samples over time or locations. Inthe case of undisturbed microphytobenthic assem-blages, P versus E curves are usually based on the irra-diance incident on the surface of the sediment (Grant1986, Pinckney & Zingmark 1993, Brotas & Catarino1995). The information thus gathered, while useful forthe characterisation of the response of the communityas a whole, describes an ‘apparent’ light response ofthe community, which is only partially determined bythe ‘inherent’ photophysiological characteristics of themicroalgal population. The characterisation of the pho-tophysiological properties of the community, determinedby the quantum yield of photosynthesis, requires thequantification not only of the chl a concentration, givenby cP, but also of the absorption cross-section of themicroalgae present in the photic zone, a*PAR (Eq. 1).The joint vertical distribution of a*PAR and c along thelight gradient within the photic zone defines the com-munity-level efficiency of light absorption, ε, a quan-tity representing the fraction of the light incident at thesurface that is absorbed for photosynthesis by thewhole microphytobenthic community:

(16)

the quantification of which allows one to fully distin-guish changes in light-harvesting efficiency (changesin a*PAR or c) from variations in energy conversion effi-ciency (changes in φP).

The viability of using F0 to trace variations in εdepends on the variability of the ratio a*650/a*PAR,which, under natural conditions, may be affected bychanges in a*PAR following variations in the spectrumof available solar light not detected in fluorescenceemission induced by monochromatic light. However,the correlations found between F0 (depending on a*650)and P versus E curve parameters (depending on a*PAR)indicate that, at least within each fortnightly cycle, thevariability in a*650/a*PAR is much lower than the vari-ability in cP. This suggests that the measurement of cP

alone may allow tracing of most of the variability in ε,and distinguishing variations in light-harvesting effi-ciency from variations in energy conversion efficiency,as the cause for the short-term variability in micro-phytobenthic photosynthesis.

One application of these results is the possibility touse estimates of cP to calculate absolute estimates of εand φP for undisturbed communities, providing thata*PAR is known. This can be illustrated considering thenumerical example presented above (cp = 5.9 mg chl am–2): using a typical value of a*PAR = 0.0075 m2 mg–1

chl a (Hartig et al. 1998), light absorption efficiencyreaches ε = 4.5%, varying between ca 2 and 14%for the range of variation of F0 observed in situ (ε =

ε = ( ) ( ) ( )

µ ( )∫ −a z c zR z

zz

zk z

0PAR

PAR

PAR

PP* e d

58


a*PARcP; Eq. 16 for a uniform photic zone); communityφPm can be calculated by additionally considering thecorresponding value of α [Eq. 13; α = 9.1 × 10–3 mmolO2 m–2 h–1 (µmol m–2 s–1)–1, estimated from the linearrelationship with F0], which results in φPm = 0.057 molO2 mol quanta–1, a value which is within the range ofvalues measured for phytoplankton under identicalnutrient concentrations (ca 10 µM NO3

–; Babin et al.1996). It may be also noted that, by normalizing α to cP,the resulting estimates of αB are within the order ofmagnitude of estimates calculated from resuspendedmicrophytobenthos: for the same example, αB = 1.8 ×10–2 mg C mg–1 chl a h–1 (µmol m–2 s–1)–1.

Microphytobenthic primary production periodicity:migratory rhythms versus physiology

Aquatic primary producers typically control theirphotosynthetic light response through changes in theefficiency of conversion of absorbed light. Light ab-sorption is largely determined by light availability andcontrolled by external processes affecting either theposition of the photosynthetic cell in the environmentallight gradient (e.g. mixing of the water column, in thecase of phytoplankton) or the optical properties of themedium (e.g. turbidity of the water column, in the caseof subtidal phytobenthos). In the case of microphyto-benthos growing on fine intertidal sediments, thecomparable dimensions of the photic zone of the sedi-ment and of the motile microalgae enables the move-ment of individual cells over the vertical light gradientto result in substantial variations in the efficiency oflight absorption at the community level. In this way,microphytobenthos communities have the ability tocontrol their photosynthetic light response not onlythrough changes in the photosynthetic yield at thelevel of the photosynthetic apparatus of individualcells, but also by actively regulating the fraction ofavailable light that is absorbed for photosynthesis.

The results of this study indicate that the effects ofmigratory rhythms in regulating the community photo-synthetic response are actually more important thanthose resulting from photophysiological adaptation,therefore identifying microalgal migratory behaviouras the main factor controlling short-term variability inthe community light utilisation efficiency. The findingof significant correlations between P versus E curveparameters and F0 in all sampling periods indicatesthat, at least on hourly to fortnightly time scales, a sig-nificant proportion of the variability in the photosyn-thetic light response of the studied communities isexplained solely by variations in the efficiency of lightabsorption, and not by variations in the efficiency oflight energy conversion. Although short-term varia-

tions in photosynthetic efficiency of microphytoben-thos have been reported (Blanchard & Cariou-Le Gall1994, Kromkamp et al. 1998), the assessment of theirrelative effect on the variability of community photo-synthesis has been limited by the impossibility ofquantifying the effects of changes in light absorptionefficiency associated with variations in productive bio-mass. Nevertheless, the existing evidence seems toconfirm the findings of the present study, as the vari-ability in biomass-independent photosynthetic lightresponse observed in other studies is much lower thanthe variability in productive biomass reported here.While hourly variations in productive biomass mayreach values near to 1 order of magnitude, maximumreported values for hourly variability in αB or Pm

B, mea-sured in resuspended samples, attain only ca 81.8and 90.6%, respectively (Blanchard & Cariou-Le Gall1994). Moreover, much lower variability was measuredin the photosynthetic efficiency of undisturbed micro-phytobenthic communities under in situ conditions(Kromkamp et al. 1998). The significant differencefound among the slopes of the regression lines, both forthe α versus F0 and for the Pm versus F0 plots, in spite ofresulting from a small number of extreme observations(November, Figs 6 & 7), suggests the occurrence of va-riations in physiological parameters that can alter thequantitative relationship between F0 and cP. Seasonalvariations in φF0 may result from changes in thylakoidstacking and membrane transparency (‘package effect’)or pigmentation (Berner et al. 1989), nutrient depletion(Geider et al. 1993), or be associated with changes intaxonomic composition, which could affect PSII:PSIstoichiometry or relative absorption cross section(Caron et al. 1987, Büchel & Wilhelm 1993).

It has been suggested that motile benthic microalgaeuse their migratory ability to actively search for opti-mal light levels for photosynthesis within the photiczone, thus maximising light absorption while avoidingexposure to photoinhibiting light levels (Admiraal1984). It may be hypothised that such behaviour repre-sents a rapid, flexible and energetically cheap way tooptimise photosynthesis, reducing the need to undergophotophysiological adaptations at the level of the pho-tosynthetic apparatus, and therefore explaining thelow variability observed in the photophysiological lightresponse of the microphytobenthos.

This study showed that the measurement of F0 onundisturbed microphytobenthic assemblages allowsquantitative assessment of the effect of the migratoryrhythms on the temporal variability of microphytoben-thic primary productivity, and distinction of variationsin light-harvesting efficiency from variations in energyconversion efficiency as the cause for the short-termvariability in community-level photosynthesis. The pre-sented results further highlight the ecological impor-

59


tance of the definition and quantification of the pro-ductive biomass of intertidal microphytobenthic com-munities. For the studied microphytobenthos of theTagus estuary, the above was seen to represent themain source of short-term variability in the communityphotosynthesis and, consequently, in the first rate-limiting step of estuarine benthic primary productivity.The application of the technique to other estuarine eco-systems or habitats would assess the extent to whichthe periodicity often observed in intertidal microphyto-benthic primary production results from migratoryrhythmicity.

Acknowledgements. The authors thank Dr Margarida Ramosfor help with oxygen microelectrode instrumentation, andPaulo Cartaxana, Teresa Cabrita and Ana Amorim-Ferreirafor help with the field and laboratory work. Dr Lília Santoskindly provided the Euglena granulata culture. This workwas supported by Junta Nacional de Investigação Científicae Tecnológica, Grants PFMRH 359/92 and PRAXIS XXIBD/5045/95 to João Serôdio. We thank Drs Ulrich Schreiberand Vanda Brotas, and 4 anonymous reviewers for criticalcomments on the manuscript. This work is a contribution tothe European Union ELOISE Programme (ELOISE No. 102)and was carried out under the framework of the Marine andScience and Technology Programme (MAST III) under theNICE Project Contract MAS3-CT96-0048, and EUROSSAMProject Contract ENV4-CT97-0436.

LITERATURE CITED

Admiraal W (1984) The ecology of estuarine-inhabiting dia-toms. Prog Phycol Res 3:269–322

Babin M, Morel A, Claustre H, Bricaud A, Kolber Z, FalkowskiPG (1996) Nitrogen- and irradiance-dependent variationsof the maximum quantum yield of carbon fixation in eutro-phic, mesotrophic and oligotrophic marine systems. Deep-Sea Res 43:1241–1272

Baillie P (1987) Diatom size distribution and community stra-tification in estuarine intertidal sediments. Estuar CoastShelf Sci 25:193–209

Bannister TT (1979) Quantitative description of steady state,nutrient-saturated algal growth, including adaptation.Limnol Oceanogr 24:76–96

Berner T, Dubinsky Z, Wyman K, Falkowski PG (1989) Pho-toadaptation and the ‘package’ effect in Dunaliella terti-olecta (Chlorophyceae). J Phycol 25:70–78

Blanchard GF, Cariou-Le Gall V (1994) Photosynthetic char-acteristics of microphytobenthos in Marennes-Oléron Bay,France: preliminary results. J Exp Mar Biol Ecol 182:1–14

Blanchard GF, Montagna PA (1992) Photosynthetic responseof natural assemblages of marine benthic microalgae toshort- and long-term variations of incident irradiance inBaffin Bay, Texas. J Phycol 28:7–14

Brotas V, Catarino F (1995) Microphytobenthos primary pro-duction of the Tagus estuary (Portugal). Neth J Aquat Res29:333–339

Brotas V, Plante-Cuny MR (1998) Spatial and temporal pat-terns of microphytobenthic taxa of estuarine tidal flats inthe Tagus Estuary (Portugal) using pigment analysis byHPLC. Mar Ecol Prog Ser 171:43–57

Büchel C, Wilhelm C (1993) In vivo analysis of slow chlorophyll

fluorescence induction kinetics in algae: progress, problemsand perspectives. Photochem Photobiol 58:137–148

Caron L, Berkaloff C, Duval JC, Jupin H (1987) Chlorophyllfluorescence transients from the diatom Phaeodactylumtricornutum: relative rates of cyclic phosphorylation andchlororespiration. Photosynth Res 11:131–139

Chamberlin WS, Booth CR, Kiefer DA, Morrow JH, MurphyRC (1990) Evidence for a simple relationship betweennatural fluorescence, photosynthesis and chlorophyll inthe sea. Deep-Sea Res 37:951–973

Colijn F (1982) Light absorption in the waters of the Ems-Dol-lard estuary and its consequences for the growth of phy-toplankton and microphytobenthos. Neth J Sea Res 15:196–216

Dubinsky Z, Berman T, Schanz F (1984) Field experiments forin situ measurement of photosynthetic efficiency and quan-tum yield. J Plankton Res 6:339–349

Falkowski P, Kiefer DA (1985) Chlorophyll a fluorescence inphytoplankton: relationship to photosynthesis and biomass.J Plankton Res 7:715–731

Falkowski PG, Green R, Kolber Z (1994) Light utilization andphotoinhibition of photosynthesis in marine phytoplank-ton. In: Baker NR, Bowes J (eds) Photoinhibition of photo-synthesis: from molecular mechanisms to the field. BIOSScientific, Oxford, p 407–432

Geider RJ, La Roche J, Greene RM, Olaizola M (1993)Response of the photosynthetic apparatus of Phaeodacty-lum tricornutum (Bacillariophyceae) to nitrate, phosphate,or iron starvation. J Phycol 29:755–766

Grant J (1986) Sensitivity of benthic community respirationand primary production to changes in temperature andlight. Mar Biol 90:299–306

Hartig P, Wolfstein K, Lippemeier S, Colijn F (1998) Photosyn-thetic activity of natural microphytobenthos populationsmeasured by fluorescence (PAM) and 14C-tracer methods:a comparison. Mar Ecol Prog Ser 166:53–62

Henley WJ (1993) Measurement and interpretation of photo-synthetic light-response curves in algae in the context ofphotoinhibition and diel changes. J Phycol 29:729–739

Kiefer DA, Chamberlin WS, Booth CR (1989) Natural fluores-cence of chlorophyll a: relationship to photosynthesis andchlorophyll concentration in the western South Pacificgyre. Limnol Oceanogr 34:868–881

Kolber Z, Falkowski PG (1993) Use of active fluorescence toestimate phytoplankton photosynthesis in situ. LimnolOceanogr 38:1646–1665

Krause GH, Weis E (1991) Chlorophyll fluorescence and pho-tosynthesis: the basics. Annu Rev Plant Physiol Plant MolBiol 42:313–349

Kromkamp J, Barranguet C, Peene J (1998) Determination ofmicrophytobenthos PSII quantum efficiency and photo-synthetic activity by means of variable chlorophyll fluores-cence. Mar Ecol Prog Ser 162:45–55

Kühl M, Jørgensen BB (1994) The light field of microbenthiccommunities: radiance distribution and microscale opticsof sandy coastal sediments. Limnol Oceanogr 39:1368–1398

Leverenz JW (1988) The effects of illumination sequence, CO2

concentration, temperature and acclimation on the con-vexity of the photosynthetic light response curve. PhysiolPlant 74:332–341

MacIntyre HL, Cullen JJ (1995) Fine-scale vertical resolutionof chlorophyll and photosynthetic parameters in shallow-water benthos. Mar Ecol Prog Ser 122:227–237

MacIntyre HL, Geider RJ, Miller DC (1996) Microphytoben-thos: the ecological role of the ‘secret garden’ of unve-getated, shallow-water marine habitats. I. Distribution,abundance and primary production. Estuaries 19:186–201

60


Morel A (1978) Available, usable, and stored radiant energy inrelation to marine photosynthesis. Deep-Sea Res 25:673–688

Paterson DM (1986) The migratory behaviour of diatom assem-blages in a laboratory tidal micro-ecosystem examinedby low temperature scanning electron microscopy. DiatomRes 1:227–239

Pinckney J, Zingmark RG (1991) Effects of tidal stage and sunangles on intertidal benthic microalgal productivity. MarEcol Prog Ser 76:81–89

Pinckney J, Zingmark RG (1993) Photophysiological responsesof intertidal benthic microalgal communities to in situlight environments: methodological considerations. Lim-nol Oceanogr 38:1373–1383

Pinckney J, Piceno Y, Lovell CR (1994) Short-term changes inthe vertical distribution of benthic microalgal biomass inintertidal muddy sediments. Diatom Res 9:33–42

Rasmussen MB, Henriksen K, Jensen A (1983) Possible causesof temporal fluctuations in primary production of theDanish Wadden Sea. Mar Biol 73:109–114

Revsbech NP, Jørgensen BB (1983) Photosynthesis of benthicmicroflora measured with high spatial resolution by theoxygen microprofile method: capabilities and limitationsof the method. Limnol Oceanogr 28:749–756

Round FE, Palmer JD (1966) Persistent, vertical-migrationrhythms in benthic microflora. II. Field and laboratorystudies on diatoms from the banks of the river Avon. J MarBiol Assoc UK 46:191–214

Schreiber U, Bilger W (1987) Rapid assessment of stress effectson plant leaves by chlorophyll fluorescence measurements.In: Tenhunen JD, Catarino FM, Lange OL, Oechel WD(eds) Plant response to stress, Springer-Verlag, Heidel-berg, p 27–53

Schreiber U, Schliwa U, Bilger W (1986) Continuous record-ing of photochemical and non-photochemical chlorophyllfluorescence quenching with new type of modulation flu-orometer. Photosynth Res 10:51–62

Schreiber U, Endo T, Mi HL, Asada K (1995) Quenching analy-sis of chlorophyll fluorescence by saturation pulse method:particular aspects relating to the study of eukaryotic algaeand cyanobacteria. Plant Cell Physiol 36:873–882

Serôdio J, Catarino F (1999) Fortnightly light and temperaturevariability in estuarine intertidal sediments and implica-tions for microphytobenthos primary productivity. AquatEcol 33:235–241

Serôdio J, Marques da Silva J, Catarino F (1997) Nondestruc-tive tracing of migratory rhythms of intertidal benthic mi-croalgae using in vivo chlorophyll a fluorescence. J Phycol33:542–553

Sosik HM, Mitchell BG (1995) Light absorption by phyto-plankton, photosynthesis pigments and detritus in theCalifornia Current System. Deep-Sea Res 42:1717–1748

Sundbäck K (1986) What are the benthic microalgae doing onthe bottom of Laholm Bay? Ophelia 4:273–286

Terashima I, Saeki T (1985) A new model for leaf photosyn-thesis incorporating the gradients of light environmentand of photosynthetic properties of chloroplasts within aleaf. Ann Bot 56:489–499

Ting CS, Owens TG (1993) Photochemical and nonphoto-chemical fluorescence quenching processes in the diatomPhaeodactylum tricornutum. Plant Physiol (Rockv) 101:1323–1330

Ting CS, Owens TG (1994) The effect of excessive irradianceon photosynthesis in the marine diatom Phaeodactylumtricornutum. Plant Physiol (Rockv) 106:763–770

61

Editorial responsibility: Otto Kinne (Editor), Oldendorf/Luhe, Germany

Submitted: February 2, 2000; Accepted: October 31, 2000Proofs received from author(s): July 16, 2001

use of in vivo chlorophyll a fluorescence to quantify ... · chl a per square metre and enables the...

Documents