ARTICLE IN PRESS

Continental Shelf Research 28 (2008) 1363– 1375

Contents lists available at ScienceDirect

Continental Shelf Research

0278-43

doi:10.1

� Corr

E-m

journal homepage: www.elsevier.com/locate/csr

Photosynthetic activity, photoprotection and photoinhibition in intertidalmicrophytobenthos as studied in situ using variable chlorophyll fluorescence

Joao Serodio �, Sonia Vieira, Sonia Cruz

Departamento de Biologia, CESAM—Centro de Estudos do Ambiente e do Mar, Universidade de Aveiro, Campus de Santiago, 3810-193 Aveiro, Portugal

a r t i c l e i n f o

Article history:

Received 20 December 2007

Received in revised form

13 March 2008

Accepted 22 March 2008Available online 28 March 2008

Keywords:

Photosynthesis

Tidal flats

Biofilms

Diatoms

Chlorophyll fluorescence

Photoprotection

Ria de Aveiro

43/$ - see front matter & 2008 Elsevier Ltd. A

016/j.csr.2008.03.019

esponding author. Tel.: +351 234370787; fax:

ail address: jserodio@ua.pt (J. Serodio).

a b s t r a c t

The photosynthetic activity of microphytobenthos biofilms was studied in situ on an intertidal mudflat

of the Ria de Aveiro, Portugal. Time series of physical variables characterizing the microenvironment at

the sediment photic zone (incident solar irradiance, temperature, salinity), photophysiological

parameters and productive biomass of undisturbed microalgal assemblages were measured during

daytime low-tide periods along one spring–neap tidal cycle, with the objective of (1) characterizing the

short-term variability in photosynthetic activity in situ, (2) relating it with the changing environmental

conditions and (3) with the operation of physiologically (xanthophyll cycle) and behaviorally (vertical

migration) based photoprotective processes, and (4) assessing the occurrence of photoinhibition. Pulse

Amplitude Modulated (PAM) fluorometry was applied to measure photosynthetic activity (the effective

and maximum quantum yield of photosystem II, DF/Fm0 and Fv/Fm; the photosynthesis index EFY; rapid

light-response curves (RLC)), the photoprotective operation of the xanthophyll cycle and photoinhibi-

tion (non-photochemical quenching, NPQ; quantum efficiency of open RCs, Fv0/Fm

0), and vertical

migration (productive biomass, Fo). The photosynthetic activity was found to be strongly affected by the

cumulative light dose received during the morning low-tide periods. The fluorescence indices

DF/Fm0 , EFY, Fv

0/Fm0 and RLC parameters were more depressed under high irradiances when clear sky

was present during the morning low tide than when foggy conditions reduced the light dose received

during a comparable period. Productive biomass exhibited maximum values in the first hours of the

morning, followed by a steep decrease when irradiance reached moderate levels, due to the downward

migration of the microalgae. This photophobic migratory response appeared to display a photo-

protective role, allowing DF/Fm0 to remain near optimum values until irradiance reached values as high

as 750 mmol m�2 s�1. The response to high light also included the formation of NPQ, expected to

represent mainly the operation of the xanthophyll cycle, which attained high values, above 5.9 for

1500 mmol m�2 s�1. Despite the photoprotection provided by energy-dissipation processes and

photophobic behaviour, the light response of most photophysiological parameters showed a clear

counter-clockwise hysteresis pattern, indicating the occurrence of photoinhibition. Hysteresis was due

to the incomplete recovery of photosynthetic activity during the afternoon low tide, and its magnitude

was dependent on the morning light doses.

& 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Microphytobenthos, the communities of benthic microalgaethat inhabit the intertidal flats of estuaries, form dense and highlyproductive diatom-dominated biofilms at the surface of the fine-grain sediments. The large amounts of microalgal biomass thataccumulates in the photic layers of the sediment during diurnallow tides and the capacity for high photosynthetic activity under awide range of environmental conditions enable intertidal micro-phytobenthos to reach community-level primary productivity

ll rights reserved.

+351 234372587.

rates comparable to those attained by phytoplankton (Underwoodand Kromkamp, 1999). Being affected by the superimposition ofdiurnal and tidal cycles, the estuarine intertidal environment ischaracterized by short-term variability in various key factorscontrolling the photosynthetic activity of microphytobenthos.During daytime low tides, benthic microalgae are frequentlyexposed to extreme conditions, which frequently may includesupersaturating irradiances, desiccation, and high salinity (Brotaset al., 2003). They are systematically subjected to extendedperiods of direct exposure to sunlight, alternating with longperiods of darkness during immersion in turbid water or night,and, during ebb or flood, experience sudden transitions in light,salinity and temperature. Moreover, the high photosynthetic ratesattained by dense biofilms may cause the local depletion of

ARTICLE IN PRESS

Table 1Notation

aRLC, bRLC Initial slope and photoinhibition parameter of a RLC (mmol–1 m2 s)

am,RLC Maximum daily aRLC value (mmol–1 m2 s)

DF/Fm0 Effective quantum yield of PSII (dimensionless)

DaRLC High light-induced decrease of aRLC (mmol–1 m2 s)

E Spectrally averaged irradiance of PAR (400–700 nm) (mmol m�2 s�1)

EFY Photosynthesis index (dimensionless)

Fo, Fm Minimum and maximum fluorescence emitted by a dark-adapted

sample (arbitrary units)

Fm0m Maximum Fm

0 value measured during a light curve (arbitrary units)

Fs, Fm0 Steady-state and maximum fluorescence emitted by a light-adapted

sample (arbitrary units)

Fv/Fm Maximum quantum yield of PSII of a dark-adapted sample

(dimensionless)

Fv0/Fm

0 Quantum efficiency of open PSII RCs (dimensionless)

NPQ Non-photochemical quenching coefficient (dimensionless)

PSII Photosystem II

RC Reaction center

rETR Relative electron transport rate (dimensionless)

rETRm,RLC Maximum relative electron transport rate in a RLC (dimensionless)

RLC Rapid rETR vs E curve; subscript denoting a parameter of a RLC

S Salinity

T Temperature (1C)

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–13751364

nutrients (Kromkamp et al., 1998; Miles and Sundback, 2000) or ofinorganic carbon (Glud et al., 1992; Cook and Røy, 2006).

The prolonged exposure to supersaturating irradiance levels isa major source of stress to the photosynthetic apparatus and apotential cause of limitation of microphytobenthos productivity.Excessive light energy absorbed by the light-harvesting complexes(LHC) and not used for photochemistry or dissipated as heat, maycause photoinhibition, the permanent or slowly reversibledamages to the photosynthetic apparatus, mainly through theinactivation of photosystem (PS) II protein D1 (Kyle et al., 1984;Falkowski et al., 1994). Diatoms respond to excessive light byoperating the xanthophyll cycle, through which the pigmentdiadinoxanthin (DD) is rapidly and reversibly converted in theenergy-dissipating form diatoxanthin (DT) (Arsalane et al., 1994;Olaizola and Yamamoto, 1994; Casper-Lindley and Bjorkman,1998). Under high light, the accumulation of DT in the LHCantennae reduces the excitation energy delivered to PSII RCs andminimizes potential damages to the photosynthetic apparatus. Inthe case of microphytobenthos dominated by epipelic diatoms, asecond form of photoprotection has been proposed to take place,based on the photophobic migratory response of motile diatomsto high irradiance. This ‘behavioural photoacclimation’ consists inthe high light-induced vertical migration of the cells, from thesurface to less illuminated zones of the sediment (Kromkampet al., 1998; Perkins et al., 2001; Serodio et al., 2001; Underwood,2002).

Under natural conditions, the effective impact of excessivelight may be enhanced by the combined effects of a number ofother environmental factors affecting biofilm condition duringlow tide, such as air temperature and humidity, and wind speed.Environmental conditions such as these are very difficult toreplicate indoors, and laboratory studies are thus likely to fail torealistically describe the responses of microphytobenthos in itsnatural environment. Considering this limitation, a growingnumber of studies on the photosynthetic performance andproductivity of intertidal microphytobenthos have been carriedout in situ (Miles and Sundback, 2000; Perkins et al., 2001;Underwood, 2002; Brotas et al., 2003; Migne et al., 2004;Honeywill et al., 2006; Spilmont et al., 2006, Migne et al., 2007).However, only a few of these have simultaneously characterizedthe photophysiological and migratory responses of microphyto-benthos to changing environmental conditions (Brotas et al.,2003; Perkins et al., 2001) and, to our knowledge, none hasaddressed quantitatively the functioning of photoprotectiveprocesses and the occurrence of photoinhibition.

The objectives of this study were (1) to characterize the short-term variability in photosynthetic activity in situ during periods oflow tide; (2) to relate these changes to the natural variability ofthe main environmental factors and (3) to relate to the operationof photoprotective mechanisms, distinguishing the relative role ofphysiological and migratory processes, and (4) to evaluate theeffectiveness of these processes in preventing photoinhibition.

The photosynthetic activity, photoprotection and photoinhibi-tion of microphytobenthos were studied in situ by using PulseAmplitude Modulated (PAM) fluorometry (Schreiber et al., 1986).This technique allows to simultaneously and in real time measurethe activity of PSII, assess the operation of the xanthophyll cycleand the occurrence of photoinhibition, and quantify verticalmigration-induced changes in the surface biomass of undisturbedbiofilms. The photosynthetic activity of biofilms was assessed bymeasuring the effective and maximum quantum yield of PSII(DF/Fm

0 and Fv/Fm, respectively; see Table 1 for notation), afluorescence-based index for gross depth-integrated photosyn-thetic rates (EFY), and by constructing rapid light-response curves(RLC). Xanthophyll cycle operation and occurrence of photoinhibi-tion were assessed by the non-photochemical quenching (NPQ)

coefficient and the quantum efficiency of open RCs (Fv0/Fm

0).Behavioural photoprotection was investigated by monitoring Fo, aproxy for the photosynthetic biomass present in the photic zone.

2. Materials and methods

2.1. Study site and periods

The study was carried out on an intertidal mudflat located nearVista Alegre, on the Ria de Aveiro, a mesotidal estuary located inthe west coast of Portugal (Fig. 1). The study site consisted of finemuddy sediment (97% particles o63 mm), colonized throughoutthe year by dense microphytobenthos assemblages, dominated bydiatoms of the genera Navicula, Nitzschia, Gyrosigma and Pleur-

osigma. Measurements were made during periods of low tide,from sunrise to sunset, on a total of 5 days, from 14 to 27 July2004, distributed along a complete spring–neap tidal cycle. Thepresented results refer to 3 days, chosen as representative of thefortnightly variability in the timing of light exposure and tidalsubmersion: 2 days when diurnal low-tide exposure occurredtwice during the photoperiod (14 and 27 July) and 1 day duringwhich a single light exposure period took place (23 July). Days 14and 27 July were chosen to illustrate the effects of contrastinglydifferent sunlight exposure conditions: clear sky during the wholeof day 14, and a dense fog during most of the morning low tide onday 27. All times are reported in UTC+1.

2.2. In situ measurements: physical variables

During each period of low tide, incident irradiance(E, spectrally averaged solar irradiance, PAR, 400–700 nm),sediment temperature (T) and salinity (S), were measured atregular intervals at the site of the measurements. Irradiance wasmeasured using a PAR sensor (LI-193SA and LI-250 light meter,Li-Cor, Lincoln, Nebraska, USA). Sediment surface temperaturewas measured non-invasively using an infrared thermometer(C.A. 872, Chauvin Arnoux, France). Salinity was measured onsediment surface interstitial water, collected from lens tissuesplaced on the sediment surface, using a salinity refractometer(S/Mill, Atago, Japan).

ARTICLE IN PRESS

Fig. 1. Map of the Ria de Aveiro, Portugal, and location of the study site (VA).

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–1375 1365

2.3. In situ measurements: variable fluorescence

Variable chlorophyll fluorescence was measured using aportable PAM fluorometer (Portable Junior-PAM, GademannInstruments GmbH, Germany). This fluorometer applied a modu-lated blue light (LED-lamp peaking at 470 nm, half-bandwidth of31 nm) as source for measuring, actinic and saturating light,emitted at a frequency of 25 Hz when measuring the minimumfluorescence level (Fo) or 1.2 kHz when measuring other fluores-cence parameters. Fluorescence was measured in situ using a1.5 mm-diameter plastic fibreoptics (Edmund Optics, UK), main-tained at a constant distance of 2 mm from the sediment surface,at a approx. 451 angle, and facing South to minimize shadingeffects. To facilitate the adjustment of the relative position of thefibreoptics and the sediment surface, measurements were madeon undisturbed sediment samples (3.6 cm diameter), collectedusing plastic corers. Samples were kept moist during the wholeperiod of measurements, by carefully adding site water asnecessary to replace evaporated interstitial water, withoutdisturbing or submerging the sediment surface. This was foundnecessary because preliminary tests showed that, mainly onsunny and windy days, core samples not kept hydrated sufferedvery rapid desiccation, not being representative of the conditionson the tidal flat. A total of 6 fibreoptics were used to measurefluorescence in the same number of replicated samples. Measure-ments on different samples were made sequentially, by main-

taining one end of each fibre fixed in its measuring positionpointing to the sample, and by connecting the free end to thefluorometer.

Every 15 min, a saturating pulse was applied and the effectivequantum yield of PSII, DF/Fm

0 ( ¼ (Fm0-Fs)/Fm

0; Genty et al., 1989),was measured on each sample. On 3 of these samples, one RapidLight Curve (RLC; White and Critchley, 1999) was constructedevery 30 min. Each samples was darkened for 10 s, using a blackplastic cover, after which it was exposed for 10 s to each of 8increasing actinic light levels (provided by the fluorometer), from80 to 1600mmol m�2 s�1. Under each light level, Fs and Fm

0 weredetermined and the relative electron transport rate was calculatedby rETR ¼ E�DF/Fm

0. The first value of the each RLC, measuredafter 10 s of darkness, was taken as an estimation of Fv

0/Fm0

( ¼ (Fm0�Fo

0)/Fm0), the quantum efficiency of open PSII RCs (Ralph

et al., 1999). The sample remained in the dark for 2 min, afterwhich Fo and the maximum quantum yield of PSII, Fv/Fm

( ¼ (Fm�Fo)/Fm), was determined, before it was returned toambient light. The fluorescence parameter Fo was used as a proxyfor productive biomass, the photosynthetic biomass present ateach moment in the photic zone of the sediment (Serodio et al.,2001). The Fo value measured after only 2 min of darkness cannotbe taken as representing the full dak-adapted state of the sample,but was shown to be a good estimator of microalgal biomass(Serodio et al., 2006a, 2007). Using the values of E and DF/Fm

0

determined before the start of the RLC, the fluorescence-basedphotosynthesis index, EFY, was calculated by (Serodio et al., 2007)

EFY ¼ E� Fo � DF=F 0m (1)

RLCs were characterized by estimating the parameters aRLC

(initial slope), rETRm,RLC (maximum rETR) and bRLC (photoinhibi-tion parameter), through fitting the model of Platt et al. (1980) torETR vs E curves. The model was fitted iteratively by minimizing aleast-squares function using MS Excel Solver. Curve fit was verygood in all cases (rX0.913 in a total of 133 RLCs).

2.4. NPQ estimation

As methodological problems impede the determination ofNPQ, generally used to quantify the operation of the xanthophyllcycle in microphytobenthic biofilms, an alternative method wasapplied in this study, based on the variation of the initial slope ofRLC, aRLC, shown to relate linearly to NPQ (Serodio et al., 2006b;Cruz & Serodio, 2008). To confirm the applicability of the method,NPQ and aRLC were compared on microphytobenthos suspensions(to avoid migratory effects on fluorescence parameters) exposedto a wide range of light conditions. Surface sediment (ca. 1 cmdeep) was collected in the days of field work and taken to thelaboratory, where microalgae were collected using the ‘lens tissuetechnique’ as described in Serodio et al. (2005), in the followingday. Variable fluorescence was measured using a PAM fluorometercomprising a computer-operated PAM-Control Unit (Walz, Ger-many) and a WATER-EDF-Universal emitter-detector unit (Gade-mann Instruments GmbH, Germany). A modulated blue light(LED-lamp peaking at 450 nm, half-bandwidth of, 20 nm) wasapplied as source for measuring, actinic and saturating light,delivered through a 1.5 mm-diameter plastic fibreoptic connectedto a fluorescence cuvette (KS-101, Walz, Germany). The fluores-cence cuvette was connected to a recirculating water bath(Frigiterm-10, Selecta, Spain) and all measurements were madeat 20 1C.

NPQ and RLCs were compared as described in Serodio et al.(2006b). Microalgae were dark incubated for 30 min, after which asaturating pulse was applied and fluorescence levels Fo and Fm

were measured. Samples were then exposed to 12 increasing

ARTICLE IN PRESS

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–13751366

irradiance levels, from 12 to 920mmol m�2 s�1. Under each level, Fs

and Fm0 were determined each 90 s, until a steady state in DF/Fm

0,was reached (minimum 7.5 min). For each light level E, DF/Fm

0 andNPQ were calculated and DF/Fm

0 vs E and NPQ vs E curves wereconstructed. Due to the possible occurrence of Fm

0 values higherthan the Fm value measured after dark adaptation, NPQ wascalculated by adapting the non-photochemical quenching coeffi-cient (Schreiber et al., 1994):

NPQ ¼ ðF 0m m � F 0mÞ=F 0m (2)

where Fm0m is the maximum Fm

0 value measured during the lightcurve (Serodio et al., 2005). One RLC was constructed after asteady state was reached (constant DF/Fm

0 values) under each lightstep of the NPQ vs E curve and immediately after NPQ had beendetermined. RLCs were constructed by exposing the sample for10 s to each of 12 increasing actinic light levels (the same used inthe NPQ vs E curves). NPQ was compared with the variation(decrease) of aRLC under high light, as calculated by (Serodio et al.,2006b)

DaRLCðEÞ ¼ am;RLC � aRLCðEÞ (3)

where am,RLC is the maximum aRLC value measured during thelight curve. Time series of NPQ were generated by applying Eq. (3)to observations of aRLC obtained in situ and estimating am,RLC fromthe maximum value observed during each day.

2.5. Taxonomic composition

The species composition of the biofilms was determined byanalysing replicated samples of the microalgal suspensions usedin the laboratory measurements. Microalgal suspensions werefixed in 1% v/v formaldehyde and viewed using bright-fieldmicroscopy for determination of the relative abundance ofdiatoms, euglenophytes and cyanobacteria. Identification ofdiatom species was carried out on subsamples oxidized usingsaturated potassium permanganate solution and digested inconcentrated HCl.

3. Results

3.1. Physical variables

During the study period, daylength was approx. 14 h 30 min(from sunrise at 6h30 to sunset at 22h00). The submersion duringhigh tide reduced the effective photoperiod to ca. 10 h (twoperiods of 7 h and 3 h, corresponding to low tide), 7 h, and 9 h30 min (4 h 30 min and 5 h) on days 14, 23 and 27 July,respectively. The days with two periods of low tide werecharacterized by clear sky throughout the day, with irradiancereaching maximum values over 1750 mmol m�2 s�1 (Fig. 2a) and1430mmol m�2 s�1 (day 27, Fig. 2c) during the low-tide periods.During the morning of day 27, foggy conditions temporarilyreduced E to values below 400 mmol m�2 s�1. Dissipation of fogbeginning at 10h00 caused a rapid increase in incident irradiance,that raised by more than 3.5-fold in 1 h. On day 23, partial coverby high clouds reduced irradiance throughout the day, whichnevertheless remained above 500 mmol m�2 s�1 during the wholelow-tide period (Fig. 2b).

The periods of low tide were also characterized by largevariations in the temperature and salinity of the surface of thesediment (Fig. 2a–c). Large increases in temperature (up to 75%,from 16 to 28 1C; Fig. 1a) were observed in the morning low-tideperiods, closely following the increase in incident solar radiation,a dependency particularly evident when the morning periods ofdays 14 (clear sky) and 27 July (fog) are compared (Figs. 2a, c).

Large variations were also observed regarding salinity, particularlyduring the afternoon period, when it continued to increase whilstirradiance and temperature were already decreasing (Fig. 2b, c).Increases of more than 100% during the same low-tide periodwere observed, with maximum values reaching above 65, andvalues of 50 being common.

3.2. Photosynthetic activity

A clear inverse relationship was found between DF/Fm0 and E

on all days (Fig. 2d–f). On day 14 July, DF/Fm0 decreased with the

morning increase in E, this decrease becoming more markedwhen E reached values above 600 mmol m�2 s�1 (Fig. 2d). ForE4900 mmol m�2 s�1, DF/Fm

0 remained very low, below 0.15.During the afternoon low tide, DF/Fm

0 recovered monotonicallyfollowing the decrease in E. A similar response pattern wasobserved on day 27, but with the difference that, during the earlierpart of the morning DF/Fm

0 started by increasing until it reached amaximum before decreasing (Fig. 2f). Maximum DF/Fm

0 valuesattained on this day were higher than those of day 14 (p ¼ 0.047,t-test), and remained relatively high during most part of themorning low tide, starting to decrease significantly only whenE4500 mmol m�2 s�1 (Fig. 2d, f). On 23 July, when low tideoccurred during the middle of the day, DF/Fm

0 decreased steadilyduring most of the exposure period, even when E remainedconstant, around solar noon (ca. 1500 mmol m�2 s�1) (Fig. 2e).Recovery of DF/Fm

0 started only when irradiance decreased below1000 mmol m�2 s�1.

Fv/Fm displayed the same overall pattern of variation of DF/Fm0,

mostly characterized by inverse variation with E (Fig. 2g–i). Fv/Fm

was generally higher than DF/Fm0, the difference between the two

decreasing with E (Fig. 2d–i). Exceptions to this tendency occurredon the morning of 27 July, during which both indices showedalmost identical values (and higher than on day 14) (Fig. 2f, i),and, at the end of the photoperiods of both days, whenDF/Fm

0 increased at a higher rate than Fv/Fm and reached valuesmarkedly higher than Fv/Fm (respectively, 14.8 and 32.9% higher;t-test, p ¼ 0.005 and 0.010). As a consequence, while the recoveryof DF/Fm

0 in comparison with early morning value was almostcomplete (for days 14 and 27 July, respectively, �23.5% and�13.0%; t-test, p ¼ 0.134 and 0.051), Fv/Fm appeared to recoveronly partially, attaining values considerably lower than thoseobserved at the beginning of the day (respectively, �56.6% and�37.8%; t-test, po0.001 and p ¼ 0.004) (Fig. 2d, f, g, i). Theirradiance level for which DF/Fm

0 equalled Fv/Fm was similar onboth days, of ca. 150–200 mmol m�2 s�1.

The pattern of variation of EFY was also strongly dependent onlight history, as evident by comparing the morning low tidesof days 14 and 27 July. In the first day, EFY displayed a bi-phasicpattern, with a marked increase due to the large increaseof E and slow decrease of DF/Fm

0. Maximum values were reachedaround 8h30, which was then followed by a gradual decrease,corresponding to the decrease of DF/Fm

0 and Fo (Fig. 3j, see below),until it remained relatively constant for the rest of the low tideperiod (Fig. 2j). On the afternoon low tide, a second peak wasobserved, due to the recovery of DF/Fm

0 and the rise in Fo

immediately after ebb, and to the later decrease in E. In contrast,on day 27 July, EFY increased during most of the morning low tide,reaching the highest value shortly before flood, and showing somedecline only after the fog has cleared out (Fig. 2l). The variationpattern observed in the afternoon low-tide period was generallysimilar to the one observed on day 14, but presenting highervalues. On 23 July, EFY showed a steep decline during the firsthours of the low-tide period, after which it remained relativelystable (Fig. 2k).

ARTICLE IN PRESS

0

250

500

750

1000

1250

1500

1750

Time of day06:000

20

40

60

80

100

120

140

0.0

0.1

0.20.3

0.4

0.50.6

0.7

0

250500

7501000125015001750

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

E (μ

mol

m-2

s-1

)

0

250

500

750

1000

1250

1500

17500

10

20

30

40

50

60

70

80

0250

500

750

1000

1250

1500

1750

200014 July 23 July 27 July

E

S

T

T (°

C),

SΔF

/Fm

'E

FY (a

.u.)

F v/F

m

10:00 14:00 18:00 06:00 10:00 14:00 18:00 06:00 10:00 14:00 18:00 22:00

Fig. 2. Time series of PAR irradiance (E), temperature (T) and salinity (S) at the sediment surface (a–c), of the effective (DF/Fm0; d–f) and maximum quantum yield of PSII

(Fv/Fm; g–i), and of the fluorescence-based proxy for photosynthetic rate EFY (j–l), during diurnal low-tide periods of 3 days of a spring–neap tidal cycle. Vertical bars

represent71 standard error (d–f: n ¼ 6; g–l: n ¼ 3).

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–1375 1367

The hourly variation of Fv0/Fm

0 and aRLC displayed a similarpattern, generally characterized by an accentuated decrease withthe morning rise in E and a partial recovery during the afternoonperiod (days 14 and 27; Fig. 3a, d and d, f), or by maintaining lowvalues during the mid-day high light exposure, that graduallyincreased under the afternoon decline in irradiance (day 23;Fig. 3b, e). As with DF/Fm

0, the light history of the morning periodconditioned significantly the hourly variation of these indices,with clear differences found between days 14 and 27 July. In themorning of day 14, when E rose rapidly, both indices decreasedsteeply to very low values, particularly Fv

0/Fm0, which remained

very close to 0 during the rest of this period (Fig. 3a). In contrast,during 27 July, a significant decrease started only shortly beforeflood, with much higher values measured for comparable E levels(Fig. 3c, f). On day 27, an increase in the two indices was observed

during the first hours of the morning period (Fig. 3c, f). This wasalso observed for aRLC on day 14, although in a much less extent(Fig. 3d).

The daily pattern of rETRm,RLC was also strongly affected bylight conditions occurring during the morning. On day 14 July, alarge initial increase was followed by a steep decline for E4600mmol m�2 s�1, after which rETRm,RLC decreased at a relatively lowrate (Fig. 3g). On day 27, rETRm,RLC increased gradually duringmost of the morning low tide, only decreasing when fogdissipated and incident irradiance rapidly increased (Fig. 3i). Onall days, rETRm,RLC decreased when E declined by the end of theday (Fig. 3g–i).

The results obtained for these 3 days were representative ofthe whole spring–neap tidal cycle. Observations made on theremaining days of the tidal cycle (summarized in Table 2) were in

ARTICLE IN PRESS

0

250

500750

1000

1250

1500

1750

0

100

200

300

400

500

600

7000

50

100

150

200

250

300

350

14 July 23 July 27 July

0.0

0.10.20.3

0.4

0.50.6

0.70.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

F v'/F

m'

rETR

m,R

LCα R

LCF o

(a.u

.)

Time of day

0

250

500

750

1000

1250

1500

1750

0

250

500750

10001250

1500

1750

E (μ

mol

m-2

s-1

)

0

250

500

750

1000

1250

1500

1750

2000

06:00 10:00 14:00 18:00 06:00 10:00 14:00 18:00 06:00 10:00 14:00 18:00 22:00

Fig. 3. Time series of the efficiency of open reaction centres (Fv0/Fm0; a–c), of rapid light-response curve parameters aRLC (d–f) and rETRm,RLC (g–i), and of the proxy for

productive biomass, Fo (j–l), during diurnal low tide periods of 3 days of a spring–neap tidal cycle. Vertical bars represent71 standard error (n ¼ 3).

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–13751368

general agreement with the range and patterns of variationpresented in detail above. DF/Fm

0, Fv/Fm, Fv0/Fm

0 and aRLC displayedthe same overall inverse variation with E, with lowest valuesgenerally coinciding with highest light levels, and highest valuesbeing observed for lower irradiances. EFY and rETRm,RLC were alsodepressed under high E levels, the higher values corresponding tomoderate irradiances (Table 2).

3.3. Productive biomass

The hourly variation of Fo, the proxy for microphytobenthosproductive biomass, was characterized by high values during theearly part of the morning low tides, which gradually decreased asE increased (Fig. 3j, l). This early morning peak corresponded tothe accumulation of large numbers of microalgae in the upper-most layers of the sediment, and was better defined on day 27

July, when morning irradiance was lower (Fig. 3l). On both days 14and 27 July, Fo started to decrease at roughly the same time(8h30–9h00), to values similar to those observed on the afternoonlight period (Fig. 3j, l). On day 14 July, this decrease in Fo becamenoticeable when irradiance reached over ca. 750mmol m�2 s�1,whilst on day 27 July it started under considerably lowerirradiance levels, of 150–200 mmol m�2 s�1 (Fig. 3j, i). An inverserelationship between Fo and E was also observed, although moreattenuated, in the afternoon low tide. An increase in Fo followingthe decrease in E also occurred on day 23, although this tendencywas latter inverted, anticipating the flood by high tide (Fig. 3k).

3.4. Light responses

To better assess the effects of light history on the hourlyvariability of the various fluorescence indices, light-response

ARTICLE IN PRESS

Table 2Range of variation (minimum and maximum values) of the physical variables (E, T, S) and fluorescence parameters and indices (DF/Fm

0 , Fv/Fm, Fv0/Fm

0 , EFY, aRLC, rETRm,RLC, Fo)

measured during diurnal low-tide periods of five days along a spring–neap tidal cycle in July 2004

Day Measuring period E T S DF/Fm0 Fv/Fm Fv

0/Fm0 EFY aRLC rETRm,RLC Fo

14 6h57–13h30 80 16 25 0.035 (1705) 0.066 (1705) 0.017 (1719) 8.1 (80) 0.097 (1705) 50.3 (1705) 111.3 (1719)

1720 30 51 0.562 (80) 0.656 (80) 0.649 (80) 58.8 (734) 0.625 (320) 193.8 (320) 430.9 (320)

18h14–21h22 1 20 35 0.068 (880) 0.220 (590) 0.059 (756) 0.1 (1) 0.192 (756) 30.4 (1) 151.0 (756)

880 37 0.481 (1) 0.419 (1) 0.428 (1) 20.6 (590) 0.432 (17) 101.4 (227) 243.7 (1)

17 8h28–15h29 85 16 35 0.029 (1300) 0.060 (1650) 0.053 (1832) 9.2 (85) 0.037 (1730) 17.8 (1730) 167.3 (1730)

1830 20 47 0.579 (85) 0.604 (116) 0.549 (85) 79.7 (811) 0.584 (85) 225.0 (158) 617.0 (811)

20 10h28–17h11 1220 17 34 0.031 (1661) 0.019 (1266) 0.009 (1655) 6.4 (1220) 0.154 (1446) 69.0 (1674) 111.3 (1635)

1690 35 66 0.119 (1542) 0.263 (1356) 0.092 (1356) 16.1 (1356) 0.207 (1356) 124.5 (1446) 278.3 (1542)

23 12h19–18h43 401 20 35 0.033 (1339) 0.219 (1172) 0.061 (1391) 26.1 (401) 0.147 (1391) 52.9 (580) 275.0 (1391)

1450 26 65 0.220 (491) 0.361 (1430) 0.245 (401) 111.5 (1430) 0.401 (401) 132.8 (1430) 404.1 (900)

27 7h10–11h05 33 18 39 0.224 (1367) 0.272 (1025) 0.184 (1025) 5.6 (46) 0.461 (1025) 68.3 (33) 239.2 (610)

1367 31 41 0.626 (56) 0.613 (70) 0.647 (70) 59.4 (610) 0.725 (70) 170.7 (610) 492.7 (124)

15h58–20h44 13 20 39 0.128 (1427) 0.185 (1060) 0.031 (880) 8.1 (46) 0.153 (1307) 64.8 (46) 192.1 (1060)

1427 32 51 0.566 (13) 0.398 (153) 0.511 (46) 44.1 (534) 0.538 (46) 132.5 (534) 338.3 (46)

Average values of three or six (DF/Fm0) replicated measurements. Numbers in parenthesis are E values at the time of the indicated minimum and maximum observations.

For days 14 and 27 results are given for the two periods of diurnal low tide.

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–1375 1369

curves were reconstructed from the time series of observationsduring the entire effective photoperiod (Fig. 4). In mostcases, a clear counter-clockwise hysteresis pattern was observed,with the decrease observed during the morning not beingmatched by the subsequent recovery during the afternoon(e.g. Fig. 4a, c). A clear effect of the morning light dose was foundfor all indices, the difference between the morning and theafternoon light responses being generally higher on day 14 thanon day 27 (Fig. 4).

In the case of DF/Fm0 and Fv/Fm (as well as aRLC, data on Fig. 3), a

higher light dose in the morning caused a larger relative decreaseof the values measured during the afternoon, particularly underlower irradiance levels (Eo500–600 mmol m�2 s�1; Fig. 4a–d).During the morning low tides, the light responses of DF/Fm

0 andFv/Fm were characterized by an initial phase when valuesremained constant or increased slightly, followed by an almostlinear decrease until E41000 mmol m�2 s�1 (Fig. 4a–d). In theafternoon period, the two indices recovered following alinear trend virtually parallel to the observed during themorning, although presenting lower values and not repeatingthe peak under low irradiance. EFY showed a bi-phasic lightresponse pattern, more accentuated during the morning period,increasing for Eo600 mmol m�2 s�1, and gradually decreasingunder the remaining irradiance range (Fig. 4e, f). Again, the 2days differed regarding the magnitude of hysteresis in the lightresponse of EFY, with a large difference (4250% for the300–600 mmol m�2 s�1 range) being measured between themorning and afternoon values of day 14, while an almostcomplete match was observed on day 27 July (Fig. 4e, f).Hysteresis was observed not only for the fluorescence indicesrelated to photosynthetic activity, but also regarding Fo, mostlydetermined by productive biomass (Fig. 4g, h). As with DF/Fm

0 andEFY, the match between the morning and the afternoon lightresponses was in general larger on 27 July.

3.5. NPQ induction

The comparison between aRLC and the NPQ level attainedunder steady state on resuspended microphytobenthos revealedthe two indices to be linearly related for all tested days. Highlysignificant (po0.001) linear regression equations were obtainedin all cases for the light range under which the induction of NPQwas observable (E4250mmol m�2 s�1) (Fig. 5a). The slope of the

regression equations of aRLC on NPQ was found not to differsignificantly among days (ANCOVA, Sokal and Rohlf, 1981) and asingle equation could be used to describe all data (Fig. 5a).

On the basis of this linear relationship, NPQ could be estimatedfrom RLCs and its short-term variability in situ could becharacterized (Fig. 5b, c). On 14 July, the NPQ induction startedearly in the morning, when irradiance levels reached over ca.500–600 mmol m�2 s�1 (Fig. 5b). NPQ increased following asaturation-like pattern, stabilizing at maximum values of5.6–5.9 for E41300 mmol m�2 s�1. On day 27, NPQ started bydecreasing from a relatively high initial value of ca. 1.2 to nearlyzero. During the rest of the morning, NPQ induction was moregradual, reaching only ca. 50% of the maximum values of day 14,although after a shorter exposure period (Fig. 5c).

The magnitude of the NPQ induction was inversely related tothe difference between DF/Fm

0 measured in situ and on suspen-sions, under optimal, stress-free conditions. This difference wassubstantially larger on day 14 July, when DF/Fm

0 measured in situ

were, apart from the initial part of the morning, much lower thanthe values measured in suspensions (a difference up to 0.40 forE ¼ 1200 mmol m�2 s�1). On day 27, DF/Fm

0 measured in situ wasmuch nearer optimal conditions, the difference between in situ

and suspensions remaining almost constant for most of the E

range (E4300 mmol m�2 s�1), and not exceeding 0.10.

3.6. Taxonomic composition

The species composition of the microalgal suspensions did notvary substantially along the study period, the assemblages beingdominated by five species of motile diatoms, which accounted for76–83% of total cell counts: Navicula cf. gregaria Donkin, Nitzschia

draveillensis Coste and Ricard, Nitzschia frustulum (Kutzing) Grunow,Nitzschia perspicua Cholnoky, Parlibellus crucicula (W. Smith)Witkowski, Lange-Bertalot and Metzeltim. Euglenophytes andcyanobacteria accounted for less than 1% of cell counts on all days.

4. Discussion

4.1. Photosynthetic activity

The photosynthetic activity of microphytobenthos duringdaytime low tides was characterized by a large short-term

ARTICLE IN PRESS

00

100

200

300

400

500

600

7000

15

30

45

60

75

90

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.0

0.1

0.2

0.3

0.4

0.5

0.6

14 July 27 July

E (μmol m-2 s-1)

F v/F

mF o

(a.u

.)E

FY (a

.u.)

ΔF/F

m'

morningafternoon

300 600 900 1200 1500 0 300 600 900 1200 1500 1800

Fig. 4. Light-response curves of effective (DF/Fm0; a, b) and maximum quantum yield of PSII (Fv/Fm; c, d), of the proxy for photosynthetic rate, EFY (e, f), and of the productive

biomass, Fo (g, h), reconstructed from the time series of data measured during the morning (open circles) and afternoon (closed circles) low-tide periods of 2 days of a

spring–neap tidal cycle. Vertical bars represent71 standard error (a, b: n ¼ 6; c–h: n ¼ 3). Arrows indicates the temporal order of the data points.

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–13751370

variability associated to the rapidly changing environmentalconditions, with DF/Fm

0 varying from physiological maxima tovalues close to zero during the course of single exposure periods.The differences found between the photosynthetic light responsesof days 14 and 27 July, with the high light-induced decrease ofDF/Fm

0 being much more pronounced on the first day, indicate thatthe pattern of variation of photosynthetic activity is stronglydetermined by the cumulative effects of prolonged exposure tohigh light. Also, the larger difference found on day 14 July betweenthe DF/Fm

0 light response of biofilms in situ and of microalgalsuspensions (considered as close to optimal), further shows thatthe photosynthetic activity at a given moment is not determined

by current irradiance level alone, but also by the previous lighthistory. Probable causes for this enhanced decrease of thephotosynthetic activity include the photoprotective downregula-tion of photosynthesis or the effects of photoinhibition, resultingfrom the continued exposure to high light. Other possible causesare the limitation of photosynthesis due to depletion of nutrientsor carbon within the dense biofilm, which can be expected as aresult of the high photosynthetic rates during the early part of themorning (Ludden et al., 1985; Glud et al., 1992; Kromkamp et al.1998; Miles and Sundback, 2000; Cook and Røy, 2006). Salt andosmotic stress may also contribute to depress photosynthesisduring prolonged low-tide exposures (Rijstenbil, 2003), but their

ARTICLE IN PRESS

0.0NPQ

0.00

0.05

0.10

0.15

0.20

0.2515 July24 July28 July

y = 0.080 x + 0.01n = 34, r2 = 0.941∗∗∗

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

NPQ

ΔF/Fm'

ΔF/Fm' suspension

ΔF/F

m'

ΔαR

LC (μ

mol

-1 m

2 s)

E (μmol m-2 s-1)

0.00

0.1

0.2

0.3

0.4

0.5

0.6

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

NP

Q

0.5 1.0 1.5 2.0 2.5

300 600 900 1200 1500 1800

Fig. 5. (a) Linear relationship between the initial slope of rapid light-response

curves (aRLC) and non-photochemical quenching coefficient (NPQ), for 3 days of a

spring–neap tidal cycle. (b, c) Light-response curves of the PSII effective quantum

yield (DF/Fm0) measured in situ and on suspensions, and of NPQ, as estimated from

aRLC, reconstructed from the time series measured during the morning low-tide

periods of days 14 July (b) and 27 July (c). Vertical bars represent71 standard error

(DF/Fm0: n ¼ 6; NPQ, DF/Fm

0 suspensions: n ¼ 3).

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–1375 1371

influence in this study was likely minimal as sample desiccationwas prevented during the measuring periods.

Because microphytobenthic biofilms are dominated by motilemicroalgae, their community-level, depth-integrated photosyn-thetic properties are also affected by migration-induced changesin the cell vertical distribution. As such, another possible cause ofthe accentuated decreases in biofilm photosynthetic activity is thedownward migration of the more physiologically competentmicroalgae, leaving in the photic zone cells in worst physiologicalcondition (e.g. non-migratory, therefore more probe to photo-damage). On the other hand, migration into the sediment couldhave accentuated the overestimation effect of depth-integrationof fluorescence emission on DF/Fm

0 when measured on intactbiofilms (Forster and Kromkamp, 2004; Serodio, 2004), particu-

larly under high irradiances, and has masked an actually largerdepressing effect of light. This depth integration effect must alsobe taken in consideration when comparing DF/Fm

0 measurementsmade on sediments and on suspensions, since it implies that theobserved differences may actually represent a larger difference inthe inherent physiological response of the microalgae.

The causes for the response of DF/Fm0 to high light may also

explain the pattern observed for rETRm,RLC, also characterized by amarked decrease under moderate to high light, both on days 14and 27 July. These results are partially in agreement with previousobservation made on intact biofilms, but contradict what isexpected for microphytobenthos cells under optimal conditions,for which an increase in rETRm,RLC with increasing irradiance wasconsistently observed (Serodio et al., 2006b; Cruz and Serodio,2008). Although determined from light-response curves too shortto allow reaching of steady state under each light step, rETRm,RLC isrelated to the maximum photosynthetic capacity, reached whenthe rate of photosynthesis is limited by the activity of the electrontransport chain or Calvin cycle enzymes (Behrenfeld et al., 2004;Ralph and Gademann, 2005). As such, increases of rETRm,RLC withirradiance have been observed and explained by the high light-induced activation of the carbon metabolism, which represents aphotoprotective response against potential photoinhibition causeby excessive light (Ralph et al., 1999).

The simultaneous decrease of DF/Fm0 and Fo, as observed on

days 14 and 27 July, caused the photosynthesis index EFY not todisplay maximum values under the highest light levels, aspredicted by productivity models based on laboratory observa-tions of the migratory control of productive biomass (Pinckneyand Zingmark, 1991; Guarini et al., 2000; Serodio and Catarino,2000). This illustrates the potentially erroneous informationretrieved from laboratory studies and the value of realistic, in situ

observations.

4.2. NPQ and photoprotection

The operation of the xanthophyll cycle is often investigated bymeasuring the quenching of variable fluorescence, NPQ, which hasbeen found to be linearly related to the relative DT content(Arsalane et al., 1994; Olaizola and Yamamoto, 1994; Serodio et al.,2005). However, due to methodological problems related to theconfounding effects of vertical migration on fluorescence para-meters, the calculation of the NPQ coefficient for intact micro-phytobenthos samples is problematic if not impossible (Serodioet al., 2005; Jesus et al., 2006). In this study, we were able tomonitor the evolution of NPQ on intact biofilms in situ byfollowing the short-term changes in DaRLC. This approach is basedon the co-variation of DaRLC and NPQ, due to their commondependence on the efficiency of light capture and energy transferfrom the LHC antennae to the PSII RCs (Serodio et al., 2006b). Thefinding of a single linear relationship between DaRLC and NPQmeasured on the different days ensured that, even if the value ofam,RLC varied among days, changes in DaRLC could be taken asproportional to changes in NPQ.

The maximum values estimated for NPQ, up to 5.9 forE41500 mmol m�2 s�1, are among the highest recorded for micro-algae or plants, for which typical maximum NPQ values remainbelow 4.0 (Rohacek, 2002; Ruban et al., 2004; Serodio et al., 2005),but similar to the maximum values documented for diatom-dominated microphytobenthos (Serodio et al., 2005, 2006b).The NPQ coefficient is determined by all non-photochemicalprocesses capable of lowering the fluorescence emission, whichinclude both rapidly reversible, photoprotective processes as thexanthophyll cycle (‘energy-dependent quenching’, qE) and slowlyreversible, photoinhibitory processes resulting from damages to

ARTICLE IN PRESS

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–13751372

the photosynthetic apparatus (‘photoinhibitory quenching’, qI;Muller et al., 2001). However, the NPQ estimated on the basis ofDaRLC may be expected to relate mainly to qE, since the formationof DT in the antennae bed is the cause for the high light-induceddecrease in light capture efficiency detected by aRLC (Serodio et al.,2006b). Moreover, the observed correlation between aRLC andFv0/Fm

0 further supports that the main features of the patterns ofvariation of aRLC are associated to the photoprotective componentof NPQ (qE), since Fv

0/Fm0 measures the quantum efficiency of

energy capture by open RCs, largely determined by downregula-tion (Rohacek, 2002). As such, the measured high NPQ values arelikely to represent an elevated capacity for photoprotection,developed as an adaptation against the damaging effects ofprolonged exposure to direct sunlight that characterizes theestuarine intertidal environment. It may be added that the highNPQ observed in this study may represent an underestimation ofthe values actually reached by the microalgae present at thesediment surface, due to the effects of depth integration ofupwelling fluorescence emitted by cells in deeper layers, exposedto lower light (Forster and Kromkamp, 2004; Serodio, 2004).

The observation of minimum NPQ values for low to moderateirradiances, and not under minimum light levels, were corrobo-rated by the early morning peaking of DF/Fm

0, Fv/Fm and Fv0/Fm

0,particularly on day 27 July, and can be explained by the formationof NPQ in the dark, and by its gradual dissipation during the firsthours of the morning. The induction of NPQ in the dark has beenreported for diatoms grown in culture, and was shown to occur toa great extent for microphytobenthos assemblages (Serodio et al.,2005, 2006b; and references therein). The most likely causes arethe occurrence of chlororespiration or the reverse operation ofATP synthase, both processes capable of forming a transthylakoi-dal proton gradient and inducing the production of energy-dissipating pigment diatoxanthin, DT (Ting and Owens, 1993;Olaizola and Yamamoto, 1994; Jakob et al., 1999; Dijkman andKroon, 2002; Lavaud et al., 2002). The production of DT in the darkwould also explain the incomplete recovery of aRLC during the endof the afternoon period, which could thus result from thecounteracting of the relaxation of NPQ following the decrease inirradiance.

4.3. Migration and photoprotection

The results showed evidence for the occurrence of photophobicdownward migration of microalgae as a response to highirradiance. On both days when ebb took place before sunriseand the low tide coincided with the beginning of the morning,microalgae accumulated at the surface but subsequentlystarted to migrate into the sediment. This pattern of behaviourwas more pronounced on day 27 July, when microalgae migratedto the surface when irradiance remained low due to foggyconditions. The downward migration started when E reached150–200 mmol m�2 s�1, a range of values that coincides with thethreshold level found to trigger a photophobic migration inlaboratory experiments (Serodio et al., 2006a). On the contrary, onday 14, descending migration appeared to start at a much higherlight level, which could lead to conclude that this migratoryresponse was not controlled by light. Although downwardmigration prior to high tide can occur independently of lightconditions as a result of a tidal endogenous rhythm (Brotas et al.,2003), this is unlikely the case, as the decrease in Fo started muchbefore (3–5 h) the actual tidal flood.

However, it must be considered that the above-mentionedresults of laboratorial experiments were obtained by testing theinduction of downward migration only after a dense biofilm hadbeen formed at the surface, under low light. The light conditions

in the morning of day 27 July resembled this protocol, since theyallowed for the biofilm to form under low light before it wasexposed to high light. In contrast, the early and rapid rise inirradiance on day 14 July makes the in situ results not directlycomparable to those obtained in the laboratory. In this case, thebiofilm was formed simultaneously with the increase in irra-diance, which may have caused the cells to accumulate at somedepth below the surface where light levels were lower. Inagreement with this is the fact that DF/Fm

0 remained nearoptimum values until surface irradiance reached values as highas 750mmol m�2 s�1. The relatively constant values of Fo duringthis period is thus likely to result from a balance between theendogenously controlled or low light-induced upward migrationand the downward photophobic response to the increase insurface irradiance.

The descending migration under high light has been inter-preted as a behavioural photoacclimation or photoprotectivestrategy, by which motile microalgae modulate their lightexposure by moving through the steep vertical light gradientwithin the sediment (Kromkamp et al., 1998; Perkins et al., 2001;Serodio et al., 2001; Underwood, 2002). Vertical photophobicmigration and the xanthophyll cycle are thus likely to act ascomplementary photoacclimation processes, the former minimiz-ing the extent of the activation of the later, both reducingpotential damages to the photosynthetic apparatus. Probablydue to the mentioned methodological problems regarding themeasurement of NPQ in biofilms, the combined operation of thesetwo processes has not been studied. The results of this studysuggest that vertical migration may display a preventive role,reducing or delaying the induction of the xanthophyll cycle. Thisidea is supported by the observations of day 14 July, when theearly morning upward migration resulted in the accumulation ofmicroalgae at subsurface layers, preventing the exposure toalready high light levels. This enabled the photosyntheticefficiency (DF/Fm

0) of the biofilm to remain near optimum values,and NPQ to remain low, until considerably high irradiance levelswere reached.

On the other hand, the migratory response to high light seemsto occur only after the induction of NPQ, as on both days theirradiance threshold for the start of downward migration washigher (although not much) than the one for the NPQ increase. Ina situation of prolonged exposure to high light, downwardmigration might represent an advantageous strategy to preventdamages resulting from the photoprotective capacity of thexanthophyll cycle being exceeded. On day 27 July, the rapidincrease in irradiance impinging on the microalgae accumulatedat the surface caused NPQ to start earlier (under lower E) than onday 14. However, the rapid and massive downward migration ofmicroalgae, together with the relatively short exposure to highlight, may have reduced the need for photoprotection through thexanthophyll cycle, and NPQ attained lower values for comparablelight levels.

4.4. Hysteresis and photoinhibition

Diatoms are known to have the capacity to develop very highlevels of non-photochemical quenching under high light stress(‘super-quenching’; Ruban et al., 2004; Lavaud et al., 2004), whichhas been shown to provide effective photoprotection againstphotoinhibition in estuarine species, both planktonic (Lavaudet al., 2007) and benthic (Cruz and Serodio, 2008). However, theobserved hysteresis patterns in the light response of mostfluorescence indices are a strong indication of the occurrence ofphotoinhibition. Counter-clockwise hysteresis caused by the ‘after-noon depression’ (Schanz and Dubinsky, 1988) of photosynthetic

ARTICLE IN PRESS

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–1375 1373

activity has been attributed to slowly reversible quenching offluorescence associated to the de novo synthesis of the D1 protein(Falkowski et al., 1994; Gorbunov et al., 2001; Levy et al., 2004).Although hysteresis may be due to other causes, like circadianrhythms, nutrient depletion, or photorespiration (Levy et al.,2004), its strong dependence on the light exposure dose found inthis study supports the idea that it was mainly due to damages tothe photosynthetic apparatus resulting from prolonged light stress.The importance of the light dose received during the morningperiod is reinforced by the observation that its effects are notreversed even when the afternoon photoperiod was relativelyshort and with moderate irradiance levels, as in the case of day 14July. On the basis of these results may be the prolonged exposureto irradiances exceeding the photoprotective capacity, which wasseen to saturate at 1300mmol m�2 s�1.

This is also supported by the observed light dose dependenceof the relative variation of Fv/Fm and DF/Fm

0. The short darkincubation (2 min) used for measuring Fv/Fm impeded the correctmeasurement of the true maximum quantum yield of PSII, theindex most commonly used to assess the extent of photoinhibition(Falkowski et al., 1994). However, the comparison of Fv/Fm withthe DF/Fm

0 level measured immediately prior to the darkening canbe informative for the distinction between qE from qI as the causefor the observed high light-induced decrease of photosynthesisefficiency. Under high light, larger differences between consecu-tive DF/Fm

0 and Fv/Fm values indicate a faster dissipation of theNPQ previously established, which can therefore be deduced to bemostly due to qE, whilst smaller differences between the twoindices are indicative of qI. The large differences found betweenDF/Fm

0 and Fv/Fm on day 14 July and the similarity observed on day27 reinforces that the difference between the 2 days is caused bylight exposure during the morning periods. On the morning of theday 14, Fv/Fm values remained very low, only slightly abovethe corresponding DF/Fm

0 levels, indicating a slow reversal of theestablished NPQ, therefore likely to be constituted by a significantqI component. In contrast, on day 27, Fv/Fm decreased only by 0.1during most of the morning period (from ca. 0.6 to 0.5; with theexception of the last measurement), remaining well above DF/Fm

0,suggesting a much larger NPQ relaxation, thus likely due to qE.This rapid relaxation of a significant fraction of NPQ is inagreement with the data available for NPQ relaxation measuredon microphytobenthos, which was shown to attain 45% in 2 minafter exposure to 1700 mmol m�2 s�1 (Serodio et al., 2005). Duringthe second low tide, the two indices became similar as irradiancedecreased, until Fv/Fm became less than DF/Fm

0 by the end of thephotoperiod. This pattern was observed on both days 14 and 27July, and can be explained by the formation of NPQ in the dark.

The finding of photoinhibition in microphytobenthos biofilmsmay be considered as unexpected, since decreases in community-level photosynthesis under high light are usually not observed(Kromkamp et al., 1998; Underwood and Kromkamp, 1999;Underwood, 2002). Hysteresis in the photosynthetic light re-sponse of microphytobenthos has been described before, but of amuch reduced importance (Serodio et al., 2007). Furthermore,hysteresis patterns were found despite the interruption of thehigh light exposure provided by high tide, which additionallyallowed to ‘reset’ the high levels of salinity, temperature and otherpotentially stressful conditions such as supersaturated oxygenconcentrations and low carbon and nutrient levels. These resultsmay thus be seen as contradicting the often put forwardhypothesis that photophobic vertical migration represents aneffective mechanism against photoinhibition. In particular, thehypothesis of ‘micromigration’, by which stressed cells in thephotic zone are continuously replaced by others in betterphysiological condition (Kromkamp et al., 1998; Underwoodet al., 2005), would predict a full recovery during the afternoon

low-tide periods and the absence (or large reduction) of hyster-esis. In contrast, the obtained results suggest that, under naturalconditions, either micromigration does not occur or it does notresult in significant beneficial effects regarding community-levelphotoinhibition. However, these results may not be generalized tothe whole year, as the photoprotection capacity of microphyto-benthos communities from the same sampling site was found tovary (increase) significantly from summer to winter (Serodio et al.,2005), which may allow a reduced sensitivity to high light andlower levels of photoinhibition.

The question remains open as to whether the hysteresisdetected for the biofilm as a whole, is due to the change in thephysiological status of the same microalgal population, or to thechange in the microalgal population present at the sedimentsurface. The method used for characterizing the species composi-tion of the biofilms did not allow for detecting short-term changesin the species present in the uppermost layers. Considering thedownward migration of large numbers of microalgae during themorning low tide, it seems possible that a surface populationchange has occurred and that the observed biofilm-level photo-inhibition resulted from the worst physiological condition of themicroalgae remaining in the photic zone during late morning andafternoon. It may be hypothesized that the large productivebiomass observed in the morning corresponds to a largepopulation of motile microalgae that surfaces only during thefirst hours of the day, when irradiance is below damaging levels,the light absorbed during this period being sufficient to satisfytheir carbon needs for the whole day. This would explain theconsiderable changes in the physiological condition of thebiofilms from the morning to the afternoon photoperiods, andalso the apparent overnight recovery of their physiological status,as denoted by the high values of the photophysiologicalparameters observed in early morning.

5. Conclusions

Benthic microalgae living on estuarine intertidal flats mustcope with large and simultaneous changes in various environ-mental factors that affect their photosynthetic performance. Inthe case of diatom-dominated biofilms, the short-term responsesto fast changing environmental conditions include not onlyphysiological processes common to other photoautotrophs butalso microalgal motility, through the regulation of the verticalposition of the cells within the of the photic zone. The obtentionof time series of chlorophyll fluorescence parameters in situ

permitted to realistically characterize both the photophysiologicaland migratory responses of microphytobenthos biofilms duringdiurnal low-tide periods, and to identify a number of aspects,some of them not previously described:

1.

Photosynthetic activity is strongly dependent on recent lighthistory, being severely depressed when the biofilm receives ahigh light dose during morning low tides.

2.

Benthic microalgae undergo photophobic downward migrationunder high light, causing the maximum accumulation of cellsat the sediment surface to occur during the early morninghours.

3.

Microalgae respond to high light by activating energy-dissipating processes based on the xanthophyll cycle, reachinglarge NPQ levels which denote a high capacity for photo-protection.

4.

The light response of most photophysiological parameters maydisplay hysteresis patterns, due to the incomplete recovery ofphotosynthetic activity during the afternoon. The magnitude ofthe hysteresis depends on the morning light history, and

ARTICLE IN PRESS

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–13751374

occurs despite the submersion in high tide during the period ofthe day of higher light stress

5.

The photoprotective mechanisms available to benthic micro-algae, both physiological and behavioural, may not besufficient to prevent the occurrence of photoinhibition.

Acknowledgements

This work was supported by Project POCTI/MAR/15318/99,funded by Fundac-ao para a Ciencia e a Tecnologia. We thank HelenaCoelho for assistance in the field and laboratory work. We thankthree reviewers for their critical comments on the manuscript.

References

Arsalane, W., Rousseau, B., Duval, J.-C., 1994. Influence of the pool size of thexanthophyll cycle on the effects of light stress in diatoms: competitionbetween photoprotection and photoinhibition. Photochemistry and Photobiol-ogy 60, 237–243.

Behrenfeld, M.J., Prasil, O., Babin, M., Bruyant, F., 2004. In search of a physiologicalbasis for covariations in light-limited and light-saturated photosynthesis.Journal of Phycology 40, 4–25.

Brotas, V., Risgaard-Petersen, N., Ottossen, L., Serodio, J., Ribeiro, L., Dalsgaard, T.,2003. In situ measurement of photosynthetic activity and respiration ofintertidal benthic microalgal communities undergoing vertical migration.Ophelia 57, 13–26.

Casper-Lindley, C., Bjorkman, O., 1998. Fluorescence quenching in four unicellularalgae with different light-harvesting and xanthophyll-cycle pigment. Photo-synthesis Research 56, 277–289.

Cook, P.L.M., Røy, H., 2006. Advective relief of CO2 limitation in microphytobenthosin highly productive sandy sediments. Limnology and Oceanography 51,1594–1601.

Cruz, S., Serodio, J., 2008. Relationship of rapid light curves of variable fluorescenceto photoacclimation and non-photochemical quenching in a benthic diatom.Aquatic Botany 88, 256–264.

Dijkman, N.A., Kroon, B.M.A., 2002. Indications for chlororespiration in relation tolight regime in the marine diatom Thalassiosira weissflogii. Journal ofPhotochemistry and Photobiology 66, 179–187.

Falkowski, P.G., Greene, R., Kolber, Z., 1994. Light utilization and photoinhibition ofphotosynthesis in marine phytoplankton. In: Baker, N.R., Bowes, J. (Eds.),Photoinhibition of Photosynthesis: From Molecular Mechanisms to the Field.Bios Scientific, Oxford, pp. 407–432.

Forster, R.M., Kromkamp, J., 2004. Modelling the effects of chlorophyll fluorescencefrom subsurface layers on photosynthetic efficiency measurements in micro-phytobenthic algae. Marine Ecology Progress Series 284, 9–22.

Genty, B., Briantais, J.-M., Baker, N.R., 1989. The relationship between the quantumyield of photosynthetic electron transport and quenching of chlorophyllfluorescence. Biochimica et Biophysica Acta 990, 87–92.

Glud, R.N., Ramsing, N.B., Revsbech, N.P., 1992. Photosynthesis and photosynthesis-coupled respiration in natural biofilms quantified with oxygen microsensors.Journal of Phycology 28, 51–60.

Gorbunov, M.Y., Kolber, Z.S., Lesser, M.P., Falkowski, P.G., 2001. Photosynthesis andphotoprotection in symbiotic corals. Limnology and Oceanography 46, 75–85.

Guarini, J.M., Blanchard, G.F., Gros, P., Gouleau, D., Bacher, C., 2000. Dynamic modelof the short-term variability of microphytobenthic biomass on temperateintertidal mudflats. Marine Ecology Progress Series 195, 291–303.

Honeywill, C., Consalvey, M., Paterson, D.M., 2006. In situ temporal studies ofmicrophytobenthos using pulse-modulated fluorescence. Nova HedwigiaSupplement 130, 265–283.

Jakob, T., Goss, R., Wilhelm, C., 1999. Activation of diadinoxanthin de-epoxidasedue to a chlororespiration proton gradient in the dark in the diatomPhaeodactylum tricornutum. Plant Biology 1, 76–82.

Jesus, B., Perkins, R.G., Consalvey, M., Brotas, V., Paterson, D.M., 2006. Effects ofvertical migrations by benthic microalgae on fluorescence measurements ofphotophysiology. Marine Ecology Progress Series 315, 55–66.

Kromkamp, J., Barranguet, C., Peene, J., 1998. Determination of microphytobenthosPSII quantum efficiency and photosynthetic activity by means of variablechlorophyll fluorescence. Marine Ecology Progress Series 162, 45–55.

Kyle, D.J., Ohad, I., Arntzen, C.J., 1984. Membrane protein damage and repair—

selective loss of a quinone-protein function in chloroplast membranes.Proceedings of the Natural Academy of Sciences USA 81, 4070–4074.

Lavaud, J., Rousseau, B., Etienne, A.-L., 2002. In diatoms, a transthylakoid protongradient is not sufficient to induce a non-photochemical fluorescencequenching. FEBS Letters 523, 163–166.

Lavaud, J., Rousseau, B., Etienne, A.-L., 2004. General features of photoprotection byenergy dissipation in planktonic diatoms (Bacillariophyceae). Journal ofPhycology 40, 130–137.

Lavaud, J., Strzepek, R.F., Kroth, P.G., 2007. Photoprotection capacity differs amongdiatoms: possible consequences on the spatial distribution of diatoms related

to fluctuations in the underwater light climate. Limnology and Oceanography52, 1188–1194.

Levy, O., Dubinsky, Z., Schneider, K., Achituv, Y., Zakai, D., Gorbunov, M.Y., 2004.Diurnal hysteresis in coral photosynthesis. Marine Ecology Progress Series 268,105–117.

Ludden, E., Admiraal, W., Colijn, F., 1985. Cycling of carbon and oxygen in layers ofmarine microphytes—a simulation model and its eco-physiological implica-tions. Oecologia 66, 50–59.

Migne, A., Spilmont, N., Davoult, D., 2004. In situ measurements of benthic primaryproduction during emersion: seasonal variations and annual production in theBay of Somme (eastern English Channel, France). Continental Shelf Research24, 1437–1449.

Migne, A., Gevaert, F., Creach, A., Spilmont, N., Chevalier, E., Davoult, D., 2007.Photosynthetic activity of intertidal microphytobenthic communities duringemersion: in situ measurements of chlorophyll fluorescence (PAM) and CO2

flux (IRGA). Journal of Phycology 43, 864–873.Miles, A., Sundback, K., 2000. Diel variation in microphytobenthic productivity in

areas of different tidal amplitude. Marine Ecology Progress Series 205, 11–22.Muller, P., Li, X.-P., Niyogi, K.K., 2001. Non-photochemical quenching. A response to

excess light energy. Plant Physiology 125, 1558–1566.Olaizola, M., Yamamoto, H.Y., 1994. Short-term response of the diadinoxanthin

cycle and fluorescence yield to high irradiance in Chaetoceros muelleri(Bacillariophyceae). Journal of Phycology 30, 606–612.

Perkins, R.G., Underwood, G.J.C., Brotas, V., Snow, G.C., Jesus, B., Ribeiro, L., 2001.Responses of microphytobenthos to light: primary production and carbohy-drate allocation over an emersion period. Marine Ecology Progress Series 223,101–112.

Pinckney, J., Zingmark, R.G., 1991. Effects of tidal stage and sun angles onintertidal benthic microalgal productivity. Marine Ecology Progress Series 76,81–89.

Platt, T., Gallegos, C.L., Harrison, W.G., 1980. Photoinhibition of photosynthesis innatural assemblages of marine phytoplankton. Journal of Marine Research 38,687–701.

Ralph, P.J., Gademann, R., 2005. Rapid light curves: a powerful tool to assessphotosynthetic activity. Aquatic Botany 82, 222–237.

Ralph, P.J., Gademann, R., Larkum, A.W.D., Schreiber, U., 1999. In situ underwatermeasurements of photosynthetic activity of coral zooxanthellae and otherreef-dwelling dinoflagellate endosymbionts. Marine Ecology Progress Series180, 139–147.

Rijstenbil, J.W., 2003. Effects of UVB radiation and salt stress on growth, pigmentsand antioxidative defence of the marine diatom Cylindrotheca closterium.Marine Ecology Progress Series 254, 37–48.

Rohacek, K., 2002. Chlorophyll fluorescence parameters: the definitions, photo-synthetic meaning, and mutual relationships. Photosynthetica 40, 13–29.

Ruban, A.V., Lavaud, J., Rousseau, B., Guglielmi, G., Horton, P., Etienne, A.-L., 2004.The super-excess energy dissipation in diatom algae: comparative analysiswith higher plants. Photosynthesis Research 82, 165–175.

Schanz, F., Dubinsky, Z., 1988. The afternoon depression in primary productivityin a high rate oxidation pond (Hrop). Journal of Plankton Research 10,373–383.

Schreiber, U., Schliwa, U., Bilger, W., 1986. Continuous recording of photochemicaland nonphotochemical chlorophyll fluorescence quenching with a new type ofmodulation fluorometer. Photosynthesis Research 10, 51–62.

Schreiber, U., Bilger, W., Neubauer, C., 1994. Chlorophyll fluorescence as anonintrusive indicator for rapid assessment of in vivo photosynthesis. In:Shulze, E.D., Caldwell, M.M. (Eds.), Ecophysiology of Photosynthesis. Springer,Berlin, pp. 49–70.

Serodio, J., 2004. Analysis of variable chlorophyll fluorescence in microphyto-benthos assemblages: implications of the use of depth-integrated measure-ments. Aquatic Microbial Ecology 36, 137–152.

Serodio, J., Catarino, F., 2000. Modelling the primary productivity of intertidalmicrophytobenthos: time scales of variability and effects of migratoryrhythms. Marine Ecology Progress Series 192, 13–30.

Serodio, J., Marques da Silva, J., Catarino, F., 2001. Use of in vivo chlorophylla fluorescence to quantify short-term variations in the productivebiomass of intertidal microphytobenthos. Marine Ecology Progress Series218, 45–61.

Serodio, J., Cruz, S., Vieira, S., Brotas, V., 2005. Non-photochemical quenchingof chlorophyll fluorescence and operation of xanthophyll cycle in micro-phytobenthos. Journal of Experimental Marine Biology and Ecology 326,157–169.

Serodio, J., Coelho, H., Vieira, S., Cruz, S., 2006a. Microphytobenthos verticalmigratory photoresponse as characterised by light-response curves of surfacebiomass. Estuarine, Coastal and Shelf Science 68, 547–556.

Serodio, J., Vieira, S., Cruz, S., Coelho, H., 2006b. Rapid light-response curves ofchlorophyll fluorescence in microalgae: relationship to steady-state lightcurves and non-photochemical quenching in benthic diatom-dominatedassemblages. Photosynthesis Research 90, 29–43.

Serodio, J., Vieira, S., Barroso, F., 2007. Relationship of variable chlorophyllfluorescence indices to photosynthetic rates in microphytobenthos. AquaticMicrobial Ecology 49, 71–85.

Sokal, R.R., Rohlf, F.J., 1981. Biometry. W.H. Freeman and Company, New York.Spilmont, N., Davoult, D., Migne, A., 2006. Benthic primary production

during emersion: in situ measurements and potential primary production inthe Seine Estuary (English Channel, France). Marine Pollution Bulletin 53,49–55.

ARTICLE IN PRESS

J. Serodio et al. / Continental Shelf Research 28 (2008) 1363–1375 1375

Ting, C.S., Owens, T.G., 1993. Photochemical and nonphotochemical fluorescencequenching processes in the diatom Phaeodactylum tricornutum. Plant Physiol-ogy 101, 1323–1330.

Underwood, G.J.C., 2002. Adaptations of tropical marine microphytobenthicassemblages along a gradient of light and nutrient availability in Suva Lagoon,Fiji. European Journal of Phycology 37, 449–462.

Underwood, G.J.C., Kromkamp, J., 1999. Primary production by phytoplankton andmicrophytobenthos in estuaries. Advances in Ecology Research 29, 93–153.

Underwood, G.J.C., Perkins, R.G., Consalvey, M.C., Hanlon, A.R.M., Oxborough, K.,Baker, N.R., Paterson, D.M., 2005. Patterns in microphytobenthic primaryproductivity: species-specific variation in migratory rhythms and photosyn-thetic efficiency in mixed-species biofilms. Limnology and Oceanography 50,755–767.

White, A.J., Critchley, C., 1999. Rapid light curves: a new fluorescence method toassess the state of the photosynthetic apparatus. Photosynthesis Research 59,63–72.