Research Article

Effects of macrophytes and terrestrial inputs on fluorescentdissolved organic matter in a large river system

Jean-FranÅois Lapierre1,2 and Jean-Jacques Frenette1,*

1 D�partement de chimie-biologie, Universit� du Qu�bec � Trois-Rivi�res, C.P. 500, Trois-Rivi�res, Qu�bec, QC,G9A 5H7, Canada

2 Present address: D�partement des sciences biologiques, Universit� du Qu�bec � Montr�al, CP 8888, Succ.Centre Ville, Montr�al, QC, H3C 3P8, Canada

Received: 26 May 2008; revised manuscript accepted: 6 January 2009

Abstract. We studied the contribution of aquaticmacrophytes and allochthonous sources to the pool offluorescent dissolved organic matter (FDOM) in alarge river system composed of several distinct watermasses that flow alongside one another in the sameriverbed. Using three dimensional fluorescence com-bined with parallel factor analysis (PARAFAC), wecharacterized FDOM found in the St. Lawrence River(Lake Saint-Pierre, Qu�bec, Canada), and frommacrophyte leaching experiments. Eight fluorescentcomponents were identified, three of which weredominant in macrophyte experiments and were sim-ilar to protein-like, autochthonous fluorophores iden-tified in previous studies. The remaining componentscorresponded to humic and fulvic acids, and aprincipal component analysis revealed that theirdistribution in Lake Saint-Pierre was different thanthat of protein-like fluorophores, suggesting a differ-

ent origin. Concentrations of dissolved organic carbonwere strongly associated with the distribution of theallochthonous components. The distribution of pro-tein-like FDOM in Lake Saint-Pierre matched that ofmacrophytes in the lake and the abundance ofallochthonous FDOM was explained by the connec-tivity with the terrestrial ecosystem. Nearshore watermasses carrying large loads of newly imported organicmatter from proximal tributaries showed the maxi-mum abundances and the older water masses, from thecenter of the lake, carried smaller quantities ofterrestrial organic matter, thus originated mainlyfrom Lake Ontario, several hundred kilometers up-stream of Lake Saint-Pierre. This study demonstratesthat macrophytes are a net source of protein-likeFDOM and could represent an important supply ofautochthonous DOM in shallow, productive environ-ments.

Key words. FDOM; macrophytes; fluvial lake; allochthonous; PARAFAC; river.

Introduction

Riverine ecosystems are fueled by organic carbonoriginating from numerous sources, including in-stream primary producers and the terrestrial land-scape. In shallow aquatic ecosystems, light penetratesmost of the water column, causing favorable growthconditions for macrophytes as well as suspended and

* Corresponding author e-mail:Jean-Jacques.Frenette@uqtr.caPublished Online First: February 16, 2009

Aquat. Sci. 71 (2009) 15 – 241015-1621/09/010015-10DOI 10.1007/s00027-009-9133-2� Birkh�user Verlag, Basel, 2009

Aquatic Sciences

attached algae. In addition to passing organic carbonto consumers, primary producers contribute, throughexcretion and decomposition, to the pool of autoch-thonous dissolved organic matter (DOM). Consider-ing the relatively high bioavailability of this DOM, itcan also contribute to overall lake and river metab-olism by sustaining bacterial activity (Chen andWangersky, 1996; Kritzberg et al. , 2004). Sobek et al.(2007) demonstrated that concentrations of dissolvedorganic carbon (DOC) in lakes are influenced pri-marily by climatic and topographic characteristics andsecondarily by catchment and lake properties. Thissuggests that factors external to the aquatic environ-ment drive the allochthonous loadings of DOM andinstream features dictate the transformation of thisterrestrial organic matter and the cycling of autoch-thonous DOM.

In marine environments, where phytoplankton isthe dominant source of organic matter, sources andforms of fluorescent dissolved organic matter(FDOM) have long been studied, using fluorescenceanalyses as a proxy of DOM (Ferrari and Mingazzini,1995; Coble, 1996; Sondergaard et al. , 2003; Stedmonand Markager, 2005a; among others). In fresh waters,high productivity and the proximity of aquatic toterrestrial ecosystems increase the complexity of theFDOM pool. Recent work has focused on theallochthonous FDOM originating from wetlands andterrestrial ecosystems (Hood et al., 2005; Mladenov etal. , 2007) and on the autochthonous FDOM derivedfrom microbial and phytoplankton sources(McKnight et al. , 2001; Cory and McKnight, 2005;Hood et al., 2005). Macrophytes provide an importantsource of DOM in several environments (Martin etal. , 2005; Fischer et al. , 2006; Maie et al. , 2006), yetlittle is known about the fluorescence properties ofthis DOM, especially in temperate aquatic ecosys-tems. Considering their abundant biomass in shallow,productive environments, macrophytes are likely tocontribute significantly to the pool of FDOM thatfuels the microbial loop. This study uses recentstatistical developments in fluorescence analyses tocharacterize the FDOM originating from macro-phytes and the terrestrial ecosystem in a large riversystem strongly influenced by wetlands, floodplainand tributaries.

Lake Saint-Pierre (lat 46812� N, long 72850� W),Qu�bec, Canada, is the largest (400 km2 with a meandischarge of approximately 10,000 m3s-1) fluvial lakeof the St. Lawrence River. Its bottom is extensivelycovered by large beds of macrophytes that, togetherwith their attached epiphytes, represent the mostimportant primary producers in the system (Vis et al. ,2007). The macrophyte distribution roughly followsthe depth profile of the lake, with maximum abun-

dances in the shallow littoral zone, decreasing towardthe center of the lake. The beds reach their maximumbiomass around mid-August then enter a senescentstate around October, when temperatures and photo-period decrease. The transition of macrophytes fromgrowth to senescence is thus likely to seasonallyinfluence the autochthonous FDOM pool composi-tion.

Lake Saint-Pierre is also strongly connected withthe terrestrial ecosystem through its extensive flood-plain (180 km2) and the confluence of seven tributa-ries with catchment areas that range from 269 km2 to23,720 km2 (mean = 6,430 km2). These contribute tothe formation of up to eight distinct water masses inthe river that flow parallel to one another, with thewater mass in the nearshore zone composed of waterimported from the nearest upstream tributary (seeFrenette et al. , 2006 for a figure illustrating thedistribution of Lake Saint-Pierre water masses). Thecentral water mass carries mainly the water flowingfrom Lake Ontario, situated 250 kilometers upstreamof Lake Saint-Pierre. Because of contrasting land usein their respective catchments, and varying residencetime and average water column depth, the watermasses differ in their light spectral characteristics,nutrient concentrations, dissolved and particulateorganic matter, and inorganic suspended matter(Frenette et al. , 2006), resulting in a strong lateralheterogeneity. The composition of the FDOM pool isthus likely to vary tremendously on a lateral gradientas well as throughout the year due to seasonal changesin the macrophyte life cycle and the hydrodynamicregime.

Methods

SamplingThree sampling surveys were performed during 2004,timed to approximately correspond with the differentstages of the macrophyte life cycle: maximum abun-dance (August 18), senescence (October 7), andsubmerged drifting (November 4). Twenty-sevensites, distributed along three lateral transects compris-ing nine sites each, were sampled on each occasion(Fig. 1) in order to capture the variability in theFDOM pool caused by the different water masses andthe various groups of macrophytes. Two litres ofintegrated depth water samples from the first 1.5 m ofthe water column were collected at each site using a 2-m long, 7-cm diameter polyethylene tube.

DOC and FDOMFrom each site, 200 mL of water was filtered through a45-mm diameter, 0.2-mm nominal pore-size polycar-

16 J.-F. Lapierre and J.-J. Frenette Sources of FDOM in a large river

bonate membrane (Isopore, Millipore). Membraneswere prerinsed with 100 mL of MilliQ water to removepotential impurities. Filtrate was stored in acid-washed borosilicate bottles and kept in the dark at 48C until analysis for FDOM fluorescence and DOCconcentration. DOC concentrations were determinedwith a total organic carbon analyser (OI Analytical,TOC-1010) by sodium persulfate digestion. Absorb-ance and fluorescence scans were performed withintwo days after sample collection.

Macrophyte excretion experimentGrowing specimens of three species of macrophytes inLake Saint-Pierre were collected during early summer2006 to measure the fluorescence signature of FDOMleached during induced senescence conditions. Pota-mogeton richardsonii represented submerged species,being the most abundant representative of that groupwithin the lake (Saint-Cyr and Campbell, 1990; Fortinet al. , 1993). Emergent macrophytes included Sag-ittaria latifolia and Scirpus sp. , which are widelydistributed within this group (Hudon, 1997). Plantswere gently rinsed with distilled water in the labo-ratory, without removing their attached epiphytes.Specimens were allowed to decay separately in thedark for eight days in separate 2-L polyethylenebottles filled with distilled water. FDOM sampleswere collected from the decaying specimens on days 1,2, 4 and 8. No DOC sample was collected, and nobactericide was added, in order to measure all possibleforms of FDOM directly or indirectly originating frommacrophytes under conditions similar to natural ones.

Optical analysesSpectral absorption by the FDOM was measured on adual-beam spectrophotometer (Shimadzu UV-2401

PC) with 1-cm quartz cells in the 190 to 900-nm range.The absorption coefficient a at 375nm was calculatedaccording to the following equation:

aCDOMð375nmÞ ¼2:303Að375nmÞ

l

where A is the absorbance and l is the path length inmeters (Kowalczuk et al. , 2006).

Three-dimensional excitation-emission matrixeswere measured (230 to 600-nm emission wavelengths,2-nm increments; 230 to 450-nm excitation wave-lengths, 5-nm increments, bandwidth = 5 nm for both)on a Varian (Eclipse) fluorometer. The signal-to-noiseratio was very good in this range but deterioratedrapidly below 230 nm. Scans were corrected for inner-filter effect (McKnight et al. , 2001) and standardizedin Raman units (R.U.) (Stedmon et al., 2003). Theywere then corrected for instrument biases using anexcitation correction spectrum derived from a con-centrated solution of rhodamine B and an emissioncorrection spectrum obtained using a ground quartzdiffuser, as recommended by the manufacturer.Raman and Rayleigh scattering were removed fromthe EEMs prior to analysis by replacing these regionsby a diagonal of NaN (not a number). The PLStoolbox (3.5) in Matlab 7.0 was used to performPARAFAC analysis on 254 excitation-emission ma-trixes (EEMs), which included the samples from LakeSaint-Pierre and from the macrophyte excretionexperiment, to identify the fluorescence componentsof the FDOM. The data set also included severalsamples from other lakes, but these were not analyzedin this study. These fluorescence components weresubsequently validated by a split-half analysis (Sted-mon et al. ,2003), and residuals were examined toensure that no systematic variation was present. Theparameters obtained from the PARAFAC model wereused to calculate an approximation of the abundanceof each component, expressed as Fmax (in R.U.), whichcorresponds to the maximum fluorescence intensityfor a particular sample.

Statistical analysesA principal component analysis using a correlationmatrix was performed on Systat (version 11, SSI, SanJose, U.S.) to explore the relationships betweenfluorescence components, optical parameters, andDOC concentrations for samples collected in LakeSaint-Pierre. Samples from excretion experimentswere not included in this analysis. Data were normal-ized prior to analysis. Factors explaining more than10 % of the variance were used.

Figure 1. Distribution of sampling sites along three south–northlateral transects, position 1 being the southernmost site.

Aquat. Sci. Vol. 71, 2009 Research Article 17

Results

Characterisation of fluorescence componentsThe PARAFAC model validated eight fluorescencecomponents (Fig. 2). All components have been iden-tified in previous studies (Table 1). Humic components1 and 7 and protein-like components 3 and 8 appear tobe present in a number of systems, unlike the remainingcomponents (Table 1). All components showed multi-ple excitation peaks for one emission peak, except forcomponent 4, with two emission peaks, and component1, with one excitation peak. Protein-like components 3,4, and 8 differed from the other five components in thatthey exhibited fluorescence in the lower wavelengthsfor both excitation (typically < 300 nm) and emission(typically< 400 nm) and were found in greater relativeabundances in the macrophyte leaching experimentsthan in the natural lake water (Fig. 3). Component 4was particularly dominant in the submerged macro-phyte leaching experiment (Fig. 3).

Component 1, a group of fluorophores that domi-nates several aquatic ecosystems (Table 1), was ap-proximately 2.5 and 8 times more abundant in FDOMfrom natural water than that leached from emergentand submerged macrophytes, respectively (Fig. 3) andrepresented the most abundant FDOM group in LakeSaint-Pierre. Considering that it represented almost5% and 15%, respectively, of the total FDOM poolleached from submerged and emergent macrophytes,and was found in higher proportions in natural water, itprobably originates from both autochthonous andallochthonous sources. It was associated with terrestrialsources of FDOM in previous studies (Stedmon andMarkager, 2005a, and references therein). Components2, 5, 6, and 7 corresponded to previously identifiedhumic and fulvic acids of terrestrial origin. Moreover,their low abundance during the macrophyte leachingexperiments demonstrated the poor contribution ofmacrophytes to the abundance of these components innatural water.

FDOM in Lake Saint-PierreThe distribution of FDOM components in Lake Saint-Pierre illustrates their contrasting origins, as can beseen with the principal component analysis (Fig. 4), inwhich the first two factors explain 83.5% of thevariance of the fluorescence components, aCDOM 375,and DOC. While all components were positivelycorrelated with factor 1, they were well separated onthe axis representing factor 2. This second factorshowed the position of fluorescence componentsalong an autochthonous to allochthonous gradient,with components 3, 4, and 8, shown with aCDOM 375 inthe lower half of factor 2. Component 1, ascribed toboth autochthonous and allochthonous sources, waspositioned in the higher part of the lower quadrantcompared to the solely autochthonous components.Components 2, 5, 6, and 7, found in the upper quadrantof factor 2, represented the allochthonous portion ofthe gradient (Fig. 4). Figure 5 illustrates the autoch-thonous/allochthonous gradient found on axis 2.Despite the fact that the principal component analysisdid not include data from the macrophyte excretionexperiment, there was a direct negative relationshipbetween the correlation of a given component withfactor 2 and its relative contribution to the totalFDOM pool leached by the macrophytes. DOC wasclose to components 6 and 7 in the upper half of factor2. Concentrations of DOC were strongly correlatedwith total fluorescence, and this relationship increasedwhen only allochthonous components (2, 5, 6 and 7)were considered (Fig. 6). It decreased considerablywhen only autochthonous components were consid-ered (Fig. 6).

Lateral variations in FDOMFor clarity, fluorescence components were groupedtogether and summed as described above, with the“autochthonous” group comprising components 3, 4,and 8, and the “allochthonous” group comprisingcomponents 2, 5, 6, and 7. Component 1 is not includedin any group since, as described, it appears to originate

Table 1. Characterisation of fluorescence components identified in this study and their correspondance with previously identifiedcomponents (non-exhaustive list). Secondary maxima in brackets. References: (1) Stedmon and Markager, 2005a; (2) Stedmon andMarkager, 2005b; (3) Cory and McKnight, 2005; (4) Ohno and Bro, 2006; (5) Fellman et al., 2008

Component Excitation maxima(nm)

Emission maxima(nm)

Previously identified componentsReference #

1 2 3 4 5

1 <230 452 C1 C2 C2 C1

2 265 446, 464 C13 <230, 265 302 C8 C4 C13 C5 C94 <230 (280) 306, 476 C8 C1, C4 C95 250 (300) 374 C36 390 (275) 482 C2 C7 C17 340 (240) 422 C5 C9 C3 C48 <230 ( 280) 338 C7 C4, C6 C3 C4 C8

18 J.-F. Lapierre and J.-J. Frenette Sources of FDOM in a large river

from both sources. Autochthonous and allochthonousFDOM, as well as DOC, achieved maximum abun-dances in the nearshore zones (Fig. 7). Higher abun-dances of both autochthonous and allochthonousFDOM were found on the south compared to thenorth shore. Lateral (shore to shore) gradients aremore pronounced for allochthonous components andcomponent 1 compared to autochthonous compo-nents and the DOC follows the same overall pattern asFDOM.

Temporal variations of FDOM and macrophyte lifecyclesOverall, the allochthonous components showed astable pattern throughout the 4-month samplingperiod (Fig. 8). We observed a considerable seasonalshift in the abundance of the autochthonous compo-nents (Fig. 8). This shift was especially pronounced forcomponent 4, which was dominant in the submergedmacrophyte leaching experiment, and showed a netincrease during October and November. Thesemonths correspond to the macrophyte senescence

Figure 2. Fluorescence signatures of the eight fluorescence components identified in this study, presented in order of decreasing percent ofexplained variation. Line plots represent the complete data set (thick line) and split-half validations (thin lines) of the component loadings.

Aquat. Sci. Vol. 71, 2009 Research Article 19

and drifting periods, respectively, in Lake Saint-Pierre. The well-defined, bell-shaped pattern of au-tumn abundances across the lake was not observed forcomponents 3 and 8. Component 8 was found in higheroverall abundances during autumn samplings, but nopattern could be observed for component 3 (Fig. 8).

Discussion

The results of the principal component analysisillustrate that the fluorescence components identifiedin this study were from various origins. There was a

Figure 3. Percent of total FDOM represented by the eightcomponents in the PARAFAC model in natural lake waters,submerged macrophyte leachate, and emergent macrophytesleachate.

Figure 4. Correlation of DOC, aFDOM(375) and FDOM compo-nents with factors 1 and 2 of a principal component analysispeformed only on samples of natural lake water. The first two axesof the analysis explain 83.5% of the variance. Percent of explainedvariation shown in brackets.

Figure 5. Relationship of the percent contribution to the FDOM inleachate from submerged and emergent macrophyte provided byeach component from the PARAFAC model, with the components�correlation with factor 2 of the principal component analysis.

Figure 6. Relationship between DOC and sum of Fmax of the eightfluorescence components identified in this study; relationshipbetween DOC and sum of Fmax for allochthonous components 2, 5,6 and 7 shown in the upper left insert; relationship between DOCand sum of Fmax for autochthonous components 3, 4 and 8 shown inthe upper right

20 J.-F. Lapierre and J.-J. Frenette Sources of FDOM in a large river

strong link between the correlation of fluorescencecomponents with factor 2 and their respective con-tribution to the FDOM pool leached from macro-phytes in the experiments (Fig. 5). Considering that

the principal components analysis did not include datafrom the macrophyte leaching experiments, the resultspresented in figure 5 do not represent an inherentautocorrelation but rather suggest that FDOM pro-duced by autochthonous and allochthonous processesin Lake Saint-Pierre possess contrasting optical prop-erties and behave differently in the system, likely dueto discrepancies in the sources and fate of thesedifferent groups of FDOM.

Components 3, 4, and 8 were the most abundantFDOM constituents leached from macrophytes. Thecombination of components 3 and 8 corresponds tocomponent 4 from Stedmon and Markager (2005b), acomponent associated with protein-like fluorescenceand algal sources. Component 4 is a combination ofcomponents 1 and 4 from Stedmon and Markager(2005b). Their component 1 corresponds to a UV-humic peak associated with both terrestrial andautochthonous sources. The very high dominance ofcomponent 4 in our excretion experiments reveals thatleaching from macrophytes produces humic substan-ces in addition to protein-like material. In our studysite, however, these differ optically from FDOMimported from the terrestrial landscape.

Although most FDOM leached from macrophyteswas similar to that excreted from phytoplankton, it

Figure 7. Lateral variations of abundances of allochthonousFDOM (sum of components 2, 5, 6, and 7), autochthonousFDOM (sum of components 3, 4, and 8), component 1, andDOC; R.U. refers to Raman units. Symbols represent averages �SE for all transects and dates at the same lateral position.

Figure 8. Lateral variations of abundances of FDOM components during periods of maximum abundance (August 18), senescence(October 7), and drifting (November 4) in the macrophyte life cycles. R.U. refers to Raman units. Symbols represent averages acrosssurveys, � SE, for all transects at the same lateral position. Autochthonous fluorescence components are identified by an asterisk.

Aquat. Sci. Vol. 71, 2009 Research Article 21

also contained certain amounts of humic and fulvicacids similar in fluorescence to those exported fromthe terrestrial landscape (Fig. 3, Table 1). This isconsistent with the higher contribution of components1, 2, 5, 6, and 7 to the FDOM leached from emergent asopposed to submerged macrophytes (Fig. 3). Theformer contain higher proportions of structural tissues(Wetzel, 1990; Balogh et al. , 2006), which aid in theformation of humic and fulvic acids in soils (McKnightand Aiken, 1998).

The much higher proportions of components 3, 4and 8 in the macrophyte leaching experiment com-pared to those found in lake water demonstrate thatmacrophytes contribute a net autochthonous FDOMinput in Lake Saint-Pierre. The production of protein-like and humic material from macrophytes suggeststhat both macrophytes and their attached epiphytesproduce autochthonous FDOM, although this couldnot be verified due to our experimental design. Thefact that the leaching experiments were performed inthe dark suggests that photosynthesis is not essentialin order for macrophytes, and possibly algae, to be asource of FDOM in the aquatic ecosystem. Also,bacteria most likely played an important role in theproduction of FDOM during the leaching experi-ments by decomposing the incubated macrophytes. Inany case, FDOM produced under dark conditions inleaching experiments was similar to that associatedwith photosynthetically active phytoplankton. Leach-ing from macrophytes is not the only source ofautochthonous FDOM in aquatic ecosystems, butthe fluorescence signatures of the macrophyte leach-ate match those autochthonous components previ-ously attributed to phytoplankton excretion (Coble,1996; Stedmon et al., 2005b). This illustrates thatdifferent autochthonous processes can produceFDOM with similar optical properties.

FDOM originating from macrophytes and theterrestrial landscape exhibited distinct optical proper-ties in Lake Saint-Pierre, but this could reflect adifferent degree of degradation or a time lag betweenthe formation of FDOM and its arrival in aquaticecosystems. Fellman et al. (2008) identified protein-like fluorophores in appreciable abundances in forestsoils. The authors also found that these fluorophoreswere the most labile compared to other FDOMcomponents, suggesting that protein-like material insoils might be decomposed by bacteria prior toexportation to aquatic environments. This could partlyexplain why allochthonous DOM has been associatedwith recalcitrant humic and fulvic molecules whilealgae, bacteria, and, in this study, macrophytes areassociated with freshly produced autochthonous,protein-like material. Our experimental design didnot allow discrimination between a direct (leaching,

excretion) or indirect (habitat for epiphytes, organicsubstrate for bacterial decomposition) contribution bymacrophytes to FDOM. The incubation conditions,however, were similar to a natural environment(except for light conditions, as discussed earlier)where production and consumption of DOM co-occur. Assuming that consumption of FDOM bybacteria in the leaching experiments was present andrepresentative of natural conditions, we conclude thatmacrophytes contribute to the standing stock ofFDOM, as measured by fluorescence analyses, inLake Saint-Pierre. However, the very strong correla-tion between DOC and allochthonous components, asopposed to autochthonous components (Fig. 6), andthe position of DOC in the top of the upper quadranton factor 2 of the PCA illustrates that DOC concen-trations in lake Saint-Pierre are mainly controlled byinputs of allochthonous DOM, as opposed to leachingfrom macrophytes or other potential autochthonoussources.

Consistent with previous works (Stedmon andMarkager, 2005; Cory and McKnight, 2005; Ohno andBro, 2006; Fellman et al. , 2008), component 1represented the most abundant FDOM group in ourstudy site. Along with appreciable proportions ofcomponent 1 in the macrophyte leaching experiments,its position on factor 2 (Fig. 4) illustrates that bothautochthonous and allochthonous inputs of FDOMdetermine its concentration in natural waters. Fur-thermore, the higher abundances of component 1 innatural water compared to that in leaching experi-ments suggest the presence of a non-macrophyte, mostlikely terrestrial (Stedmon and Markager 2005a),source of this component. If macrophytes were theonly source, one would expect lower concentrations ofcomponent 1 in natural water because additionalinputs of FDOM arriving from tributaries wouldgradually diminish the proportion of component 1 inthe total FDOM pool.

The FDOM characteristics in Lake Saint-Pierrewere influenced by the connectivity with the flood-plain and inflowing tributaries, each draining itsrespective catchment. In the nearshore zone, whereallochthonous FDOM concentrations reached theirmaxima, the tributaries drain large catchments thatare heavily impacted by human activities such asagriculture (Frenette et al. , 2006) and flow directlyinto the fluvial lake. Water masses in the center of thelake (positions 4, 5, and 6), where the lowest amountsof allochthonous FDOM were found, originate fromthe Ottawa River and Lake Ontario, both situatedover 250 km upstream (Frenette et al. , 2006); LakeOntario�s hydraulic residence time is around 6 years.Recent findings showed that, in systems with highresidence times, concentrations and bioreactivity of

22 J.-F. Lapierre and J.-J. Frenette Sources of FDOM in a large river

FDOM are markedly lower at the outlet compared tothe inlet (Mari et al. , 2007). Similarly, rivers carrysmaller quantities of FDOM when a headwater lake issituated upstream because of in-situ processing ofhumic and fulvic acids (Larson et al. , 2007). Ourresults, therefore, suggest that the concentration andcomposition of allochthonous FDOM in Lake Saint-Pierre is a function of a) DOM loading from thefloodplain and its tributaries, and b) in situ trans-formation of DOM during transport from sourcewaters to Lake Saint-Pierre. Thus, we expect thattributaries deliver large quantities of young allochth-onous FDOM, which will likely be more bioreactivecompared to that found in the older water masses atthe center of the lake.

Autochthonous FDOM represented roughly 20 %(averaged for all samplings) of the total FDOM poolin Lake Saint-Pierre based on the fluorescencemeasurements. The distribution of autochthonousFDOM was strongly linked to the distribution ofmacrophytes (Jean Morin, unpublished data), to thedepth and phosphorus patterns in the system (Hudonand Carignan, 2008) and to primary productivitypatterns (Vis et al. , 2007). Although the lateraldistribution of autochthonous and allochthonousFDOM components were somewhat similar withmaxima in the nearshore zones (Fig. 7), our attribu-tion of a fluorescence component to an autochthonousor allochthnous origin is consistent with its presumedsource, i.e. primary production or input from tribu-taries or the floodplain. The principal componentsanalysis (Fig. 4) further illustrated the contrastingdistribution of the different FDOM components,which was explained by differences in their sources,as shown in figure 5. The increase in autochthonousFDOM during October and November suggests thatsenescence, and thus decomposition of macrophytes,also constitutes an important input of autochthonousFDOM in Lake Saint-Pierre and, potentially down-stream, because of the autumn drift of dead plantswith advected water masses.

Even though macrophytes do not represent a maincarbon source for herbivores or direct grazers (Miller,2004), they contribute strongly to the detritivores�food web through carbon and nutrient cycling and bysupporting bacterial consumers. Macrophytes aremajor constituents of freshwater and marine wetlandsand dominate in shallow and productive environmentssuch as fluvial Lake Saint-Pierre. These plant com-munities are largely influenced by factors such aswater level (Hudon et al., 2005) and nutrient enrich-ment (Hudon, 2004), and thus, ultimately, by humanactivities. Along with increasing eutrophication ofinland waters, the impact of macrophytes as a majorcarbon source for microbes is likely to increase in

years to come as their abundances and distributionsgrow.

Acknowledgments

We are very grateful to Marianne Lefebvre, DanielPouliot, Patrice Thibeault and Pourvoirie Gladu fortheir help in the field and in the laboratory. Thanks toPhilippe Massicotte and three anonymous reviewersfor detailed, constructive reviews of this paper. TheNatural Sciences Research Council of Canada(NSERC) and the Fonds Qu�b�cois de la Recherchesur la Nature et les Technologies (FQRNT) providedfunds to J.-J. F. for this research, as did a postgraduateNSERC fellowship to J.-F. L. This is a contributionfrom the “Groupe de Recherche Interuniversitaire enLimnologie” (GRIL).

References

Balogh, K. V., M. Presing, L. Voros and N. Toth, 2006. A study ofthe decomposition of reed (Phragmites australis) as a possiblesource of aquatic humic substances by measuring the naturalabundance of stable carbon isotopes. International Review ofHydrobiology 91: 15–28.

Chen, W. and P. J. Wangersky, 1996. Rates of microbial degradationof dissolved organic carbon from phytoplankton cultures.Journal of Plankton Research 18: 1521–1533.

Coble, P. G., 1996. Characterisation of marine and terrestrial DOMin seawater using excitation-emission matrix spectroscopy.Marine Chemistry 51: 325–346.

Cory, R. M. and D. M. McKnight, 2005. Fluorescence spectroscopyreveals ubiquitous presence of oxidized and reduced quinonesin dissolved organic matter. Environmental Science & Tech-nology 39: 8142–8149.

Fellman, J. B., D. V. D�Amore, E. Hood and R. D. Boone., 2008.Fluorescence characteristics and biodegradability of dissolvedorganic matter in forest and wetland soils from coastaltemperate watersheds in southeast Alaska. Biogeochemistry88: 169–184.

Ferrari, G. M. and M. Mingazzini, 1995. Synchronous fluorescencespectra of dissolved organic matter (DOM) of algal origin inmarine coastal waters. Marine Ecology Progress Series 125:305–315.

Fischer, H., C. Mille-Lindblom, E. Zwirnmann and L.J. Tranvik,2006. Contribution of fungi and bacteria to the formation ofdissolved organic carbon from decaying common reed (Phrag-mites australis). Archiv Fur Hydrobiologie 166: 79 –97.

Fortin, G.R., L. Saint-Cyr and M. Leclerc, 1993. Distribution ofsubmersed macrophytes by echo-sounder tracings in LakeSaint-Pierre, Quebec. Journal of Aquatic Plant Management31: 232–240.

Frenette, J.-J., M. T. Arts, J. Morin, D. Gratton and C. Martin, 2006.Control of the underwater light climate in fluvial Lac Saint-Pierre. Limnology and Oceanography 51: 2632–2645.

Hood, E., M. W. Williams and D.M. McKnight, 2005. Sources ofdissolved organic matter (DOM) in a Rocky Mountain streamusing chemical fractionation and stable isotopes. Biogeochem-istry 74: 231–255.

Hudon, C., 1997. Impact of water level fluctuations on St.Lawrence River aquatic vegetation. Canadian Journal ofFisheries and Aquatic Sciences 54: 2853–2865.

Hudon, C. 2004. Shift in wetland plant composition and biomass

Aquat. Sci. Vol. 71, 2009 Research Article 23

following low-level episodes in the St. Lawrence River:looking into the future. Canadian Journal of Fisheries andAquatic Sciences 61: 603–617.

Hudon, C. and R. Carignan., 2008. Cumulative impacts ofhydrology and human activities on water quality in the St.Lawrence River (Lake Saint-Pierre, Quebec, Canada). Cana-dian Journal of Fisheries and Aquatic Sciences 65: 1165–1180.

Hudon C., P. Gagnon, J.-P. Amyot, G. L�tourneau, M. Jean, C.Plante, D. Rioux and M. DeschÞnes, 2005. Historical changes inherbaceous wetlands distribution induced by hydrologicalconditions in Lake Saint-Pierre (St. Lawrence River, Quebec,Canada). Hydrobiologia 539: 205–224.

Kowalczuk, P., C.A. Stedmon and S. Markager, 2006. Modelingabsorption by CDOM in the Baltic Sea from season, salinityand chlorophyll. Marine Chemistry 101: 1–11.

Kritzberg, E. S., J. J. Cole, M.L. Pace, W. Gran�li and D.L. Bade,2004. Autochthonous versus allochthonous carbon sources ofbacteria: Results from whole-lake 13C addition experiments.Limnology and Oceanography 49: 588–596.

Larson, J. H., P. C. Frost, Z. Y. Zheng, C.A. Johnston, S. D.Bridgham, D. M. Lodge and G.A. Lamberti, 2007. Effects ofupstream lakes on dissolved organic matter in streams.Limnology and Oceanography 52: 60 –69.

Maie, N., R. Jaff�, T. Miyoshi and D. L. Childers., 2006.Quantitative and qualitative aspects of dissolved organiccarbon leached from senescent plants in an oligotrophicwetland. Biogeochemistry 78: 285–314.

Mari, X., E. Rochelle-Newall, J. P. Torreton, O. Pringault, A. Jouonand C. Migon, 2007. Water residence time: A regulatory factorof the DOM to POM transfer efficiency. Limnology andOceanography 52: 808–819.

Martin, C., J.-J. Frenette and J. Morin, 2005. Changes in the spectraland chemical properties of a water mass passing throughextensive macrophyte beds in a large fluvial lake (Lake Saint-Pierre, Quebec, Canada). Aquatic Sciences 67: 196–209.

Miller, B. C., 2004. Biological Oceanography, Blackwell Publish-ing, 402 pp.

McKnight, D. M. and G. R. Aiken, 1998. Sources and age of aquatichumus. In D. O. Hessen and L. J. Tranvik (eds.), Aquatic humicsubstances: Ecology and Biogeochemistry, Springer Verlag,New York, pp. 9–39.

McKnight, D. M., E. W. Boyer, P. K. Westerhoff, P.T. Doran, T.Kulbeand D. T. Andersen, 2001. Spectrofluorometric charac-terization of dissolved organic matter for indication ofprecursor organic material and aromaticity. Limnology andOceanography 46: 38–48.

Mladenov, N., D. M. McKnight, S. A. Macko, M. Norris, R. M. Coryand L. Ramberg, 2007. Chemical characterisation of DOM inchannels of a seasonal wetland. Aquatic Sciences 69: 456–471.

Ohno, T. and R. Bro., 2006. Dissolved organic matter character-ization using multiway spectral decomposition of fluorescencelandscapes. Soil Science Society of America Journal 70: 2028–2037.

Saint-Cyr, L. and P. G. C. Campbell, 1990. R�sultats de biomassedes plantes aquatiques submerg�es du lac Saint-Pierre, fleuveSaint-Laurent, obtenus lors de I��chantillonnage du moisd�aout 1990. INRS ETE, Sainte-Foy, rapport d��tape no 1.

Sobek, S., L. J. Tranvik, Y. T. Prairie, P. Kortelainen and J.J. Cole,2007. Patterns and regulation of dissolved organic carbon: Ananalysis of 7 500 widely distributed lakes, Limnology andOceanography 52: 1208–1219.

Sondergaard M., C. A. Stedmon and N. H. Borch, 2003. Fate ofterrigenous dissolved organic matter (DOM) in estuaries:Aggregation and bioavailability. Ophelia 57: 161–176.

Stedmon, C. A., S. Markager and R. Bro, 2003. Tracing dissolvedorganic matter in aquatic environments using a new approachto fluorescence spectroscopy. Marine Chemistry 82: 239–254.

Stedmon, C. A. and S. Markager, 2005a. Resolving the variability indissolved organic matter fluorescence in a temperate estuaryand its catchment using PARAFAC analysis. Limnology andOceanography 50: 686–697.

Stedmon, C. A. and S. Markager, 2005b. Tracing the production anddegradation of autochthonous fractions of dissolved organicmatter by fluorescence analysis. Limnology and Oceanography50: 1415–1426.

Vis, C., C. Hudon, R. Carignan and P. Gagnon, 2007. SpatialAnalysis of production by macrophytes, phytoplankton andepiphyton in a large river system under different water-levelconditions. Ecosystems 10: 293–310.

Wetzel, R. G., 1990. Land-water interfaces: Metabolic andlimnological regulators. Internationale Vereinigung for theor-etische und angewandte Limnologie. Verhandlungen 24: 6–24.

To access this journal online:http://www.birkhauser.ch/AS

24 J.-F. Lapierre and J.-J. Frenette Sources of FDOM in a large river