PRODUCTION BACTÉRIENNE ET STRUCTURE DU RÉSEAU … · Mme Patricia Ramlal (Freshwater Institute, Department of Fisheries and Oceans, Winnipeg, Canada) ... Merci aussi à M. Julian

CATHERINE VALLIÈRES PRODUCTION BACTÉRIENNE ET STRUCTURE DU

RÉSEAU ALIMENTAIRE MICROBIEN DANS LE FLEUVE MACKENZIE ET L’OCÉAN ARCTIQUE

CÔTIER

Mémoire présenté

à la Faculté des études supérieures de l’Université Laval dans le cadre du programme de maîtrise en biologie

pour l’obtention du grade de maître ès sciences (M.Sc.)

FACULTÉ DES SCIENCES ET DE GÉNIE UNIVERSITÉ LAVAL

QUÉBEC

2007 © Catherine Vallières, 2007

Résumé Les pergélisols nordiques contiennent une quantité significative de carbone organique. Or,

son utilisation par les microorganismes dans les rivières et l’océan Arctique est peu connue.

Au cours de l’Arctic River-Delta Experiment (ARDEX), des variables environnementales et

microbiologiques furent mesurées le long d’un transect de 300 km entre le fleuve

Mackenzie et la mer de Beaufort (juillet-août 2004). Nous voulions étudier les gradients

fluviaux et estuariens de la structure et de l’activité de la communauté microbienne et

évaluer l’influence des UV et de l’approvisionnement en carbone sur les processus

bactériens. La communauté microbienne a changé le long du transect et les bactéries

attachées aux particules jouaient un rôle majeur dans le fleuve et la zone de transition. Le

métabolisme bactérien était limité par la disponibilité du carbone dans le fleuve Mackenzie.

La photodégradation a augmenté la labilité du carbone organique dans le fleuve, mais l’a

diminuée dans la mer de Beaufort.

Abstract Globally significant quantities of organic carbon are stored in northern permafrost soils, but

little is known about how this carbon is processed by microbial communities once it enters

rivers and is transported to the coastal Arctic Ocean. As part of the Arctic River-Delta

Experiment (ARDEX), we measured environmental and microbiological variables along a

300 km transect across the Mackenzie River and coastal Beaufort Sea in July-August 2004

to investigate the river and estuarine gradients in microbial community structure and

activity, and to evaluate the influence of UV exposure and carbon supply on bacterial

processes in these ecosystems. Microbial community structure changed along the transect

and the contribution of particle-attached bacteria was significantly higher in riverine and

transition zone stations. Experimental results showed that bacterial metabolism was carbon

limited in the Mackenzie River. Photodegradation increased organic carbon biolability in

the Mackenzie River and decreased it in the Beaufort Sea.

Avant-propos Ce mémoire comporte une introduction générale à la matière étudiée suivie de deux

chapitres rédigés en anglais sous forme d’articles scientifiques et, finalement, d’une

conclusion générale. Les résultats du premier chapitre ainsi qu’une partie du deuxième

chapitre seront éventuellement soumis pour publication à un journal scientifique sous le

titre «Bacterial production and microbial food web structure in a large arctic river and the

coastal Arctic Ocean». J’écrirai cet article avec les précieux conseils de mon directeur, M.

Warwick F. Vincent. Mme Leira Retamal (département de biologie de l’Université Laval et

Centre d’études nordiques, Québec, Canada) sera co-auteure de cet article pour m’avoir

gracieusement prêté ses données de chlorophyll a et de matière en suspension et dissoute

du fleuve Mackenzie et de la mer de Beaufort. Mme Patricia Ramlal (Freshwater Institute,

Department of Fisheries and Oceans, Winnipeg, Canada) sera également co-auteure pour

ses mesures de dioxyde de carbone effectuées lors de la mission ARDEX. M. Chris Osburn

(National Research Council-Naval Research Laboratory, Washington, DC) sera l’autre co-

auteur pour ses expériences de photodégradation du carbone organique dissous.

Les résultats récoltés lors de la croisière sur le fleuve Mackenzie et la mer de Beaufort ont

été présentés oralement lors de l’atelier de l’Arctic River Delta Experiment (ARDEX) en

mars 2005 à Québec et sous forme d’affiche à l’occasion de l’atelier du Canadian Shelf

Exchange Study (CASES) en février 2006 à Winnipeg. Ces résultats ont aussi été présentés

sous forme d’affiche au congrès estival de l’American Society of Limnology and

Oceanography (ASLO) en juin 2006 à Victoria. Pour cette affiche, j’ai reçu l’une des 10

mentions Outstanding Student Poster Award qui étaient remises aux affiches efficacement

et clairement organisées, montrant des innovations et des révélations scientifiques et

démontrant un design expérimental et des méthodes de haute qualité.

Pour m’avoir permis de mener à bien mon projet de maîtrise, je désire témoigner ma

reconnaissance à tous les organismes et les personnes ayant supporté financièrement,

logistiquement et moralement celui-ci. Grâce à eux, j’ai pu participer à deux magnifiques

missions de terrain qui furent formatrices autant du point de vu scientifique qu’humain.

iv

En tout premier lieu, je veux exprimer ma gratitude envers mon directeur de projet, M.

Warwick F. Vincent, pour m’avoir fait confiance et qui a toujours su m’encourager au

cours de toutes les étapes de mes études de maîtrise. Je tiens à remercier Mme Isabelle

Laurion, ma co-directrice, pour ses conseils et pour avoir donné un temps précieux pour la

lecture de mon mémoire. Merci aussi à M. Julian Dodson pour ses commentaires.

Merci à l’appui financier du Conseil de recherche en sciences naturelles et en génie du

Canada (CRSNG), du Centre d’études nordiques (CEN), du Programme de formation pour

scientifiques dans le nord (PFSN) du ministère des Affaires indiennes et du Nord Canadien

et du département de biologie de l’université Laval.

Merci au soutien logistique de l’Aurora Research Institute d’Inuvik, de la Garde côtière du

Canada et de la station de recherche du CEN à Kuujjuarapik-Whapmagoostui. Je remercie

le personnel de ces institutions pour avoir aidé, de près ou de loin, au bon déroulement de

mon projet. Je pense entre autre à l’équipage du NGCC Nahidik et à M. Claude Tremblay

de la station de Kuujjuarapik-Whapmagoostui pour leur aide de tous les jours. Je tiens

également à souligner l’incontournable participation de M. Laurent-Étienne Desgagnés sans

qui je n’aurais pas pu échantillonner dans la Baie d’Hudson. Merci à M. Chris Osburn et à

Mme Patricia Ramlal pour leur collaboration lors de la mission ARDEX. Merci également

à M. Craig Emmerton pour ses nombreuses données de nutriments du fleuve Mackenzie et

de la mer de Beaufort.

Un merci tout spécial aux membres des laboratoires Vincent et Laurion! Merci à Leira

Retamal qui a toujours été là du début à la fin pour répondre à mes innombrables questions.

Merci à Milla Rautio qui m’a apporté une aide inestimable sur le NGCC Nahidik. Merci à

Julie Breton pour sa participation à mon échantillonnage de la Grande Rivière de la

Baleine. Merci à Marie-Ève Garneau pour m’avoir appris les rudiments des techniques de

production bactérienne. Merci à Marie-Josée Martineau et à Christine Martineau, nos

inestimables techniciennes, et à tous les autres membres du labo qui ont fait partie de ma

vie quotidienne à l’université.

Finalement, merci à ma famille et à mes amis pour tout le reste!

Table des matières RÉSUMÉ ...................................................................................................................................II ABSTRACT ...............................................................................................................................II AVANT-PROPOS ......................................................................................................................III TABLE DES MATIÈRES ............................................................................................................. V LISTE DES FIGURES................................................................................................................ VII LISTE DES TABLEAUX............................................................................................................. XI LISTE DES ANNEXES .............................................................................................................. XII INTRODUCTION GÉNÉRALE.....................................................................................................14 CHAPTER 1: GRADIENTS IN BACTERIAL PRODUCTION AND MICROBIAL COMMUNITY STRUCTURE IN THE MACKENZIE RIVER AND ITS ESTUARY .....................................................37

Résumé..............................................................................................................................37 Abstract.............................................................................................................................37 Introduction.......................................................................................................................38 Methods ............................................................................................................................40

Sampling .......................................................................................................................40 Physical characteristics of the water column ...............................................................40 Particulate and dissolved material ...............................................................................40 Chlorophyll a ................................................................................................................41 Microbial community structure ....................................................................................41 Heterotrophic prokaryote production...........................................................................44 Statistical analysis ........................................................................................................45

Results...............................................................................................................................46 Sampling conditions and meteorological data .............................................................46 Hydrographic gradients................................................................................................47 Environmental characteristics and gradients...............................................................47

Surface ......................................................................................................................47 Bottom ......................................................................................................................51 Surface versus bottom waters ...................................................................................53

Microbial population gradients ....................................................................................53 Heterotrophic picoplankton ......................................................................................53 Picophytoplankton ....................................................................................................53 Protists ......................................................................................................................54 Total microbial community ......................................................................................56

Heterotrophic activity gradients...................................................................................56 Overall relationships at the surface between microbial and environmental characteristics...............................................................................................................57

Discussion.........................................................................................................................58 Spatial gradients ...........................................................................................................59 Microbial community structure ....................................................................................61 Heterotrophic picoplankton 3H-leucine uptake ............................................................65

Bacterial carbon cycling ...........................................................................................67 Metabolic balance of the system...................................................................................69 Implications for the future of the Arctic........................................................................71

vi

CHAPTER 2: FACTORS CONTROLLING BACTERIAL PRODUCTION IN TWO HIGH LATITUDE RIVERS..............................................................................................................................................95

Résumé..............................................................................................................................95 Abstract.............................................................................................................................95 Introduction.......................................................................................................................96 Methods ............................................................................................................................97

Mackenzie River and Beaufort Sea...............................................................................97 Great Whale River and Hudson Bay.............................................................................98

Physical and chemical characteristics.......................................................................98 Bacterial production..................................................................................................99

Mackenzie River experimental design ........................................................................100 Carbon limitation ....................................................................................................100 Photodegradation effect on carbon biolability........................................................101

Great Whale River experimental design.....................................................................101 Carbon and phosphorus limitation..........................................................................101

Results.............................................................................................................................101 Mackenzie River and Beaufort Sea.............................................................................101 Great Whale River and Hudson Bay...........................................................................102

Sampling conditions ...............................................................................................102 Environmental characteristics and gradients ..........................................................102

Response of bacterial activity to glucose addition in the Mackenzie River / Beaufort Sea ecosystems............................................................................................................103 Response of bacterial activity to the exposure of DOC to sunlight in the Mackenzie River / Beaufort Sea ecosystems .................................................................................103 Response of bacterial activity to glucose and phosphorus addition in the Great Whale River / Hudson Bay ecosystems ..................................................................................104

Discussion.......................................................................................................................104 Effect of carbon on bacterial metabolism...................................................................105 Effect of phosphorus addition on bacterial metabolism .............................................105 Effect of photochemistry on bacterial metabolism .....................................................106 Comparison of Mackenzie River and Great Whale River...........................................107

CONCLUSION GÉNÉRALE ......................................................................................................115 RÉFÉRENCES ........................................................................................................................118 ANNEXES .............................................................................................................................126

Liste des figures INTRODUCTION Figure I.1 Le réseau alimentaire microbien et ses composantes. Traduit et adapté de

Vincent et Hobbie (2000). ............................................................................................32 Figure I.2 Illustration du continuum de taille de la matière organique retrouvée dans les

milieux aquatique et marin et les différentes méthodes d’isolement. Tiré et traduit de Hedges (2002)...............................................................................................................32

Figure I.3 Représentation idéalisée des taux de décomposition de biomolécules importantes. Tiré et traduit de Canfield (2005). ...........................................................33

Figure I.4 Représentation de l’océan Arctique et de ses principaux affluents. La largeur des flèches est proportionnelle à la décharge en DOC qui est donnée en 1012 g C an-1. Tiré de Dittmar et Kattner (2003).........................................................................................33

Figure I.5 Illustration des changements de l’anomalie de décharge des rivières arctiques, de l’index NAO (North Atlantic Oscillation) hivernal et de la température moyenne de l’air de surface globale (SAT) de 1936 à 1999. Tiré de Peterson et al. (2002). ...........34

Figure I.6 Cycle hydrologique du fleuve Mackenzie montrant son apport d’eau dans l’océan Arctique pour les années 1973 à 1990. Tiré de Macdonald et al. (1998). La zone grise montre la période de la croisière d’échantillonnage ARDEX 2004. La ligne pointillée montre le débit moyen du fleuve Saint-Laurent à Québec (Vincent et Dodson 1999)................................................................................................................34

Figure I.7 Situation géographique du bassin de drainage du fleuve Mackenzie. Tiré du site Internet Mackenzie River Basin Board (Online)...........................................................35

Figure I.8 Détail du delta du Mackenzie. Tiré de Emmerton (2006). .................................35 Figure I.9 Illustration de la distribution de salinité dans l’estuaire du Mackenzie en été. La

zone ombrée montre la distribution typique des sédiments en suspension associés avec la plume du Mackenzie et la re-suspension du fond. Tiré de Carmack et Macdonald (2002)............................................................................................................................36

Figure I.10 La Grande Rivière de la Baleine. ......................................................................36 CHAPTER 1 Figure 1.1 Sampling site and stations. White dots: Mackenzie River stations. Black dots:

transition zone stations. Gray dots: Beaufort Sea stations. The picture shows the 20 m and 50 m isobaths. ........................................................................................................78

Figure 1.2 Tests for the determination of the sonication length which optimized the bacterial counts at two different stations. .....................................................................79

Figure 1.3 Saturation curve and temporal series for the calibration of the 3H-leucine uptake measurements. The measurements have been made at the riverine station R3. The error bars represent the standard deviation for each measurements (n = 3). The dotted lines show the 3H-leucine concentration and the incubation length used during the present study..............................................................................................................................79

Figure 1.4 Mean air temperature and precipitations that occurred during the ARDEX cruise sampling period. Data are shown for three sites. Norman Wells is located upstream of Inuvik and is shown in order to indicate the weather in the southern part of the watershed. The dashed lines show the beginning of the cruise. Data are from the

viii

National Climate Data and Information Archives of Environment Canada (Online-a).......................................................................................................................................80

Figure 1.5 Mackenzie River discharge between July 26th and August 5th 2004 at two different stations. Data are from the Water survey of Canada of Environment Canada (Online-b)......................................................................................................................81

Figure 1.6 Water column characteristics of the stations. The dashed lines show the sampling depths. S: Salinity curve; C: Chl a curve; T: temperature curve. **Note that the Chl a fluorometer was not calibrated and that the data are given in relative units.82

Figure 1.7 Surface water properties at each station along ARDEX transect. ......................83 Figure 1.8 Bottom water properties at each station along ARDEX transect. ......................84 Figure 1.9 Salinity and temperature data separated according to sampling zones (River, TZ

and Sea) and depths (Surface, Bottom). Plots entitled SURFACE and BOTTOM compare the average value of each zone for one depth. Those entitled RIVER, TRANSITION ZONE and SEA compare the average value of each depth for one zone. The errors bars represent the standard deviations of the data. The letters show the results of the multiple comparison tests. n.s. = no significant difference. * = significant difference (P ≤ 0.05). ** = highly significant difference (P ≤ 0.01). .................................................85

Figure 1.10 Turbidity, SPM, POM and <3 µm Chl a / total Chl a ratio data separated according to sampling zones (River, TZ and Sea) and depths (Surface, Bottom). Plots entitled SURFACE and BOTTOM compare the average value of each zone for one depth. Those entitled RIVER, TRANSITION ZONE and SEA compare the average value of each depth for one zone. The errors bars represent the standard deviations of the data. The letters show the results of the multiple comparison tests. n.s. = no significant difference. * = significant difference (P ≤ 0.05). ** = highly significant difference (P ≤ 0.01). .............................................................................................................................86

Figure 1.11 DOC, POC, PON, Total Chl a and <3 µm Chl a data separated according to sampling zones (River, TZ and Sea) and depths (Surface, Bottom). Plots entitled SURFACE and BOTTOM compare the average value of each zone for one depth. Those entitled RIVER, TRANSITION ZONE and SEA compare the average value of each depth for one zone. The errors bars represent the standard deviations of the data. The letters show the results of the multiple comparison tests. n.s. = no significant difference. * = significant difference (P ≤ 0.05). ** = highly significant difference (P ≤ 0.01). ...87

Figure 1.12 Picophytoplankton abundance along the ARDEX transect between the Mackenzie River and the Arctic Ocean. .......................................................................88

Figure 1.13 Bacteria and picophytoplankton data separated according to sampling zones (River, TZ and Sea) and depths (Surface, Bottom). Plots entitled SURFACE and BOTTOM compare the average value of each zone for one depth. Those entitled RIVER, TRANSITION ZONE and SEA compare the average value of each depth for one zone. The errors bars represent the standard deviations of the data. The letters show the results of the multiple comparison tests. n.s. = no significant difference. * = significant difference (P ≤ 0.05). ** = highly significant difference (P ≤ 0.01). ...........................89

Figure 1.14 Protist community structure and abundance across the freshwater-saltwater transition zone of the Mackenzie River and Beaufort Sea. Crypt.: Cryptophyceae; Chloro.: Chlorophyceae; Bacillario.: Bacillariophyceae; Dino.: Dinophyceae; Chryso.: Chrysophyceae; Raphido.: Raphidophyceae; Eugleno.: Euglenophyceae; Prymnesio.: Prymnesiophyceae; Prasino.: Prasinophyceae; Choano.: Choanoflagellates; Flag.:

ix

Flagellates; Auto.: Autotroph or autotrophic; Hete.: Heterotroph or heterotrophic; Ident.: Identified; and Unident.: Unidentified. .............................................................90

Figure 1.15 Protist community structure and biomass across the freshwater-saltwater transition zone of the Mackenzie River and Beaufort Sea. Crypt.: Cryptophyceae; Chloro.: Chlorophyceae; Bacillario.: Bacillariophyceae; Dino.: Dinophyceae; Chryso.: Chrysophyceae; Raphido.: Raphidophyceae; Eugleno.: Euglenophyceae; Prymnesio.: Prymnesiophyceae; Prasino.: Prasinophyceae; Choano.: Choanoflagellates; Flag.: Flagellates; Auto.: Autotroph or autotrophic; Hete.: Heterotroph or heterotrophic; Ident.: Identified; and Unident.: Unidentified. .............................................................91

Figure 1.16 Protist community and total microbial community (protists and picoplankton) structure in terms of abundance and carbon biomass. Nano. : Nanoplankton. Micro. : Microplankton. Picophyto. : Picophytoplankton. Pico. : Picoplankton. * Heterotrophic picoplankton data missing. ...........................................................................................92

Figure 1.17 Heterotrophic picoplankton production as 3H-leucine uptake rates. Top panels: Total and <3 µm 3H-leucine uptakes (pmol L-1 h-1). Bottom panels: Percentage of the total 3H-leucine uptake that is due to particulate-attached bacteria (>3 µm). The error bars represent the standard deviation for each measurement (n = 3)............................93

Figure 1.18 Total 3H-leucine uptake, <3 µm 3H-leucine uptake and >3 µm / total 3H-leucine uptake ratio data separated according to sampling zones (River, TZ and Sea) and depths (Surface, Bottom). Plots entitled SURFACE and BOTTOM compare the average value of each zone for one depth. Those entitled RIVER, TRANSITION ZONE and SEA compare the average value of each depth for one zone. The errors bars represent the standard deviations of the data. The letters show the results of the multiple comparison tests. n.s. = no significant difference. * = significant difference (P ≤ 0.05). ** = highly significant difference (P ≤ 0.01)................................................................94

CHAPTER 2 Figure 2.1 Sampling sites in the Great Whale River. Between EST and BAY there was a

salinity front. The BAY site was more saline than the EST station. ..........................111 Figure 2.2 Response of bacterial activity to glucose addition as the ratio of treatment 3H-

leucine uptake to control in the Mackenzie River, estuary and the Beaufort Sea. The error bars represent the standard deviation of the two ratios. * Significantly >1 with P < 0.05. ** Significantly >1 with P < 0.01. .................................................................112

Figure 2.3 Bacterial 3H-leucine uptake in filtered water (0.2 µm) that had been exposed to sunlight for 3 days and subsequently inoculated with a bacterial community. Each bar represents one bottle. Error bars represent the analytical standard deviation of 3 replicates for 3H-leucine uptake measurements in each bottle. ..................................113

Figure 2.4 Response of bacterial activity to glucose addition (C), phosphorus addition (P) and simultaneous glucose and phosphorus addition (C+P) as the ratio of treatment to control 3H-leucine uptake in the Great Whale River and estuary. The letters a, b and c are the results of ANOVA and Fisher-LSD tests run to compare the different treatments (control, glucose, phosphorus, glucose + phosphorus) at each station. n.s. means that no significant difference was detected between the 4 treatments. Different letters correspond to significant difference between the treatments (P < 0.05). In GRB1 the glucose addition treatment was not significantly different from the control.....................................................................................................................................114

x

CONCLUSION Figure C.1 Schéma récapitulatif illustrant le destin du carbone organique dissous (DOC)

dans le milieu aquatique. ............................................................................................117

Liste des tableaux CHAPTER 1 Table 1.1 Autotrophic and heterotrophic picoplankton community abundance along the

ARDEX transect. ..........................................................................................................73 Table 1.2 Observed protist taxa at each station at the surface along ARDEX transect. ......74 Table 1.3 Percentage of autotrophs and heterotrophs in the protist community in terms of

abundance and carbon biomass. Numbers in parenthesis are the percentages obtained after the exclusion of the unknown group. ...................................................................76

Table 1.4 Percentages of autotrophs and heterotrophs in the microbial community (picoplankton and protists) in terms of carbon biomass. Numbers in parenthesis are the percentages obtained after the exclusion of the unknown group..................................76

Table 1.5 Comparisons between bacterial metabolism (net bacterial production plus bacterial respiration) and net primary production (L. Retamal, unpublished). BGE are from Meon and Amon (2004). ......................................................................................77

CHAPTER 2 Table 2.1 Initial environmental and biological characteristics of ARDEX stations. SRP data

are from Emmerton (2006). ........................................................................................110 Table 2.2 Initial environmental and biological characteristics of GRW stations. .............110

Liste des annexes CHAPTER 1 Annex 1.1 Environmental characteristics of the stations. SPM: Suspended particulate

matter; POC: Particulate organic carbon; PON: Particulate organic carbon; DOC: Dissolved organic carbon; -: unavailable data. (± range; n = 2).................................126

Annex 1.2 Chlorophyll a biomass measured during the ARDEX cruise. Chla: total chlorophyll a; Chla<3µm: <3 µm chlorophyll a fraction. Each value is the mean of duplicates ± SD...........................................................................................................127

Annex 1.3 Heterotrophic picoplankton production as 3H-leucine uptake rates measured during the ARDEX cruise. Each value is the mean of duplicates ± SD. ....................127

Annex 1.4 Surface variables mean and S.D. for each zone. River: n = 4. TZ: n = 3. Sea: n = 3. Exceptions are mentioned in the table. .............................................................128

Annex 1.5 Bottom variables mean and S.D. for each zone. River: n = 3. TZ: n = 2. Sea: n = 3. Exceptions are mentioned in the table. .............................................................129

Annex 1.6 Spearman correlation coefficients. Bold: P ≤ 0.01. x: no significant correlation.....................................................................................................................................130

Annex 1.7 Protist and cyanobacteria photomicrographs. Station numbers are indicated. .131 Annex 1.8 Volumetric and integrated net heterotrophic picoplankton carbon production.

....................................................................................................................................132 Annex 1.9 Percentage saturation of CO2 in water compared to atmospheric value.

% saturation = [CO2]water / [CO2]air * 100%. Concentrations were in mmol m-3 corrected for solubility and temperature. (Ramlal, unpublished data). ......................133

CHAPTER 2 Annex 2.1 Surface water properties of the Great Whale River stations.............................134 Annex 2.2 Bacterial abundance in surface waters of the Great Whale River stations. ......135 Annex 2.3 Saturation curve and temporal series for the calibration of the 3H-leucine uptake

measurements. The measurements were made on water from the Great Whale River sampled on July 19th from the shore upstream of the Cris dock. The error bars represent the analytical standard deviation of the method (n = 3). The dotted lines show the 3H-leucine concentration and the incubation length used during the present study............................................................................................................................135

Annex 2.4 Heterotrophic picoplankton uptake of 3H-leucine and percentage of the activity due to particle-attached bacteria in the Great Whale River (2005). The error bars represent the analytical error (S.D. of triplicate measurements of the same water sample)........................................................................................................................136

Annex 2.5 Response to glucose addition at three stations in the ARDEX cruise 2004. The graphs show measurements done in each bottle. C: Control bottle. A and B: treatment bottles that have received 5 µM of glucose. Dashed line: Level of control 3H-Leu uptake for total community. Dotted line: Level of control 3H-Leu uptake for < 3 µm fraction. The error bars represent the analytical standard deviation of the method (n = 3). ................................................................................................................................137

Annex 2.6 Nutrient addition experiments in the Great Whale River system. The error bars represent the SD of the treatment (n = 3). The letters show the results of the Fisher-

xiii

LSD multiple comparison tests. Different letters mean significant differences (P < 0.05). ...................................................................................................................138

Introduction générale Introduction Les modèles de circulation globale prédisent que le réchauffement climatique se fera

ressentir plus fortement aux hautes latitudes nordiques (Moritz et al. 2002). Des

manifestations de ce phénomène sont déjà observées dans l'Arctique, tel la diminution de la

couverture de glace marine (Stroeve et al. 2005). Un autre exemple spectaculaire a

récemment exposé aux yeux du public l’impact des hausses de température dans l’Arctique

canadien. Entre 2000 et 2002, la plate-forme de glace Ward Hunt de l'île d'Ellesmere,

Nunavut, s’est fracturée causant le drainage d'un lac épiplate-forme, un écosystème rare

d'eau douce (Mueller et al. 2003). Cette fracturation est un signe très localisé du

réchauffement global de la planète et il témoigne de la sensibilité des écosystèmes arctiques

à celui-ci. Il est donc essentiel de développer une connaissance fondamentale des

écosystèmes nordiques afin de mieux comprendre leurs réactions potentielles aux

changements environnementaux.

Globalement, plus de la moitié du carbone organique emmagasiné dans le sol se retrouve

dans le bassin de drainage de l’océan Arctique (Dixon et al. 1994). Le pergélisol est un des

réservoirs significatif du carbone planétairement et il répond aux changements climatiques

d’une façon simple et unique : avec un réchauffement son étendue diminue, causant une

perte rapide de carbone; avec un refroidissement, le réservoir du permafrost se reconstitue

lentement de nouveau (Zimov et al. 2006). Ainsi, la fonte du pergélisol pourrait mener au

relâchement de ce carbone organique qui ruissellerait vers les lacs, les rivières et l’océan. Il

serait alors soumis à la dégradation microbienne et une partie de ce carbone serait

éventuellement relâché sous forme de CO2 vers l’atmosphère (Kling et al. 1991). Or, la

réactivité du carbone organique et l’écologie microbienne dans les écosystèmes aquatiques

arctiques restent mal connues. Une question de grand intérêt dans le contexte de

changement climatique global est de savoir si les eaux nordiques sont une source ou un puit

de gaz à effet de serre, essentiellement à cause de la grande étendue des milieux humides de

haute latitude (Vincent et Hobbie 2000).

15

L’étude présentée ici cherche à mieux comprendre l’écologie microbienne des écosystèmes

fluviaux, estuariens et côtiers arctiques ainsi que l’importance des microorganismes dans le

flux de carbone de ces écosystèmes de hautes latitude. Dans cette section, j’introduis les

concepts fondamentaux touchés par l’étude. En premier lieu, je décris la structure du réseau

alimentaire microbien en m’attardant à chacun de ses constituants. Ensuite, je présente

l’écologie microbienne des écosystèmes aquatiques en mettant l’emphase sur le recyclage

de la matière organique et en décrivant certaines méthodes utilisées pour mesurer le

métabolisme hétérotrophe microbien. Suivent une description des fleuves et des estuaires

arctiques et un portait des sites d’études visités lors de ma recherche, soient les systèmes du

fleuve Mackenzie et de la Grande Rivière de la Baleine. Finalement, j’explique, dans les

deux dernières sections, les hypothèses et les objectifs de mon étude ainsi que

l’organisation de mon mémoire.

Réseau alimentaire microbien Un des principaux avancements des deux dernières décennies en écologie aquatique est la

découverte que de minuscules organismes unicellulaires jouent un rôle majeur à la base des

réseaux alimentaires aquatiques et marins. Généralement, >90% de la matière organique

totale produite dans un écosystème ou entrant dans ce dernier est métabolisée, mais n’est

jamais consommé par les métazoaires mangeurs de particules (Wetzel 2001). Ces

microorganismes unicellulaires comprennent entre autre le picoplancton hétérotrophe

(Bactéries et Archaea), les picocyanobactéries, les picoeucaryotes, les nanoflagellés, les

ciliés et les autres protistes. Ils font partie de ce qui est appelé le réseau alimentaire

microbien (Fig. I.1) (Sigee 2005; Wetzel 2001). Le concept de boucle microbienne est

simplement un modèle illustrant la trajectoire du carbone et des nutriments à travers les

composantes microbiennes des communautés aquatiques pélagiques (Wetzel 2001). Le

réseau alimentaire microbien joue un rôle primordial dans la transformation de la matière

organique dissoute (DOM) en matière organique particulaire (POM) et ultimement dans sa

dégradation en nutriments et en CO2 par la respiration.

Ce réseau alimentaire comprend tous les niveaux trophiques : a) des bactéries hétérotrophes

qui utilisent la matière organique comme source d’énergie et qui l’intègre dans leur

biomasse ou qui la respire; b) des picocyanobactéries autotrophes qui sont spécialisées dans

16

la photosynthèse; c) des nanoflagellés et des ciliés qui phagocytent les particules

organiques (vivantes et non vivantes); et d) des flagellés mixotrophes qui tirent leur énergie

en partie de la photosynthèse et en partie de l’utilisation du carbone organique dissous et

des particules organiques (vivantes et non vivantes) (Vincent et Hobbie 2000). Le réseau

alimentaire microbien constitue une grande fraction de la biomasse totale des lacs et des

rivières arctiques. En transformant la DOM en POM, les organismes du réseau alimentaire

microbien permettent de transmettre cette source d’énergie au réseau alimentaire classique

duquel font partie les organismes supérieurs (Wetzel 2001).

Écologie bactérienne aquatique et dégradation de la matière organique Les organismes vivants ne constituent qu’une très faible proportion de la matière organique

présente dans l’environnement aquatique. La majorité du carbone organique dans l’eau est

constitué d’un mélange de produits végétaux, microbiens et animaux dans différents stades

de décomposition (Wetzel 2001) qu’on nomme le bassin des détritus (Canfield et al. 2005).

Dans les eaux naturelles, le carbone organique se retrouve sous un spectre continu de taille:

molécules libres, macromolécules, agrégats et organismes (Hedges 2002) (Fig. I.2). La

distinction entre le carbone organique particulaire (POC) et dissous (DOC) est arbitraire et

varie selon les différents auteurs. Au milieu du vingtième siècle, les filtres disponibles

avaient des pores d’une taille minimale de 0,45 - 1,0 µm. C’est de là que vient la définition

opérationnelle qui veut que le matériel dissous passe à travers de tels filtres et que la

matière particulaire n’y passe pas (Hedges 2002). Cependant, de tels filtres laissent passer

une bonne partie des petites bactéries qui sont alors considérées comme faisant partie du

matériel dissous. C’est pourquoi, dans le cas présent, la séparation entre le particulaire et le

dissous a été fixée à 0,2 µm (une taille largement utilisée), ce qui permet d’exclure la

majorité des bactéries de la fraction dissoute. Les phases dissoute et particulaire du carbone

organique ne sont pas indépendantes l’une de l’autre, mais sont reliées par une série

dynamique de flux réversibles assurés en grande partie par le métabolisme des organismes

aquatiques (Wetzel 2001).

Le carbone organique peut avoir une origine soit autochtone (découlant de la photosynthèse

in situ) soit allochtone (provenant de la photosynthèse dans le bassin versant). Dans

plusieurs systèmes lentiques (lacs et plans d’eau stagnante), l’écologie est dominée par la

17

fixation autochtone du carbone par le phytoplancton et l’entrée de carbone dans la chaîne

alimentaire se fait donc à l’interne (Sigee 2005). Au contraire, dans les systèmes lotiques, la

majorité du carbone organique provient des végétaux supérieurs terrestres, des terres

humides et du littoral (Wetzel 2001). La quantité et la qualité du carbone organique dans

l’eau d’un lac ou d’une rivière dépendent des caractéristiques du bassin versant (type de

végétation, perturbations), de la saison, du climat et de l’hydrologie (Williamson et al.

1999).

Les composantes du cycle de la matière organique dans la biosphère sont sa production, sa

transformation et sa dégradation. Dans les écosystèmes aquatiques, le picoplancton

hétérotrophe participe à ces trois étapes et il est principalement responsable de la

dégradation de cette matière organique (del Giorgio et Davis 2003). Différentes raisons

expliquent pourquoi ce sont les microorganismes qui dominent le recyclage de la matière

organique morte (Canfield et al. 2005):

1. Ils peuvent hydrolyser un assemblage divers de composés organiques dont plusieurs

sont inutilisables par les animaux supérieurs.

2. À cause de leur petite taille, et donc de leur ratio surface-volume élevé, ils peuvent

utiliser le carbone et les nutriments dissous en faible concentration.

3. Les petites cellules peuvent créer des contacts étroits avec les surfaces solides ce qui

leur permet de capter plus facilement les produits générés par l’activité de leurs

exoenzymes sur ces mêmes surfaces.

4. Certains types de microorganismes peuvent maintenir un métabolisme efficace et

minéraliser les substances organiques même dans des conditions anoxiques.

Dans les milieux aquatiques, la matière organique est présente sous forme dissoute et

particulaire, mais, généralement, la fraction dissoute est dominante (Wetzel 2001). Malgré

cela, le picoplancton hétérotrophe montre souvent une affinité pour la matière organique

particulaire. Cela serait dû au microhabitat bénéfique qu’offrent les particules et à la

présence de nutriments organiques adsorbés à leur surface (Droppo et al. 1998). Ainsi, les

particules créent une hétérogénéité spatiale dans la distribution de la matière organique, des

minéraux re-minéralisés et des espèces de microorganismes dans l’eau (Azam et al. 1993).

18

Lors de la dégradation de la matière particulaire, le picoplancton hétérotrophe obtient les

composés organiques dissous dont il a besoin par l’entremise de l’activité de ses

exoenzymes sur la POM (Wetzel 2001). Même si les bactéries attachées ne dominent pas

toujours dans la communauté bactérienne, la fraction des bactéries associées aux particules

montre souvent une plus grande activité métabolique que les bactéries libres (Crump et

Baross 2000; Crump et al. 1998; Vincent et al. 1996). Or, en général, la proportion de la

population totale de bactéries qui sont attachées aux particules tend à augmenter avec la

charge de particules en suspension (Fletcher 1991). Un désavantage possible de

l’attachement des bactéries aux particules est une plus grande susceptibilité au broutage par

le zooplancton (Fletcher 1991).

Le bassin de matière organique morte contient un vaste spectre de molécules qui ne sont

pas toutes facilement décomposables. Wetzel (2001) divise le carbone organique en deux

catégories: le carbone non humique et le carbone humique. La fraction non humique

comprend des glucides, des protéines, des peptides, des acides aminés, des lipides, des

résines et des pigments qui sont généralement labiles et qui sont donc rapidement utilisés

par les bactéries. La partie humique, pouvant former de 70 à 80 % du carbone organique,

est composée de molécules habituellement colorées de jaune à noir (et, dans ce cas, appelée

matière organique dissoute colorée ou CDOM), a un plus haut poids moléculaire et est

généralement plus récalcitrante à la dégradation biologique. Le carbone humique provient

en grande partie de l’activité microbienne sur le matériel végétal terrestre. Parce que la

plupart de la DOM est constituée de composés polymériques de haut poids moléculaire,

seule une petite partie de la DOM est rapidement utilisable dans les eaux naturelles (Wetzel

2001). À cause de cette hétérogénéité, la dégradation microbienne de la matière organique

s’étend sur une échelle temporelle allant des heures aux centaines d’années, tout dépendant

de la labilité des molécules impliquées (Canfield et al. 2005) (Fig. I.3). Les substances non

humiques sont les plus rapidement dégradées et le temps de décomposition va des heures

aux semaines. Les molécules humiques nécessitent un temps de dégradation de plusieurs

années. Certaines caractéristiques structurales de la matière organique sont même

reconnues pour diminuer sa biodégradabilité. Ainsi, la polymérisation extensive et les

embranchements multiples peuvent créer des liens qui ne sont pas hydrolysés facilement et

les composés hétérocycliques, polycycliques et aromatiques sont intrinsèquement difficiles

19

à dégrader (Canfield et al. 2005). Un composé typique contenant ces structures est la

lignine provenant de la dégradation du matériel structural des plantes vasculaires.

Or, la dégradation de la matière organique n’est pas simplement due aux microorganismes.

Il existe aussi des processus physiques contribuant à modifier la structure de la matière

organique et aussi de la dégrader en CO2. Un de ces phénomènes est la photodégradation de

la CDOM par les radiations solaires, en particulier par les UV et la lumière bleue. Les

manifestations de ces réactions sont le photoblanchissement de la CDOM (perte

d’absorption dans les UV et le bleu) et la production de photoproduits (Bertilsson et

Tranvik 2000; Lean 1998; Moran et Covert 2003). L’absorption de radiations solaires par la

CDOM génère une variété d’intermédiaires photochimiques qui peuvent potentiellement

altérer les processus métaboliques prenant place dans le milieu aquatique (Lean 1998;

Moran et Covert 2003). Ces intermédiaires comprennent des états excités de la DOM, des

électrons solvatés, des radicaux organiques, des ions superoxydes, des radicaux hydroxyles

et des radicaux peroxydes. Par contre, l’absorption de radiations solaires peut mener à la

photodégradation partielle du CDOM avec la génération d’acides gras volatiles et autres

composés simples qui servent d’excellents substrats pour la dégradation bactérienne

(Moran et Covert 2003; Wetzel 2001). Ces photoproduits biologiquement labiles font partie

d’une des quatre catégories suivantes identifiées jusqu’à maintenant: 1) des composés

organiques de faible poids moléculaire, 2) des gaz carbonés, 3) de la matière organique

blanchie non identifiée et des 4) composés riches en azote et en phosphore (Moran et Zepp

1997). L’étude de Tranvik et Bertilsson (2001) expose bien la dualité des effets de la

photodégradation sur le métabolisme microbien. Ils ont montré que l’effet net des

radiations solaires sur la biodisponibilité du DOC peut être modelé sur la base des

caractéristiques de celui-ci. Le DOC d’origine algale produit récemment semble être

majoritairement transformé en composés de plus faible qualité pour les microbes (Benner et

Biddanda 1998; Tranvik et Bertilsson 2001). Au contraire, le matériel humique plus âgé est

dégradé en produits stimulants la croissance bactérienne (Tranvik et Bertilsson 2001).

Leurs résultats suggèrent une interrelation élaborée entre l’activité microbienne et les

réactions photochimiques avec des implications pour la balance entre la dégradation et la

préservation de la matière organique morte. Brisco et Ziegler (2004) et Obernosterer et al.

(1999) sont arrivés à des conclusions similaires en montrant que le DOC plus réfractaire est

20

rendu plus labile et que le DOC labile est rendu plus réfractaire à la suite d’une exposition

aux radiations solaires.

Une étude récente par Bélanger et al. (2006) nous donne une idée de l’importance de la

photodégradation du DOC par rapport à d’autres processus de production ou d’utilisation

du CO2. Son travail mené dans le sud-est de la mer de Beaufort, Arctique canadien, a

montré que la photoproduction de carbone inorganique dissous (DIC) est équivalente à

10% du taux de respiration bactérienne, à 8% du taux de production primaire et à 2.8% du

DOC déchargé annuellement par le fleuve Mackenzie. Dans un contexte d’un océan sans

glace qui laisserait passer toutes les radiations solaires, la production de DIC par la

photodégradation atteindrait 6.2% du DOC déchargé par le Mackenzie.

En bref, la biodisponibilité de la matière organique dépend de facteurs intrinsèques à celle-

ci incluant ses caractéristiques chimiques (distribution du poids moléculaires, contenu en

nutriments et types de composés) qui sont déterminées par sa source et son état de

diagenèse (del Giorgio et Davis 2003). Or, l’utilisation de la matière organique et sa labilité

apparente sont également affectées par des facteurs extrinsèques qui régulent le

métabolisme des bactéries. Ainsi, la température, la disponibilité des nutriments

inorganiques, les interactions trophiques dans le réseau alimentaire microbien et la

composition phylogénétiques des assemblages bactériens ont une influence sur le

métabolisme bactérien et donc sur l’utilisation de la matière organique (del Giorgio et

Davis 2003). Par exemple, les bactérivores prédominant dans le milieu aquatique sont les

protistes, souvent les flagellés et les protozoaires ciliés (Wetzel 2001). Le picoplancton

hétérotrophe et les autres microorganismes sont aussi la cible des infections virales qui

causent de 40 à >50% des mortalités (Wetzel 2001). Les infections virales compromettent

l’intégrité des cellules et provoquent ainsi le relâchement de DOM dans le milieu.

Étant donné l’importance du picoplancton hétérotrophe dans le recyclage de la matière

organique, un grand intérêt de la recherche microbiologique est le développement de

techniques permettant d’étudier son métabolisme. Les méthodes les plus employées

aujourd’hui pour mesurer la production bactérienne ont recours à la prise de radio-isotopes

par les bactéries. La 3H-thymidine, un acide nucléique, est le premier radio-isotope à avoir

été utilisé pour mesurer la production bactérienne (Fuhrman et Azam 1980). Cette

21

technique est basée sur la prémisse que seules les bactéries en croissance active incorporent

cette molécule dans leur ADN. Une autre technique est basée sur l’incorporation d’un acide

aminé, la 3H-leucine, dans les protéines bactériennes (Kirchman et al. 1985). La production

de biomasse bactérienne peut être calculée à partir du taux de synthèse des protéines parce

que les protéines comptent pour une fraction importante et assez constante de la biomasse

bactérienne (environ 60%) (Kirchman 1993). Les deux méthodes donnent des taux

métaboliques bactériens similaires. Dans la présente étude, nous avons employé la méthode

de la 3H-leucine qui est environ dix fois plus sensible que celle de la thymidine et qui

permet d’évaluer le taux de production bactérienne sans avoir à mesurer la taille des

cellules (Wetzel et Likens 2000). L’incorporation de leucine dans les protéines est mesurée

en suivant l’apparition de radioactivité dans le matériel bactérien (Kirchman 1993) grâce à

différentes méthodes d’isolement. Après des tests préliminaires effectués en 2002 (Garneau

et al. 2006), nous avons adopté la méthode de microcentrifugation décrite par Smith et

Azam (1992). Cette technique n’a pas recours à la capture des organismes sur des filtres

comme en 2002, ce qui permet d’éviter l’adsorption des radio-isotopes qui se faisait sur les

filtres et donc de minimiser l’interférence avec l’absorption biologique. Pour exprimer la

productivité en terme d’unités de carbone, des facteurs de conversion appropriés doivent

être utilisés (Simon et Azam 1989). Par contre, les facteurs de conversion varient

grandement dans la littérature et sont donc encore un sujet de discussion.

Fleuves et estuaires arctiques Le bassin de drainage de l’océan Arctique (Arctic Ocean River Basins - AORB) couvre

15,5 x 106 km2 (Rachold et al. 2004) ce qui en fait, sur une base volumétrique, le bassin

océanique recevant le plus grand apport d’eau douce et de matière d’origine terrestre au

monde (Dittmar et Kattner 2003; Peterson et al. 2002). En effet, bien que l’océan Arctique

ne contienne que 1.0% de toute l’eau des océans du monde, il reçoit 11% du drainage

global d’eau douce (Rachold et al. 2004). L’apport total d’eau douce par les rivières y est

de 3299 km3 a-1 (Rachold et al. 2004) et le plus grand flux de carbone organique terrigène

vers l’océan se fait via celles-ci sous forme dissoute (Hansell et al. 2004). Les plus grandes

rivières arctiques en termes de décharge d’eau douce sont, en ordre: Yenisei, Lena, Ob,

Mackenzie, Pechora et Kolyma (Dittmar et Kattner 2003) (Fig. I.4).

22

Le bilan de l’océan arctique en terme d’apports terrestres est cependant en train de changer.

En effet, comme le souligne Peterson et al. (2002), la décharge d’eau douce des fleuves

eurasiens a augmenté de 7% entre 1936 et 1999 (Fig. I.5). Ce changement est corrélé avec

les variations de l’Oscillation Nord Atlantique et avec la hausse de la température globale

de l’air de surface. De plus, il y a eu une augmentation de 65% de la concentration de DOC

en 12 ans dans les milieux aquatiques drainant les bassins versants des Royaumes Unis

(Freeman et al. 2001). Cet enrichissement serait dû à une hausse de la température qui

stimulerait les exportations de DOC des tourbières vers les écosystèmes aquatiques (Evans

et al. 2002; Freeman et al. 2001) et à des changements dans l’hydrologie des milieux

(Tranvik et Jansson 2002). Comme on sait que plus de la moitié du carbone organique

mondial emmagasiné dans le sol se retrouve dans le basin de drainage de l’océan Arctique

(Dixon et al. 1994) et que la plupart des flux de carbone aquatique se font par des

mouvements de DOC provenant du sol, à travers les lacs et les rivières (Vincent et Hobbie

2000), il importe de s’interroger sur l’importance des rivières nordiques dans le transport

du carbone vers l’océan.

La concentration en matière organique des rivières arctiques est parmi les plus élevées du

monde (Dittmar et Kattner 2003). Le contenu en composés dissous et particulaires est

hautement dépendant du type de bassin de drainage et du régime hydrologique des rivières

(Cauwet 2002). Une des caractéristiques des rivières arctiques est qu’elles possèdent une

grande variabilité dans leurs taux de décharge saisonniers. En effet, elles évacuent plus de

90% de leur décharge annuelle entre mai et juillet et leur concentration en carbone

organique augmente parallèlement à la décharge en eau (Dittmar et Kattner 2003). En

général, les concentrations de DOC dépassent de beaucoup celles de POC, sauf dans le

fleuve Mackenzie qui est exceptionnellement riche en matière en suspension et qui

possèdent des niveaux comparables de POC et de DOC (Dittmar et Kattner 2003). Il existe

une relation inverse entre le contenu en POM de la matière en suspension totale (TSM) et la

concentration de cette dernière : le pourcentage de POM dans la TSM décroît avec une

augmentation logarithmique de la TSM, ce qui pourrait être dû à une réduction de la

production primaire en présence d’une haute concentration de matière en suspension

(Ittekkot et Laane 1991). Le ratio DOC/POC diminue également avec une augmentation en

TSM (Ittekkot et Laane 1991).

23

Un estuaire est défini comme un lieu où l’eau de l’océan est considérablement diluée par

l’eau douce provenant d’un cours d’eau. Comme l’eau douce est moins dense que l’eau

salée, on observe généralement dans les estuaires un système à deux couches. La couche

supérieure d’eau douce s’écoule vers le large en amenant avec elle un peu d’eau salée de la

couche inférieure, ce qui crée un vide. Ainsi, dans la couche inférieure, l’eau salée entre

dans l’estuaire pour remplir le vide (Mann 2000). Dans la couche d’eau douce, aussi

appelée plume, une suite de transformations biogéochimiques affectant le matériel

organique et inorganique, dissous et particulaire, se produit sur des échelles de temps et

d’espace variables (Dagg et al. 2004). Les charges de surface des particules transportées par

les rivières sont modifiées lorsqu’elles rencontrent l’eau salée. Ce phénomène amène la

floculation des particules dans une zone de l’estuaire qui est appelée zone de turbidité

maximale (Mann 2000). L’augmentation de taille des particules, suite à la floculation, les

amène à sédimenter hors de la couche d’eau douce, diminuant du même coup la turbidité de

cette dernière. La majeure partie (60-90%) de la matière particulaire déchargée par les

rivières est précipitée dans les estuaires et la zone de plateau continental adjacente, ce qui

explique pourquoi la zone de transition entre l’eau douce et l’eau salée est appelée le filtre

marginal (marginal filter) (Lisitsyn 1995; Vetrov et Romankevich 2004). De plus, le

mélange avec l’océan dilue la CDOM fluviale, ce qui augmente encore plus la pénétration

de la lumière (Dagg et al. 2004).

Ces changements dans les propriétés physiques et optiques ont un impact majeur sur les

processus chimiques et biologiques prenant place dans les plumes fluviales (Dagg et al.

2004). Par exemple, des études faites dans le fleuve Saint-Laurent ont montré que la

communauté bactérienne de la zone de transition estuarienne différait de celle de la zone

d’eau douce par un plus grand pourcentage de bactéries attachées au matériel particulaire

en suspension (Painchaud et al. 1995; Vincent et al. 1996). De plus, un gradient de

production primaire est généralement observé dans les estuaires dont le côté fluvial est très

turbide et concentré en nutriments et dont le côté marin est clair et moins riche en

nutriments. Du côté fluvial, la production primaire est limitée par la disponibilité de la

lumière et du côté marin, par les nutriments. Conséquemment, un niveau de production

primaire maximal est généralement observé entre les deux zones (Eisma et Cadée 1991).

Cette croissance phytoplanctonique accrue stimule ensuite les réseaux alimentaires

24

classique et microbien (Dagg et al. 2004). De plus, une augmentation dans les

transformations photochimiques et microbiennes de la DOM fluviale réfractaire amène un

relâchement de nutriments qui stimulent à leur tour la production phytoplanctonique, les

réseaux alimentaires et les cycles des éléments (Dagg et al. 2004).

Les apports fluviaux contribuent substantiellement au budget du DOC de l’océan Arctique

et dominent les budgets des plateaux continentaux (Dittmar et Kattner 2003). L’apport des

rivières en matière organique labile représente 13% de la production primaire dans les

estuaires, et seulement 1% de celle des zones côtières (Ittekkot et Laane 1991). Cela

signifie que la production primaire estuarienne est la source d’énergie principale pour les

organismes hétérotrophes dans les estuaires et les zones côtières.

Comme les estuaires sont la transition entre les environnements d’eau douce et d’eau salée,

un changement de communauté biologique se produit dans ceux-ci. De nombreuses études

ont été menées dans le fleuve Saint-Laurent dans le but de mieux comprendre les gradients

biologiques se produisant dans sa zone de transition. Frenette et al. (1995) ont vu une

augmentation du nombre de taxa et de la taille des cellules de phytoplancton et de

protozoaires en se déplaçant vers l’eau salée. Winkler et al. (2003) ont observé dans

l’estuaire du Saint-Laurent des changements dans la structure de la communauté de

protistes : les espèces hétérotrophes ne composaient que 20 % du total dans l’eau douce,

mais elles dominaient la biomasse dans les salinités plus élevées (jusqu’à 6 psu). Vincent et

al. (1996) ont aussi constaté un changement brusque de communauté dans la zone de

transition eau douce / eau salée du Saint-Laurent : une biomasse importante de

phytoplancton en faible salinité était remplacée par une grande biomasse de zooplancton et

d’ichtyoplancton en plus forte salinité. Les estuaires sont donc des régions de forte

productivité primaire et secondaire qui agissent comme des pouponnières riches en

nourriture pour plusieurs populations de poissons (Vincent et Dodson 1999).

Le fleuve Mackenzie et son estuaire À travers le Canada, c’est le bassin versant du fleuve Mackenzie qui a connu le plus grand

réchauffement au cours des 100 dernières années (Macdonald et Yu 2006). La perte du

pergélisol, spécialement dans la zone discontinue, pourrait altérer le couplage hydrologique

25

entre le fleuve Mackenzie et son bassin versant avec des conséquences sur le transport de

l’eau, des sédiments, du carbone organique et des nutriments (Macdonald et Yu 2006).

Le fleuve Mackenzie est le quatrième fleuve arctique en importance en terme d’apport

d’eau douce dans l’océan Arctique (330 km3 an-1) (Macdonald et al. 1998). Pour ce qui est

de sa décharge sédimentaire, il est classé au premier rang des fleuves arctiques en

transportant 124 Mt an-1 de sédiments vers l’océan (Carson et al. 1998; Rachold et al.

2004). Le débit du fleuve Mackenzie montre une très forte variation saisonnière (Fig. I.6)

passant de 3000-5000 m3 s-1 en hiver à 40 000 m3 s-1 lors du dégel au début juin, qui est

aussi sa période de décharge sédimentaire maximale (Macdonald et Yu 2006). Pour

comparaison, le fleuve Saint-Laurent décharge annuellement 400 km3 d’eau douce dans

l’océan Atlantique et a un débit moyen à la hauteur de la ville de Québec de 12 600 m3 s-1

(Vincent et Dodson 1999), avec des valeurs maximales de 15 000 à 25 000 m3 s-1 au

printemps (Vincent et al. 1996). La température de l’eau du Mackenzie est aussi fortement

variable variant entre 0oC et 20oC.

Le fleuve Mackenzie prend sa source dans le Grand Lac des Esclaves et son bassin versant

est le plus grand au Canada, s’étendant sur 1,8 x 106 km2 (Macdonald et al. 1998) (Fig. I.7).

La partie ouest du bassin possède un relief très élevé et procure la plus grande partie des

sédiments au fleuve Mackenzie. La portion centrale du bassin draine les Plaines intérieures

et la partie est se situe dans le Bouclier canadien. Le relief et l’apport de sédiments de ces

deux dernières portions sont relativement faibles (Macdonald et al. 1998). Le tiers le plus

au nord du bassin possède très peu d’arbres tandis que la partie sud est couverte par la forêt

boréale (Droppo et al. 1998). Ainsi, le fleuve Mackenzie n’est pas simplement arctique,

mais possède un caractère arctique et tempéré (Macdonald et Yu 2006). Le fleuve

Mackenzie est couvert de glace de la fin septembre à la fin juin dans sa partie la plus

nordique (Droppo et al. 1998). Contrairement aux autres fleuves arctiques, le Mackenzie

transporte annuellement une plus grande quantité de carbone organique sous forme

particulaire que dissoute. Sa concentration en DOC est de 5,2 mg L-1 et en POC, de

7,2 mg L-1 (moyenne multi annuelle) (Rachold et al. 2004).

Le delta du Mackenzie (Fig. I.8) couvre 12 170 km2 (Droppo et al. 1998). Le fleuve y

décharge 1.8 Mt a-1 de POC avec une quantité approximativement égale de DOC

26

(1.3 Mt a-1) (Macdonald et al. 1998). Le fleuve Mackenzie s’y écoule par trois principaux

chenaux. Les deux tiers du volume empruntent le Middle Channel, un sixième, le East

Channel et le dernier sixième, le West Channel. Lors de la débâcle à la fin du printemps /

début de l’été, plus de 95 % du delta est inondé à cause des embâcles (Droppo et al. 1998).

Le Mackenzie envahit alors les lacs qui parcellent son delta.

La structure de l’estuaire du Mackenzie est décrite par Carmack et Macdonald (2002) (Fig.

I.9). En juillet-août, le fleuve Mackenzie s’écoule sur un plateau continental libre de glace.

Sa décharge d’eau douce reste importante, ce qui est mis en évidence par la présence de

plumes et de fronts turbides et chauds. La plume envahit la côte et le plateau continental et

la forme de celle-ci est grandement affectée par les vents. Les marées sont de faibles

amplitudes dans l’estuaire du Mackenzie et ont un effet minime sur la dispersion de la

plume. La structure de la plume est observable par imagerie satellitaire pendant l’été et elle

peut s’étendre à plus de 400 km de la côte. L’eau du fleuve Mackenzie reste douce jusqu’à

ce qu’elle passe au-dessus du seuil de 2 m caractéristique des estuaires arctiques. À cet

endroit, l’apport d’eau douce forme un estuaire partiellement mélangé avec une couche

d’eau saumâtre entraînant l’eau salée du fond avec une forte pycnocline à 10-20 m. Du côté

océanique de la rampe de 2 m, une couche d’eau fluviale turbide de 2 à 4 m d’épaisseur

s’étend vers le large sur 5 à 10 km. La salinité augmente de 0 à 25 psu à l’intérieur de cette

zone très turbide et plusieurs processus associés avec le comportement non-conservatif de

la matière organique (floculation, sédimentation, etc.) se déroulent à cet endroit. À cause de

la faible profondeur de cette région, le vent des tempêtes peut produire un mélange vertical

important qui augmente la re-suspension des sédiments du fond.

Dans cette zone de transition de l’eau douce à l’eau salée, le comportement du carbone

organique particulaire (POC) semble non-conservatif et il est assumé que celui du carbone

organique dissous (DOC) soit conservatif (Dittmar et Kattner 2003; Macdonald et Yu

2006). En d’autres termes, il y aurait une perte de POC dans l’estuaire, mais pas de DOC.

Cependant, plusieurs aspects restent mal compris en ce qui a trait au comportement de la

matière organique et des sédiments dans les estuaires. Dans la mer de Beaufort, environ

50% de l’apport sédimentaire est piégé dans le delta et environ 40% sur le plateau et le

reste s’échappe du bord du plateau (Macdonald et al. 1998).

27

Pour ce qui est des communautés planctoniques de l’estuaire du Mackenzie, Parsons et al.

(1988; 1989) suggèrent qu’ils en existent deux distinctes : la première se trouve à

l’embouchure du fleuve, dans l’eau peu salée, et est constituée majoritairement de bactéries

et d’amphipodes et la seconde se situe plus au large et est caractérisée par du

phytoplancton, des copépodes, des hydroméduses et des cténophores.

La Grande Rivière de la Baleine La Grande Rivière de la Baleine (Fig. I.10) est une rivière subarctique de la côte sud-est de

la Baie d’Hudson. Elle prend sa source au lac Bienville et s’écoule sur le lit granitique du

Bouclier canadien précambrien. Son bassin versant (42 736 km2) est couvert par la forêt

boréale qui est dominée par les épinettes (Picea spp.), le mélèze (Larix laricina (Du Roi)

K. Koch) et le peuplier faux-tremble (Populus tremuloides Michx.) entrecoupée de zones

de lichens et de tourbières. Le peuplier baumier (Populus balsamifera L.), le myrique

baumier (Myrica gale L.) et l’aulne rugueux (Alnus rugosa (Du Roi) Spreng.) se retrouvent

dans la zone riparienne de la Grande Rivière de la Baleine (Hudon 1994).

La Grande Rivière de la Baleine a un débit moyen d’environ 700 m3 s-1 (décharge

d’environ 22 km3 a-1) allant de 200 m3 s-1 en mars à 1300 m3 s-1 en juin (Ingram 1981). Elle

est classée cinquième en importance comme source d’eau douce dans l’est de la Baie

d’Hudson après les rivières La Grande, Nottaway, Eastmain et Rupert (Hudon 1994). La

dispersion de sa plume d’eau douce est surtout contrôlée par les vents (Ingram 1981). À

marée basse, un front peut être observée à la périphérie de la plume sous la forme d’une

accumulation de mousse séparant des eaux de couleurs différentes. Ce front est caractérisé

par un changement de salinité de 12 ‰ (Ingram 1981) qui est accompagné par un

changement dans la quantité mais non dans la qualité du CDOM (Retamal et al. 2007).

Hypothèses et objectifs Comme nous l’avons souligné précédemment, une question de grand intérêt dans le

contexte actuel de changements climatiques globaux est « Est-ce que les eaux nordiques

sont une source ou un puit de gaz à effet de serre? » (Vincent et Hobbie 2000).

Actuellement, de grands efforts de recherche scientifique sont en branle à travers

l’Arctique, mettant l’emphase sur les changements climatiques et le cycle du carbone. La

28

majeure partie de mon projet s’imbriquait dans le programme multidisciplinaire Arctic

River Delta Experiment (ARDEX, un programme satellite du projet Canadian Arctic Shelf

Exchange Study (CASES)) qui s’intéressait à la matière organique dissoute dans le fleuve et

le delta du Mackenzie et son effet sur les communautés microbiennes, la pénétration des

radiations solaires dans l’eau, la photochimie et l’écologie du zooplancton. Mon étude se

concentrait sur l’écologie microbienne, en particulier sur la structure des communautés

microbiennes fluviales et côtières et leur métabolisme hétérotrophe, et était l’une des

premières du genre se faisant dans le système du fleuve Mackenzie. Les données ont été

obtenues dans le but de mieux comprendre le rôle des environnements fluviaux dans les

zones polaires, spécifiquement dans la production et le relâchement de CO2 vers

l’atmosphère. ARDEX permettra de mieux comprendre les impacts potentiels des

changements de climat sur les écosystèmes nordiques et les mécanismes de rétroactions de

ceux-ci avec la biosphère. L’autre partie de mon projet s’est déroulé dans la Grande Rivière

de la Baleine et avait pour but de faire certaines comparaisons avec le système du

Mackenzie, de l’autre côté de l’Arctique canadien, concernant les facteurs de contrôle du

métabolisme hétérotrophe microbien.

Objectifs généraux

Ce projet avait deux grands objectifs principaux. Ceux-ci se basaient sur la prémisse que le

picoplancton hétérotrophe (Bactéries et Archaea) est le principal utilisateur du carbone

organique dans le milieu aquatique. Nous nous sommes intéressés à son métabolisme pour

évaluer l’importance de la dégradation biologique du carbone organique en CO2. Nous nous

sommes également intéressés à la structure de la communauté microbienne puisque le

picoplancton hétérotrophe est à la base du réseau alimentaire microbien et que les

interactions de celui-ci avec les autres composantes du réseau affectent l’efficacité du

recyclage de la matière organique (bactérivorie, compétition, etc.). Finalement, nous

voulions étudier les changements dans la production du picoplancton hétérotrophe et dans

la structure de la communauté microbienne au travers de la transition d’un important fleuve

arctique vers un milieu marin côtier, sachant que le carbone organique y subit des

transformations majeures. Nos objectifs généraux se définissaient donc ainsi :

29

A) Définir les changements dans la production bactérienne et dans la structure de

la communauté microbienne le long du gradient eau douce / eau salée du fleuve

Mackenzie et de la mer de Beaufort.

B) Mieux comprendre les facteurs contrôlant le métabolisme des bactéries dans le

système du Mackenzie et le système de la Grande Rivière de la Baleine dans le but

de mieux cerner comment les changements climatiques pourraient affecter celui-ci.

Hypothèses de recherche

Pour chacun de ces objectifs généraux, nous avons développé les hypothèses de recherche

suivantes :

A. Gradients microbiens :

A1) Le métabolisme bactérien augmente d’Inuvik jusqu’à la mer à cause de

l’augmentation de la biodisponibilité du carbone organique dissous présent dans le

milieu.

A2) Il existe un changement brusque dans la structure de la communauté

microbienne au sein de la zone de transition eau douce / eau salée. D’une

communauté dominée par les bactéries hétérotrophes dans le fleuve et dans la zone

de mélange de l’eau douce et de l’eau salée, où la turbidité est élevée, il y a passage

à une communauté dominée par les picocyanobactéries et le phytoplancton dans

l’eau salée.

A3) Les particules jouent un rôle majeur dans l’activité totale de la communauté

bactérienne.

B. Facteurs de contrôle :

B1) Le métabolisme bactérien est limité par la disponibilité du carbone organique

dans les fleuves arctiques.

B2) La photochimie des UV sur le CDOM augmente la biodisponibilité de celui-ci

dans le système du Mackenzie.

B3) Les facteurs contrôlant la production bactérienne ne sont pas les mêmes entre

la Grande Rivière de la Baleine et le Fleuve Mackenzie à cause de leurs différences

physico-chimiques.

30

Approches et objectifs spécifiques

Pour adresser les hypothèses A1 à A3, nous avons participé à la croisière d’échantillonnage

du programme ARDEX sur le NGCC Nahidik, entre le 26 juillet et le 2 août 2004, dans le

fleuve Mackenzie et la mer de Beaufort. Nous avons couvert un transect de 300 km allant

d’Inuvik à une station située 50 km au large de la côte dans la mer de Beaufort et qui

traversait le gradient eau douce / eau salée entre le fleuve et la mer. À cette période de

l’année, le débit du fleuve était environ la moitié de son débit maximal (Fig. I.6). Les

variables suivantes ont été mesurées:

1. L’abondance des bactéries hétérotrophes, en utilisant la microscopie à épifluorescence.

2. L’abondance des picocyanobactéries et des picoeucaryotes autotrophes, grâce à la microscopie à épifluorescence.

3. L’abondance des protistes, avec une méthode développée pour les estuaires turbides.

4. Les paramètres physicochimiques : température, salinité, turbidité, matière particulaire en suspension (SPM), matière organique particulaire (POM), carbone et azote organique particulaire (POC et PON) et carbone organique dissous (DOC).

5. La concentration de chlorophylle a (Chl a) dans l’eau comme une mesure de la biomasse du phytoplancton.

6. La prise de 3H-leucine par le picoplancton hétérotrophe total et la fraction <3 µm comme indicateur de la production hétérotrophe bactérienne.

Les hypothèses B1 à B3 ont été évaluées lors de la croisière ARDEX 2004 et lors d’une

mission d’échantillonnage, en 2005, sur la Grande Rivière de la Baleine dans le Nord

québécois. Les mesures suivantes ont été effectuées:

1. Les changements dans la prise de 3H-leucine par les bactéries suite à l’ajout de glucose dans le fleuve Mackenzie, son estuaire et la mer de Beaufort.

2. Les changements dans la prise de 3H-leucine à la suite de l’exposition aux radiations solaires d’eau filtrée sur 0,2 µm dans le fleuve Mackenzie et la mer de Beaufort.

3. Les changements dans la prise de 3H-leucine par les bactéries à la suite de l’ajout de glucose et de phosphore dans la Grande Rivière de la Baleine et son estuaire (le phosphore a été ajouté pour augmenter la portée des expériences).

31

Organisation du mémoire Ce mémoire de maîtrise comporte deux chapitres. Le premier se veut une description de la

structure de la communauté microbienne et du métabolisme du picoplancton hétérotrophe

dans le fleuve Mackenzie et la mer de Beaufort. Le second chapitre présente les résultats

des expériences effectuées dans le but de mieux comprendre les facteurs contrôlant

l’activité bactérienne.

32

Figure I.1 Le réseau alimentaire microbien et ses composantes. Traduit et adapté de Vincent et Hobbie (2000).

Figure I.2 Illustration du continuum de taille de la matière organique retrouvée dans les milieux aquatique et marin et les différentes méthodes d’isolement. Tiré et traduit de Hedges (2002).

33

Figure I.3 Représentation idéalisée des taux de décomposition de biomolécules importantes. Tiré et traduit de Canfield (2005).

Figure I.4 Représentation de l’océan Arctique et de ses principaux affluents. La largeur des flèches est proportionnelle à la décharge en DOC qui est donnée en 1012 g C an-1. Tiré de Dittmar et Kattner (2003).

34

Figure I.5 Illustration des changements de l’anomalie de décharge des rivières arctiques, de l’index NAO (North Atlantic Oscillation) hivernal et de la température moyenne de l’air de surface globale (SAT) de 1936 à 1999. Tiré de Peterson et al. (2002).

Figure I.6 Cycle hydrologique du fleuve Mackenzie montrant son apport d’eau dans l’océan Arctique pour les années 1973 à 1990. Tiré de Macdonald et al. (1998). La zone grise montre la période de la croisière d’échantillonnage ARDEX 2004. La ligne pointillée montre le débit moyen du fleuve Saint-Laurent à Québec (Vincent et Dodson 1999).

35

Figure I.7 Situation géographique du bassin de drainage du fleuve Mackenzie. Tiré du site Internet Mackenzie River Basin Board (Online).

Figure I.8 Détail du delta du Mackenzie. Tiré de Emmerton (2006).

36

Figure I.9 Illustration de la distribution de salinité dans l’estuaire du Mackenzie en été. La zone ombrée montre la distribution typique des sédiments en suspension associés avec la plume du Mackenzie et la re-suspension du fond. Tiré de Carmack et Macdonald (2002).

Figure I.10 La Grande Rivière de la Baleine.

CHAPTER 1: GRADIENTS IN BACTERIAL PRODUCTION AND MICROBIAL COMMUNITY STRUCTURE IN THE MACKENZIE RIVER AND ITS ESTUARY

Résumé Lors de la mission Arctic River-Delta Experiment (ARDEX), nous avons mesuré des

variables environnementales et microbiennes suivant un transect de 300 km s’étendant dans

le fleuve Mackenzie et la mer de Beaufort pour étudier les gradients dans la structure et

dans la production de la communauté microbienne entre ces deux zones. La concentration

bactérienne moyenne en surface était de 6,7 x 105 cellules ml-1. L’abondance

picophytoplanctonique diminuait vers la mer, variant de 51 à 0,03 x 103 cellules ml-1. La

communauté de protistes montrait des changements d’espèce et de taxons. La production

bactérienne fractionnée mesurée par la prise de 3H-leucine variait de 25 à 134 pmol L-1 h-1.

Le pourcentage de la production due aux bactéries attachées diminuait avec une diminution

du carbone organique particulaire, contribuant de 97%, à Inuvik, à 16%, à un site marin. Le

fleuve Mackenzie était nettement hétérotrophe et la mer de Beaufort tendait vers

l’autotrophie.

Abstract As part of the Arctic River-Delta Experiment (ARDEX), we measured environmental and

microbiological variables along a 300 km transect across the Mackenzie River and coastal

Beaufort Sea in July-August 2004 to investigate the river and estuarine gradients in

microbial community structure and activity. Surface bacterial concentrations averaged 6.7 x

105 cells ml-1. Picophytoplankton abundance decreased towards marine sites varying

between 51 and 0.03 x 103 cells ml-1 and there were changes in protist community structure.

Size-fractionated bacterial production as measured by 3H-leucine uptake varied between 25

and 134 pmol 3H-leucine L-1 h-1. The percentage of production by attached (particles

>3 µm) versus total bacteria diminished with decreasing particulate organic carbon, from

97% at Inuvik to 16% at a marine site. The combined analysis of several variables

measured during ARDEX showed that the Mackenzie River was net heterotrophic and that

the Beaufort Sea tended towards autotrophy at the time of sampling.

38

Introduction Global warming has become a subject of increasing concern, especially for the Polar

Regions (Moritz et al. 2002). Unique ecosystems that depend on ice cover are especially

vulnerable and have begun disappearing (Mueller et al. 2003; Serreze et al. 2000). Dixon et

al. (1994) estimated that more than half of the global organic carbon pool is stocked in the

catchment area of the Arctic Ocean. Permafrost melting has the potential to liberate this

organic carbon in the watershed of lakes and rivers, and make it available for microbial

breakdown to CO2 (Kling et al. 1991). The arctic rivers discharge annually 3299 km3 y-1

into the Arctic Ocean or approximately 11% of the global river discharge (Rachold et al.

2004). There is therefore much interest in identifying the role of large arctic rivers in

greenhouse gas production, given that these are the principal links between land and the

Arctic Ocean, and likely have a strong influence on arctic coastal ecosystems.

Over the past 100 years within Canada, the greatest warming has been observed in the

Mackenzie Basin (Macdonald and Yu 2006). The Mackenzie River is the fourth largest

arctic river in term of freshwater discharge (330 km3 y-1) and has the largest catchment area

in Canada (1.8 x 106 km2) (Macdonald et al. 1998). However, there is lack of knowledge

about this system, especially concerning its microbial ecology. Spears and Lesack (2006)

give information about microbial dynamics in floodplain lakes of the Mackenzie River,

however little work has been conducted on the main body of the river. Some preliminary

sampling in 2002 (Garneau et al. 2006; Wells et al. 2006) provided insights into the

molecular ecology of Bacteria and Archaea in the Mackenzie River, but little is known

about the microbial food web of this ecosystem.

Freshwater-saltwater transition zones lie at the interface between rivers and the sea. They

integrate upstream and downstream processes, are one of the most biologically productive

sections of the river, and are a prime site for monitoring fluvial and estuarine health

(Vincent and Dodson 1999). Estuaries are also regions of major and complex

biogeochemical transformations of organic and inorganic, dissolved and particulate

material (Dagg et al. 2004). A river estuary is thus a major step of the hydrological and

biogeochemical pathway from land to ocean. One important process taking place in the

freshwater-saltwater transition zone is the flocculation of particles of different sizes (Eisma

39

and Cadée 1991). The formation and enlargement of such particles increase their

sedimentation rates out of the buoyant freshwater plume, causing the diminution of

turbidity in the marine side of the estuaries. The microbial community structure also

undergoes pronounced changes along the salinity gradients (bacterioplankton (Bouvier and

del Giorgio 2002; Crump et al. 2004; Galand et al. 2006; Garneau et al. 2006; Selje and

Simon 2003), phytoplankton and other protists (Frenette et al. 1995)), although little is

known about this aspect of high latitude estuaries.

The Mackenzie River is the largest arctic river in terms of sediment discharge (124 Mt a-1;

Rachold et al. 2004) and it deposits 65 Mt of sediments in its delta annually (Macdonald et

al. 1998). Contrary to other large arctic rivers, the Mackenzie carries annually more

particulate organic carbon (POC) than dissolved organic carbon (DOC) (Rachold et al.

2004). Particles are thus a major player of the biogeochemistry of the Mackenzie River and

estuary. Aggregates constitute important microhabitats for microorganisms and, generally,

the proportion of particle-bound bacteria increases with the suspended particle

concentration (Fletcher 1991). In the St Lawrence River transition zone, particle-attached

bacteria dominate total bacterial production (Vincent et al. 1996). In the study of Crump et

al. (1998), 90% of the bacterial activity was associated with particles larger than 3 µm in

the Columbia River Estuary.

Our aim in the present study was to define the gradients in bacterial activity and in

microbial community structure across the freshwater-saltwater transition zone between the

Mackenzie River and the Arctic Ocean. We hypothesized that there are major changes in

microbial community structure across this interface: from a heterotrophic community in the

river and the estuary where the turbidity is high, to a community dominated phytoplankton

in the marine zone. We surmised that bacterial metabolism increases towards the marine

stations due to the increasing lability of the dissolved organic carbon in the ocean, and that

particle-attached bacteria (relative to free living cells) would account for a large fraction of

total bacterial production. Another goal of this study was to integrate our measurements

with data from other ARDEX projects in order to address the question posed by Vincent

and Hobbie (2000): “Are northern waters a source or a sink of greenhouse gases?”

40

To investigate those questions, a sampling cruise was carried on between July 26th and

August 2nd 2004 along a transect crossing the Mackenzie River and the Beaufort Sea

ecosystems. We sampled 10 stations and measured physicochemical variables, chlorophyll

a biomass, the abundance of heterotrophic bacteria, picocyanobacteria, autotrophic

picoeukaryotes and protists, and heterotrophic picoplankton production as determined by 3H-leucine uptake and incorporation into bacterial protein.

Methods

Sampling Sampling was carried on aboard the CCGS Nahidik in the Mackenzie River and the

Beaufort Sea within the framework of ARDEX (Arctic River Delta Experiment), a

multidisciplinary satellite program of CASES (Canadian Arctic Shelf Exchange Study).

Water samples were collected along a 300 km transect between Inuvik, NWT, Canada, and

a station 50 km offshore in the Beaufort Sea across the freshwater-saltwater transition zone

(TZ) (Fig. 1.1). Surface samples were collected using a clean plastic bucket and depth

samples using a 6.2 L Kemmerer sampler or a peristaltic pump. We sampled 10 stations of

which 8 were sampled at the surface and near the bottom and 2 were sampled only at the

surface.

Physical characteristics of the water column A CTD (Conductivity, Temperature and Depth) profiling logger (RBR, Canada) was used

to profile the water column conditions. The logger was also equipped with a fluorometer to

measure the chlorophyll a concentrations.

Particulate and dissolved material Turbidity was measured using an Aquafluo handheld fluorometer (Turner Designs inc.,

USA) and expressed as nephelometric turbidity units (NTU) calibrated against standards

from GFS Chemicals Inc. (2.0, 10, 40 and 100 NTU). Samples for suspended particulate

matter (SPM) were filtered in duplicate onto pre-ashed and pre-weighed GF/F filters (0.7

µm, 47 mm) which were folded, wrapped in aluminum foil and stored at -80oC. Filters were

then dry for 24 h at 60oC and re-weighed for determination of SPM mass. Filters were

subsequently combusted at 500oC for 1.5 h and weighed to obtain the mass of particulate

41

inorganic matter (PIM). Particulate organic matter (POM) was obtained by subtracting PIM

from SPM.

Particulate organic carbon (POC) and particulate organic nitrogen (PON) samples were

filtered in duplicates onto pre-combusted GF/C filters (1.2 µm, 25 mm) which were folded

and kept frozen (-80oC) in aluminum foil until further processing. POC and PON

concentration were analyzed by high temperature oxidation using an elemental analyzer

LECO CHNS-932 (Institut national de la recherche scientifique, Centre Eau, Terre et

Environnement (INRS-ETE) Québec City, Canada) with a detection limit of 0.03 mg L-1 for

carbon and of 0.005 mg L-1 for nitrogen. Filters were acidified with HCl fumes overnight

and allowed to dry at 65°C prior to analysis in tin or silver sleeves.

Dissolved organic carbon (DOC) samples were obtained by filtering water through 0.2 µm

cellulose acetate filters (47 mm) (previously flushed with the sample). The filtrate was kept

at 4oC in acid washed amber glass bottles thoroughly rinsed with the sample, until further

analysis by high temperature combustion. The samples were bubbled with CO2-free

nitrogen for 7 min to ensure the removal of all the dissolved inorganic carbon (DIC).

Analyses were done using a Shimadzu TOC Analyzer 5000A (detection limits of

0.05 mg L-1) at the INRS-ETE, Québec City, Canada.

Chlorophyll a Water samples were fractionated through Poretics 3 µm polycarbonate filters (0.7 µm,

47 mm). Total (Chl a) and fractionated chlorophyll a (< 3 µm Chl a) water samples were

then filtered in duplicate onto Whatman GF/F glass fiber filters (47 mm). After folding,

filters were wrapped in aluminum foil and stored at -80oC. Filters were extracted with hot

ethanol (95 %) and Chl a concentrations were determined by fluorometry before and after

acidification using a Cary Eclipse spectrofluorometer calibrated against a standard curve of

Chl a made with a Cary 300 Bio U.V. spectrophotometer.

Microbial community structure Picophytoplankton (picocyanobacteria and picoeukaryotes) samples were filtered without

previous fixation onto Anodisk filters (0.2 µm, 25 mm) under gentle pressure. 10 ml (river

and estuarine stations) to 30 ml (marine stations) were filtered and filters were mounted

42

between slides and cover slips using Aquapoly/Mount (Polyscience, Inc.). The slides were

stored at -20oC for up to 20 months before analysis. The samples were counted under a

Zeiss Axioskop 2 epifluorescence microscope using green and blue lights at 1000x

magnification with immersion oil. Picocyanobacteria fluoresce bright orange or red under

green light and yellow or pale red under blue light contrary to photosynthetic

picoeukaryotes that fluoresce deep red in both cases (MacIsaac and Stockner 1993). A

minimum of 15 fields and 400 cells were counted where possible (not possible at the

bottom of R5d, R5a and R7 and at all depths of R8 and R9).

Heterotrophic picoplankton samples (Bacteria and Archaea) were preserved with

formaldehyde (2%, final conc.) in acid washed clear glass bottles previously rinsed with the

samples and stored in the dark at 4oC (for up to 10 months). Due to the presence of large

amounts of sediment, the samples from the river and estuarine stations were sonicated for

15 s (the duration was determined after sonication tests; see Fig. 1.2). For sonication,

samples were poured in acid washed glass test tubes, rinsed with milli-Q water and

samples, and placed in an ultrasonic bath (Bransonic 220, 117 volts, 50-60 Hz, 125 watts)

filled with water and ice (to prevent warming of samples). Samples where then filtered onto

Nuclepore black polycarbonate membranes (0.22 µm, 25 mm) placed on cellulose acetate

backing filters (0.8 µm, 25 mm) under low pressure. DAPI (Porter and Feig 1980) (5 µg L-1

final conc.) was added when only 2 ml of sample remained in the filter tower, and then left

to stain for 15 min before final filtration. Filters were mounted between microscope slides

and cover slips with non fluorescent immersion oil and stored at -20oC until counting. The

counting was made using a Zeiss Axioskop 2 epifluorescence microscope, under UV light

and 1000x magnification with immersion oil. A minimum of 15 fields and 400 cells were

counted when possible.

Protist samples were preserved with paraformaldehyde (0.5 g L-1 final conc.) and

glutaraldehyde (0.5 % final conc.) (Tsuji and Yanagita 1981) in HDPE Nalgene bottles and

stored in the dark at 4oC for up to 18 months. Flagellates and protozoa were counted,

measured and identified using a combined system of fluorescence, Nomarski interference

and Utermöhl sedimentation (FNU) (Lovejoy et al. 1993) for riverine stations R3 and R4,

the TZ station R5d and R5a and coastal stations R8 and R9. Between 3.6 ml and 100 ml of

43

samples were concentrated in Utermöhl sedimentation chambers, depending on the

concentration of cells and sediments. For the sedimentation process, chambers were kept at

4oC in the dark in a refrigerator. Sedimentation duration depended of the settling column

volume and varied from 12 to 48 h. After the sedimentation, DAPI was gently added (0.1

µg ml-1 final conc.) and left to stain for a minimum of 2 h. The analyses were done with a

Zeiss Axiovert 100 inverted epifluorescence microscope under 1000x magnification using

immersion oil. Counts and identification of river and estuarine samples were laborious due

to high sediment load. For stations R3, R4 and R5d, more than 250 fields and less than 100

cells were observed. At other stations, counts were easier with fewer fields and more cells

observed. Also, differentiation between autotrophic and heterotrophic organisms was

sometime difficult because of bleaching of some pigments due to long samples storage.

Another problem encountered was the non homogeneous sedimentation of diatoms cells

with long setae or colonies (e.g., Chaetoceros sp., Thalassiosira sp., etc.) that created large

aggregations of cells near the chamber walls and thus causing some bias in the estimation

of abundance. Cells were separated in size classes as nanoplankton (2 to 20 µm) and

microplankton (>20 µm). Documentation used to identify organisms included Lebour

(1925), Smith (1950), Bourelly (1968; 1970), Findlay and Kling (1979), Sournia (1986),

Ricard (1987), Chrétinnot-Dinet (1990), Tomas (1997), Bérard Therriault et al. (1999), and

Wehr and Sheath (2003). Mixotrophic organisms were classified as autotrophs. Ciliates

were classified as heterotrophs. However some of the ciliates contained orange-red

fluorescent structures (strong fluorescence under green light, and lower fluorescence under

blue light) which may represent ingested autotrophic cells or functional plastids.

Picophytoplankton was counted separately from protists because the FNU method is not

suitable for picoplankton cells enumeration given the very long sedimentation times they

would require and because they are subject to convective motions in settling chambers

(Lovejoy et al. 1993).

Plankton biovolumes were calculated as in Hillebrand et al. (1999), with biovolume

estimates calculated for each protist taxon. Because many protists species could only be

measured once, missing dimensions were estimated from identification guides, if possible,

or from similar organisms observed during counting. For picophytoplankton, dimensions

used were from Bertrand and Vincent (1994) who studied the picophytoplankton

44

community in another large river, the St. Lawrence. Picocyanobacteria and picoeukaryotes

biovolumes were calculated as spheres with diameters of 1.25 µm and 1.5 µm respectively.

Protist and picophytoplankton carbon biomass was estimated for each group with the

equations given in Menden-Deuer and Lessard (2000): pg C cell-1 = 0.216 x V0.939 for non-

diatom cells and pg C cell-1 = 0.288 x V0.811 for diatoms, where V is the mean cell

biovolume of a group. Heterotrophic picoplankton carbon biomass was estimated with the

widely used value of 20 fg C per cell (Lee and Fuhrman 1987).

Heterotrophic prokaryote production Water samples were fractionated onto Poretics 3 µm polycarbonate filters (47 mm) to

separate free living bacteria from particulate-attached bacteria. The pore size of 3 µm was

used because the size range of bacteria is 0.2 to 2 µm and that we did not want to exclude

the larger bacteria by using filter with 2 µm pores. The filters were previously flushed with

sample water and were changed whenever clogging was apparent.

The 3H-leucine (3H-Leu) incorporation method was used as a measurement of protein

synthesis by heterotrophic picoplankton (Bacteria and Archaea). For each sampling, 5

sterile microvials (2 ml) received 1.25 ml of water sample. Two samples were killed with

trichloroacetic acid (TCA; 5% final conc.) to serve as controls. Microvials were then

inoculated with 3H-Leu (specific activity: 152 Ci mmol-1, Amersham Biosciences). Because

it was not possible to get the results of saturation curve while on the ship, we used a

standard final concentration of 10 nM as proposed by Simon and Azam (1989). Microvials

were incubated in the dark at the simulated in situ temperature for 2 h. Difference from true

in situ temperature was usually very small and generally less than 2oC. One exception was

at R5b, 0 m, where the incubation temperature was 4oC lower than in situ (for incubation

temperatures see Annex 1.3, and for a comparison with in situ temperature see Fig. 1.7 and

1.8, top panels). Protein synthesis was stopped by the addition of TCA (5% final conc.).

The microvials were then stored at 4oC to be processed in the next 24 h or frozen (-20oC) to

be processed later.

To eliminate unincorporated 3H-Leu, a microcentrifugation protocol was followed,

modified from Smith and Azam (1992). After a first centrifugation step (12 min,

13 000 rpm), supernatant was aspirated with a Pasteur pipette connected to a vacuum

45

pump. Care was taken not to aspirate the pellet and to aspirate all remaining supernatant

drops. The pellet was rinsed by adding 1 ml of TCA 5% followed by a second

centrifugation round (12 min, 13 000 rpm). Supernatant was aspired and microvials stored

at -20oC. Microvials received 1 ml of scintillation liquid (OptiPhase ‘HiSafe’ 2; Wallac

Scintillation products) and were then vortexed. After 24 h at ambient temperature, samples

were radio-assayed in a Beckman LS 6500 scintillation system.

The equation used to transform DPM in rate of leucine incorporation per volume

(mmol Leu L-1 h-1) was:

where :

• dpm is disintegrations per minute • 4.5 x 10-13 is a constant representing the number of Curies (Ci) per dpm. • SA is 3H-Leu specific activity in Ci mmol-1. • t is incubation length in h. • V is incubation volume in L.

At station R3, a time series and a saturation curve experiments were performed. For the

time series, the tubes received 10 nM of 3H-Leu and were incubated for 70, 130, 200, 250

and 335 min. Results of this experiment showed that 3H-Leu uptake was linear for at least

335 min (Fig. 1.3) indicating that a 2 h incubation was suitable. For the saturation curves,

tubes received 6.5, 10, 15, 20 and 25 nM of 3H-Leu and were incubated for 2 h (Fig. 1.3).

These measurements demonstrated that a concentration of 10 nM was below required to

saturate 3H-Leu uptake rates of the heterotrophic picoplankton communities, at least in the

river stations. Thus, the bacterial production rates should be considered as conservative

estimates.

Statistical analysis The stations were separated in three categories according to their surface salinities in order

to evaluate general trends in the data set: river (R1 to R4; salinity 0 to 1 psu), transition

zone (TZ) (R5d to R5a; salinity 1 to 10 psu) and sea (R7 to R9; salinity >20 psu). Surface

46

and bottom samples were also considered separately. Statistical analyses were run on

SigmaStat 3.11 for Windows (2004 SysStat Software, Inc.) Correlations between pairs of

variables were tested using Spearman rank order correlation (rs coefficient) because of

strong deviation from normality for many variables. One-way ANOVA or Kruskal-Wallis

one way analyses of variance on ranks was used to test differences among sampling

regions. When there was a significant difference among the sampling zones, Fisher-LSD or

Dunn’s multiple comparison procedure was run to isolate the group or groups that differ

from the others. To test the difference between the surface and the bottom data, t-test or

Mann-Whitney rank sum test was used.

Results

Sampling conditions and meteorological data Air temperature (Fig. 1.4), precipitation (Fig. 1.4) and wind speed data were obtained from

the National Climate Data and Information Archive of Environment Canada website

(Online-a). Data from Inuvik and Tuktoyaktuk show the weather conditions that took place

along the ARDEX transect, while data from Norman Wells exposes the weather affecting

the river watershed upstream. The week prior to the cruise, moderate precipitations

occurred upstream in the watershed as shown by Norman Wells data. During the cruise, the

watershed received large amounts of rain. On July 26th and 27th (sampling of R1 and R4),

the weather was mostly sunny or cloudy with moderate winds. On July 27th, at

Tuktoyaktuk, wind was coming from the south with average speed of 12.5 km h-1. July 28th

(sampling of R9) was rainy with air temperatures around 15oC and north-westerly winds of

18.5 km h-1 (mean speed). On July 29th (no sampling), there was rain accompanied by

storm winds coming from the north (mean speed: 25 km h-1; max speed 35 km h-1) and cold

air temperatures accompanied by surface waves up to 2 m which likely induced bottom

resuspension at R8 and R7. July 30th (sampling of R8 and R7) was cold and cloudy with

mild to moderate south-easterly winds. August 1st (sampling of R5d and R5a) and 2nd

(sampling of R3 and R4) were rainy and misty with moderate to strong winds.

The large amount of rain received during the sampling period induced changes in the

Mackenzie River discharge (Fig. 1.5). Data from two stations of the Water Survey of

Canada (Environment Canada, Online-b) show the increase in river discharge that occurred

47

between the sampling of R1-R4 (July 26th and 27th) and R2-R3 (August 1st). The discharge

increased by 550 m3 s-1 and by 44 m3 s-1 at Arctic Red River and Inuvik (East Channel),

respectively.

Hydrographic gradients Large changes in water column structure occurred across the freshwater-saltwater transition

zone of the Mackenzie River towards offshore Beaufort Sea station R9 (Fig. 1.6). In the

river, there was no stratification of the water column. Water temperature, salinity and Chl a

concentration were constant from the surface to the bottom. In the transition zone (TZ),

there was an intrusion of more saline, colder sea water at the bottom of the water column,

but there was no evident difference in the Chl a concentration between the surface and the

bottom. Stratification of the water column increased along the transect towards the offshore

Beaufort Sea. In the Beaufort Sea stations, water column was highly stratified with the

warm brackish waters of the buoyant river plume overlying the saline and cold layer of

Arctic Ocean waters. The depth of the river plume diminished with increasing distance

from the coast. Also, in the coastal stations, the Chl a concentration was not constant across

the water column. The Chl a concentration was slightly higher in the bottom, more saline

marine layer at R7 and R8, and at R9, there was a marked deep maximum Chl a layer at

21 m, as is often found in the Arctic Ocean (Lovejoy et al. 2007).

Environmental characteristics and gradients

Surface Surface water temperature decreased from the riverine to marine stations (Fig. 1.7 and 1.9;

Annexes 1.1 and 1.4). Mean temperatures of the river, the TZ and the sea differed

significantly from one another (ANOVA, F = 86.321, P < 0.001; Fisher-LSD, P < 0.001)

(Fig. 1.9). River surface water temperature was warm and varied slightly between the

stations (17.9 ± 0.7oC, mean ± S.D.). In the TZ, temperature was colder than in the river

with a mean of 13.5 ± 1.8oC. At the offshore stations, surface temperature continued to

drop and averaged 8.4 ± 1.0oC.

Across the river stations, surface water salinity was 0.13 psu (Fig. 1.7 and 1.9; Annexes 1.1

and 1.4) indicating fresh water of moderate solute content. Surface waters in the TZ were

48

brackish (mean of 4.77 ± 3.32 psu) and increased of 6.62 psu between R5d and R5a. In the

marine stations, water was more saline, but still strongly freshwater influenced with an

average value of 24.69 ± 2.66 psu. The average salinity of the TZ was not significantly

different from the river (Kruskal-Wallis, H = 8.018, P = 0.001; Dunn’s, Q = 1.514,

P > 0.05) nor from the sea (Dunn’s, Q = 1.214, P > 0.05) (Fig. 1.9). Surface water salinity

and temperature were highly negatively correlated (rs = -0.976, P < 0.001) (Annex 1.6).

The Mackenzie River carried a high load of suspended particulate matter (SPM) with an

average of 53.6 ± 11.4 mg L-1 in its surface waters (Fig. 1.7 and 1.10; Annexes 1.1 and 1.4).

In the TZ, there was a sharp decrease of SPM load with a loss of 28.8 mg L-1 along the 10.1

km separating R5d and R5b (TZ mean of 44.9 ± 19.4 mg L-1). In the coastal zone, SPM

concentration continued to decrease with a loss of 14.9 mg L-1 between R7 and R9 (mean

of 26.6 ± 7.9 mg L-1). Mean SPM values were not significantly different between the three

environments (ANOVA, F = 3.514, P = 0.088) (Fig. 1.10) However, the power of the

performed test (0.342) is well below the desired level of 0.800. A trend could be observed

along the transect, with a pronounced decrease of the SPM load in the surface waters

between the Mackenzie River and the Beaufort Sea, with the sharpest decrease occurring in

the TZ. Also, surface SPM showed a strong negative correlation with salinity (rs = -0.830,

P < 0.001) and a positive relationship with temperature (rs = 0.745, P = 0.011) (Annex 1.6).

River surface particulate organic matter (POM) (Fig. 1.7 and 1.10; Annexes 1.1 and 1.4)

varied between 4.3 ± 2.3 mg L-1 (R4) and 10.2 ± 9.7 mg L-1 (R2) with a mean value of

6.2 ± 2.7 mg L-1. In the TZ, mean POM concentration was of 5.0 ± 0.7 mg L-1 and it

continued to decrease in the marine stations to an average of 3.8 ± 0.45 mg L-1. POM was

significantly different between the 3 zones (Kruskal-Wallis, H = 5.982, P = 0.034) (Fig.

1.10). POM was correlated with salinity (rs = -0.685, P = 0.025) and temperature

(rs = 0.673, P = 0.029) (Annex 1.6). There was also a strong relationship between POM

concentration and SPM load (rs = 0.794, P = 0.0038) (Annex 1.6).

Surface turbidity was high in the river (70.7 ± 10.8 NTU) (Fig. 1.7 and 1.10; Annexes 1.1

and 1.4). A turbidity peak was observed at R5d with a value of 97.1 NTU followed by a

sharp drop of 67.4 NTU at R5a. Turbidity continued to decrease towards the offshore

stations to reach an average 8.8 ± 1.5 NTU. The turbidity was significantly different

49

between the 3 zones (Kruskal-Wallis, H = 5.791, P = 0.046) (Fig. 1.10). There was a strong

relationship between turbidity and SPM concentration (rs = 0.79, P < 0.001) (Annex 1.6).

Turbidity was also correlated with salinity (rs = -0.697, P = 0.022) and temperature

(rs = 0.697, P = 0.022) (Annex 1.6).

Particulate organic carbon (POC) (Fig. 1.7 and 1.11; Annexes 1.1 and 1.4) was also high

and averaged 1.22 ± 0.15 mg L-1 in the surface waters of the river. The maximum POC

concentration was in the TZ at R5d which was followed by a major drop towards R5b.

After this point, the decrease in POC load was almost linear and, at R9, the concentration

was below the detection limit of the analyzer (≤ 0.05 mg L-1). POC values were

significantly different between the river and the TZ (ANOVA, F = 7.753, P = 0.017; LSD,

P = 0.025), and between the river and the sea (ANOVA, P = 0.017; LSD, P = 0.007) (Fig.

1.11). POC load was strongly correlated with salinity (rs = -0.855; P < 0.001) and

temperature (rs = 0.818; P = 0.0015). It also showed a strong relationship with SPM

concentration (rs = 0.915; P < 0.001), POM load (rs = 0.709; P = 0.0186) and turbidity

(rs = 0.903; P < 0.001) (Annex 1.6).

The organic content of the seston, measured as the percentage of POM in SPM

(%POM/SPM) was 12.5 ± 7.9 % in the river (8.6 ± 1.3% without the anomalous R2 value),

12.1 ± 3.3% in the TZ and 14.8 ± 2.6 % in the sea. The percentage of POC in SPM

(%POC/SPM) was 2.3 ± 0.4% in the river, 2.1 ± 0.4% in the TZ and 0.5 ± 0.3% in the sea.

%POC/SPM was negatively correlated with salinity (rs = -0.794; P = 0.0038) and positively

correlated with temperature (rs = 0.818; P = 0.0015). The percentage of POC in POM

(%POC/POM) was 22.7 ± 9.6% in the river, 19.2 ± 10.18% in the TZ and 4.0 ± 3.1% in the

sea. %POC/POM was negatively correlated with salinity (rs = -0.830; P < 0.001) and

positively correlated with temperature (rs = 0.770; P = 0.0069).

Particulate organic nitrogen (PON) (Fig. 1.7 and 1.11; Annexes 1.1 and 1.4) was highly

variable in the river, varying between 0.07 ± 0.03 mg L-1 (R2) and 0.22 ± 0.03 mg L-1 (R4)

which was the highest concentration found at the surface throughout the transect. The mean

PON value in the Mackenzie River was 0.14 ± 0.06 mg L-1. In the TZ, PON averaged

0.11 ± 0.02 mg L-1. In the sea, PON concentration was very low and values were below the

detection point of the analyzer, <0.03 mg L-1 at R7 and ≤ 0.01 at R8 and R9. The

50

differences between the river and the TZ (ANOVA, F = 7.292, P = 0.019; LSD, P = 0.042),

and between the river and the sea were significant (ANOVA, F = 7.292, P = 0.019; LSD,

P = 0.007) (Fig. 1.11). PON was highly correlated with salinity (rs = -0.839; P < 0.001) and

to temperature (rs = 0.760; P = 0.0087). PON was also strongly related to SPM load

(rs = 0.900; P < 0.001), turbidity (rs = 0.796; P = 0.00381) and POC concentration

(rs = 0.900; P < 0.001) (Annex 1.6).

Dissolved organic carbon (DOC) (Fig. 1.7 and 1.11; Annexes 1.1 and 1.4) was also highly

variable in the river and in the TZ. In the river, it varied between 3.2 mg L-1 (R2) and

4.8 mg L-1 (R1), and averaged 3.8 ± 0.7 mg L-1. In the TZ, mean DOC concentration was

4.0 ± 0.8 mg L-1 with a large peak of 4.9 mg L-1 at R5a. It subsequently dropped to a mean

value of 2.6 ± 0.3 mg L-1 in the marine zone, where DOC concentrations were less variable.

There was no significant difference between the 3 zones (ANOVA, F = 4.472, P = 0.056)

(but the power of the test (0.455) was well below the desired level of 0.800) (Fig. 1.11),

and DOC was correlated with PON concentration (rs = 0.650; P = 0.038) (Annex 1.6) but

not with temperature nor salinity.

Total Chl a concentration (Fig. 1.7 and 1.11; Annexes 1.2 and 1.4) averaged 2.9 ± 0.3

µg L-1 in the Mackenzie River. In the TZ, Chl a rose to the maximum value met at the

surface along the transect with 3.98 ± 0.03 µg L-1 at R5a. The marine stations

concentrations dropped to values below 1 µg L-1. The percentage of the < 3 µm fraction in

total Chl a biomass (Fig. 1.7; Annexes 1.2 and 1.4) was the lowest is the river with a mean

value of 3.4 ± 2.4%. It increased in the TZ to a mean of 13.2 ± 4.2% and it rose 4-fold in

the sea to reach a mean of 55.7 ± 12.5%. There was no significant difference among the 3

zones for total Chl a (Kruskal-Wallis, H = 5.400, P = 0.051) and for < 3 µm Chl a fraction

(ANOVA, F = 2.044, P = 0.23) (Fig. 1.11). However, the percentage of < 3 µm fraction in

total Chl a biomass significantly differed between the river and the sea (Kruskal-Wallis,

H = 6.250, P = 0.011; Dunn’s, P < 0.05) (Fig. 1.10). Total Chl a concentration was

correlated with POM load (rs = 0.700, P = 0.030) and DOC concentration (rs = 0.717,

P = 0.025) (Annex 1.6). The Chl a fraction < 3 µm was not correlated with any of the

environmental characteristics (Annex 1.6). The percentage of < 3 µm in total Chl a was

highly correlated with salinity (rs = 0.857, P = 0.0018) and temperature (rs = -0.81,

51

P = 0.0096) and with other correlates of the latter two variables: SPM concentration

(rs = -0.833, P = 0.0053), POM load (rs = -0.929, P < 0.001), turbidity (rs = -0.762,

P = 0.021), POC amount (rs = -0.762, P = 0.021) and PON load (rs = 0.766, P = 0.021)

(Annex 1.6).

Bottom River bottom water temperatures (Fig. 1.8 and 1.9; Annexes 1.1 and 1.5) were relatively

constant between the stations with a mean value of 17.5 ± 0.3 oC. In the TZ, mean

temperature decreased to 12.6 ± 1.9 oC and, in the marine stations, it further dropped to

0.1 ± 1.7 oC. River and sea zones were significantly different in terms of bottom water

temperature (ANOVA, F = 119.510, P < 0.001; LSD, P < 0.001). Temperature differences

were also significant between the river and the TZ (LSD, P = 0.013), and between the TZ

and the sea (LSD, P < 0.001) (Fig. 1.9).

In the Mackenzie River, bottom water salinity (Fig. 1.8 and 1.9; Annexes 1.1 and 1.5) had a

mean value of 0.13 psu. It increased in the transition zone to a mean of 10.90 ± 6.19 psu.

The highest augmentation of salinity was between R5a and R7 with a rise of 14.55 psu. Sea

bottom salinity averaged 31.19 ± 1.22 psu. Mean water salinity were significantly different

between the river and the sea bottom (Kruskal-Wallis, H = 6.250, P = 0.011; Dunn’s,

P < 0.05) (Fig. 1.9).

Mean SPM load (Fig. 1.8 and 1.10; Annex 1.1 and 1.5) was of 52.1 ± 19.3 mg L-1 in the

river and of 63.6 ± 31.1 mg L-1 in the TZ. A steep rise was observed at station R7 with a

SPM concentration of 166.8 mg L-1 probably associated with bottom sediment resuspension

due to winds. Towards the offshore stations, SPM amount dropped by a factor of 6. Bottom

SPM load was not significantly different between the three zones (Fig. 1.10). Bottom POM

concentration (Fig. 1.8 and 1.10; Annex 1.1 and 1.5) followed the same pattern of

distribution as SPM, but was an order of magnitude lower in concentration.

In the river, bottom turbidity (Fig. 1.8 and 1.10; Annex 1.1 and 1.5) had a mean value of

71.3 ± 6.1 NTU. The only turbidity measurement available for the TZ came from R5d and

was 48.6 NTU. As for SPM and POM, turbidity showed a sharp increase at R7 with a value

of 125.3 NTU and an important drop toward R9 where it reached 31.3 NTU.

52

Bottom POC concentration (Fig. 1.8 and 1.11; Annex 1.1 and 1.5) had a mean value of

1.63 ± 0.30 mg L-1 in the river. It decreased to 0.91 ± 0.64 mg L-1 in the TZ. At R7, it rose

to 1.98 ± 0.17 mg L-1 and decreased by an order of magnitude towards R9.

The %POM/SPM was relatively stable across zones, averaging 12.1 ± 3.2% in the river,

10.0 ± 4.2% in the TZ and 12.7 ± 2.7% in the sea. The %POC/SPM was of 3.1 ± 0.5% in

the river, 1.4 ± 0.4% in the TZ and of 0.9 ± 0.3% in the sea. The %POC/POM was of 26.1

± 4.6% in the river, 15.6 ± 10.0% in the TZ and 7.7 ± 4.0% in the sea.

River PON concentration averaged 0.19 ± 0.09 mg L-1 at the bottom (Fig. 1.8 and 1.11;

Annex 1.1 and 1.5). In the transition zone, mean PON load decreased to 0.12 ± 0.01 mg L-1.

Towards the offshore stations, PON concentration decreased to fall below the detection

limit of the instrument (≤ 0.01 mg L-1) at R9.

Bottom DOC concentration showed a continuous decrease along the transect (Fig. 1.8 and

1.11; Annex 1.1 and 1.5). In the river, the mean concentration was of 3.5 ± 0.3 mg L-1.

There was a diminution in DOC load in the TZ where it averaged 3.0 ± 0.2 mg L-1. In the

sea, the concentration continued to decrease, with a mean of 1.6 ± 0.3 mg L-1. Bottom DOC

concentration was significantly different between the river and the sea (ANOVA,

F = 48.159, P < 0.001; LSD, P < 0.001) and between the TZ and the sea (ANOVA,

F = 48.159 P < 0.001; LSD, P = 0.002) (Fig. 1.11).

Bottom total Chl a was variable across the transect (Fig. 1.8 and 1.11; Annex 1.2 and 1.5).

In the river, it ranged from 1.99 ± 0.05 µg L-1 (R4) to 3.34 ± 0.07 µg L-1 (R2) with an

average of 2.89 ± 0.78 µg L-1. It was quite stable in the TZ with a mean value of

2.99 ± 0.36 µg L-1. However, in the sea, the Chl a concentration decreased at R7, increased

sharply at R8 with the highest Chl a concentration measured in the transect

(6.64 ± 0.13 µg L-1) and diminished by a factor of almost 5 at R9. The percentage of

< 3 µm fraction in total Chl a biomass was generally low in the river (mean of 3.8 ± 1.2%)

and the TZ (mean of 3.9 ± 0.8%) (Fig. 1.8 and 1.10; Annex 1.2 and 1.5). However, it was

over 10% at R7 and R8, and then subsequently dropped to 1.6% at R9.

53

Surface versus bottom waters There was no significant difference between the surface and the bottom environmental

characteristics in either the river or the TZ (t-tests, P > 0.05) (Fig. 1.9, 1.10 and 1.11).

However, in the sea, there were significant differences between the surface and the bottom

for some characteristics (Fig. 1.9, 1.10 and 1.11), reflecting the stratification of the water

column in this zone. Mean surface salinity increased significantly of 6.5 psu towards the

bottom (t = -3.849, P = 0.018) (Fig. 1.9). Surface temperature was significantly higher than

the bottom temperature by 8.3oC (t = 7.232, P = 0.002) (Fig 1.9). The percent contribution

of < 3 µm Chl a to total Chl a biomass was 6 times higher at the surface than at the bottom

(t = 5.684, P = 0.005) (Fig. 1.10). Finally, DOC concentration decreased significantly from

2.6 mg L-1 at the surface to 1.57 mg L-1 at the bottom (t = 4.543, P = 0.01) (Fig. 1.11).

Microbial population gradients

Heterotrophic picoplankton In terms of cell concentrations, heterotrophic picoplankton dominated the picoplankton

community by 1 to 4 orders of magnitude over the picophytoplankton community

(picocyanobacteria and picoeukaryotes) (Table 1.1; Fig. 1.13). Across the transect,

heterotrophic picoplankton abundance ranged between 4.3 x 105 cells ml-1 (R9) and

9.6 x 105 cells ml-1 (R2) at the surface and between 1.7 x 105 cells ml-1 (R9) and

9.2 x 105 cells ml-1 (R2) at the bottom. There was no significant difference in heterotrophic

picoplankton abundance between the three zones and between the surface and the bottom in

each zone (Fig. 1.13).

Picophytoplankton Surface autotrophic picoplankton abundance (Table 1.1; Fig. 1.12 and 1.13) decreased

along the transect towards the marine stations ranging between 45.1 x 103 cells ml-1 in the

river (R2) to 0.03 x 103 cells ml-1 in the sea (R9). Between R7 and R9, picophytoplankton

abundance dropped of 2 orders of magnitude. Autotrophic picoplankton abundance was

significantly different between the river and the sea (ANOVA, F = 10.350, P = 0.008; LSD,

P = 0.003), and between the river and the transition zone (ANOVA, F = 10.350, P = 0.008;

LSD, P = 0.021) (Fig. 1.13). The bottom picophytoplankton community showed the same

pattern of diminution towards the coastal zone than at the surface, but concentrations

54

dropped of 2 orders of magnitude between R7 and R8 and the concentrations were not

significantly different between the three zones (ANOVA, F = 5.605, P = 0.053). The

difference between the surface and the bottom for picophytoplankton abundance was not

significant in the river, in the TZ and in the sea (Fig. 1.13). Autotrophic picoplankton was

composed of two different categories: the picocyanobacteria and the picoeukaryotes (Figure

1.12). The contribution of the picoeukaryotes to picophytoplankton abundance increased in

the coastal zone. In the river, picoeukaryotes represented < 1% of the total

picophytoplankton community but at the surface of R8, in the sea, it composed 26% of it.

The largest contribution of picoeukaryotes to autotrophic picoplankton abundance was at

the bottom of R9 where it reached 58%. In terms of carbon biomass, picoeukaryotes

composed 70% of picoplankton community at R9, 21 m. In the river and the TZ, the

picocyanobacteria were observed alone, in pairs or in large colonies (even more than 24

individuals per colony). Solitary cells seemed to represent a minority of the community in

the river and the TZ, but in the sea they were the majority of the picophytoplanktonic cells.

Many kinds of colonies could be observed: spherical cells were in tight groups or slack

clusters (probably maintained together by mucilage) and rod cells were arranged in chains.

There were also two types of picocyanobacteria: phycoerythrin rich cells which fluoresced

orange under green light and yellow under blue light, and phycocyanin rich cells which

fluoresced orange-red under green light and pale red under blue light as described by

Bertrand and Vincent (1994).

Protists Protist community structure (Fig. 1.14, Table 1.2 and Annex 1.7) showed large changes

across the transect both in terms of concentration and species composition. The protists

were significantly more abundant in the TZ and the river than in the sea where their

concentration dropped by 1 order of magnitude (Fig. 1.15) (R+TZ vs S, t = 4.520,

P = 0.011). At each station, at least 25% of the organisms could not be classified in a

specific taxonomic group, to a maximum of 52.5% of unidentified organisms at R4. On

average, the most abundant identified groups in the Mackenzie River were in order the

Bacillariophyceae (24% of total protist count), Cryptophyceae (10%), identified

heterotrophic groups (10%) and Chrysophyceae (8%). In the TZ, Chlorophyceae (19%),

Cryptophyceae (17%), Bacillariophyceae (16%) and Chrysophyceae (5%) were the most

55

abundant groups. The Beaufort Sea presented high abundance of autotrophic Dinophyceae

(16%), Prymnesiophyceae (14%), Chrysophyceae (10%), Prasinophyceae (9%),

Bacillariophyceae (8%) and identified heterotrophic groups (8%).

The shift in protist community composition occurred in the TZ and it was well illustrated

by the change in species distribution along the transect. Protist species (Table 1.2 and

Annex 1.7) identified at R3 and R4 were typical of freshwater habitats. At stations R8 and

R9, only marine species were recognized. However, in the TZ, both freshwater and marine

species were observed. Marine species were more abundant at R5a with the diatoms

Chaetoceros sp. and Pseudo-nitzschia sp., the Dinophyceae Heterocapsa rotundata

(Lohmann) Hansen, a Raphidophyceae species, an Ebriidean species (Ebria tripartita

(Schumann) Lemmermann) and the heterotrophic flagellate cf. Telonema subtile

Griessmann. At R5d, one marine species of diatom, Chaetoceros sp., and a marine

heterotrophic flagellate, cf. Telonema subtile Griessmann, provided evidence of mixing of

riverine and coastal sea waters.

In terms of biomass, the protist community showed a different pattern across the

freshwater-saltwater transition zone compared to their abundance. There was no significant

difference in protist carbon biomass between the three zones (Kruskal-Wallis, H = 2.000,

P = 0.533). However, a biomass peak was evident at R5a with 118.3 ng C ml-1, which was

3 (R3) to 7 (R4) times higher than the biomass found at other stations (Fig. 1.16). The

taxonomic groups contributing the most to biomass were Bacillariophyceae (42%),

identified heterotrophic groups (11%) and Cryptophyceae (8%) in the Mackenzie River. In

the TZ, they were Raphidophyceae (28%), identified heterotrophic groups (23%),

Bacillariophyceae (16%) and Cryptophyceae (12%). In the Beaufort Sea, the dominant

groups in terms of biomass were autotrophic Dinophyceae (45%), Prymnesiophyceae

(18%), Bacillariophyceae (11%) and ciliates (9%).

Heterotrophic protists (excluding those of unknown trophic status) were 35 ± 21% of total

protists carbon biomass in the river. It was similar in the TZ with 25 ± 4% of carbon

biomass as heterotrophic protists. In the Beaufort Sea, the percentage of heterotrophic

protists diminished to 12 ± 3% of the total protist carbon biomass. The difference between

the three zones was not significant (Kruskal-Wallis, H = 3.429, P = 0.333).

56

Total microbial community If we consider all the microbial compartments together, heterotrophic picoplankton were

the numerically most abundant microorganisms, constituting 96% (R3) to 99.9% (R9) of

total abundance. However, it contributed from 10% (R5a) to 26% (R8) of total carbon

biomass. The largest percentage of carbon biomass was that of nanoplanktonic and

microplanktonic protist cells (Fig. 1.16). The highest carbon biomass was found in the TZ

at R5a with almost 135 ng C ml-1.

The percentage contribution of heterotrophs to total microbial community carbon biomass

was not significantly different between the river plus TZ and the sea (t = 0.452, P = 0.682).

The percentage of carbon biomass as heterotrophic microbes varied between 32% (R9) and

41% (R5d) (when the organisms with unknown trophic state were excluded).

Heterotrophic activity gradients Surface measurements of total microbial 3H-Leu uptake (Fig. 1.16 and 1.17; Annex 1.3)

showed 2 levels of activity in the Mackenzie River. Stations R1 and R4 had heterotrophic

activity (mean of 125.9 pmol L-1 h-1) more than 3 times higher than R2 and R3 (mean of

38.5 pmol L-1 h-1). The first two stations of the TZ showed an activity level similar to R2

and R3, but production rates rose downstream to reach 106.2 pmol L-1 h-1 at R5a. In the sea,

the total picoplankton 3H-leu uptake rate averaged 57.3 ± 15.5 pmol L-1 h-1.

Surface free living bacterial 3H-Leu uptake rates (<3 µm fraction) (Fig. 1.16 and 1.17;

Annex 1.3) were smallest in the river zone with a mean of 3.8 ± 1.5 pmol L-1 h-1. It

increased in the TZ to average 18.8 ± 10.8 pmol L-1 h-1. In the Beaufort Sea, the mean 3H-leu uptake rate of the free living bacteria doubled compared to the TZ mean rate

(41.0 ± 19.7 pmol L-1 h-1).

In the surface waters of the Mackenzie River, particle-attached bacteria (> 3 µm) (Fig. 1.16

and 1.17; Annex 1.3) were responsible for the largest fraction of the total heterotrophic

picoplankton activity, accounting for 94 ± 5% of the activity. In the TZ, the importance of

the particle-bound bacteria in the heterotrophic production decreased to 74 ± 2%.

Heterotrophic bacterial production in the Beaufort Sea was not dominated by the particle-

attached bacteria which accounted for 31 ± 16% of the activity.

57

The 3 zones were not significantly different from one another for surface total 3H-Leu

uptake rates (Fig.1.17). However, there were significant differences between the river and

the sea for surface free living bacterial uptake rates (Kruskal-Wallis, H = 6.444 P = 0.008;

Dunn’s, P < 0.05) and for the contribution of particle-bound bacteria to total heterotrophic

activity (Kruskal-Wallis, H = 7.000, P = 0.005; Dunn’s, P < 0.05).

In the river and the TZ, there was no significant difference between the surface and the

bottom values for the heterotrophic microbial activity data (Fig. 1.17). In the sea, < 3 µm 3H-leu uptake rates were significantly higher at the surface than at the bottom (t = 3.400,

P = 0.027) and particle-attached bacteria accounted for a higher fraction of the total

heterotrophic activity at the bottom (almost 3 times higher) than at the surface (t = -6.268,

P = 0.003).

Bottom waters of the TZ had significantly higher total 3H-Leu uptake relatively to the sea

(ANOVA, F = 12.389, P = 0.012; LSD, P = 0.004) and the river (ANOVA, F = 12.389,

P = 0.012; LSD, P = 0.03) (Fig. 1.17).

Overall relationships at the surface between microbial and environmental characteristics Surface picophytoplankton abundance showed a strong negative correlation to salinity

(rs = -0.867, P < 0.001) and a strong positive relationship with temperature (rs = 0.867,

P < 0.001) (Annex 1.6). It was also correlated with particulate material load as SPM

(rs = 0.661, P = 0.03), POM (rs = 0.806, P = 0.003) and POC concentrations (rs = 0.636,

P = 0.04). Total Chl a (rs = 0.817, P = 0.004), the contribution of < 3 µm fraction to total

Chl a (rs = -0.929, P < 0.001), and bacterial abundance (rs = 0.750, P = 0.04) were also

correlated with picophytoplankton abundance. Picophytoplankton concentration was also

highly correlated with NO3 concentrations (rs = 0.863; P < 0.0001) determined by

Emmerton (2006) during ARDEX cruise.

Total 3H-Leu uptake was correlated with total nitrogen (TN) (Emmerton 2006) (rs = 0.736;

P = 0.01) but not with other variables (Annex 1.6). Free living heterotrophic picoplankton 3H-Leu uptake (<3 µm) was negatively correlated with temperature (rs = -0.817, P = 0.004),

POM concentration (rs = -0.833, P=0.002), NTU (rs = -0.683, P = 0.04) and

picophytoplankton abundance (rs = -0.883, P < 0.001) (Annex 1.6). It had a positive

58

relationship with salinity (rs = 0.767, P = 0.01), < 3 µm Chl a fraction (rs = 0.762,

P = 0.021) and the percentage of < 3 µm Chl a in total Chl a (rs = 0.929, P < 0.001). It was

also negatively correlated with %POC/SPM (rs = -0.700; P = 0.03) and NO3 (Emmerton

2006) (rs = -0.845; P < 0.0001).

Attached bacteria 3H-Leu uptake (total – free living 3H-Leu) was negatively correlated with

salinity (rs = -0.817; P = 0.004) and with the percentage of < 3 µm Chl a in total Chl a

(rs = -0.690; P = 0.047). It showed a positive relationship with temperature (rs = 0.850;

P < 0.0001), DOC (rs = 0.867; P < 0.0001), total Chl a (rs = 0.700; P = 0.03),

picophytoplankton (rs = 0.767; P = 0.01), %POC/SPM (rs = 0.733; P = 0.02), %POC/POM

(rs = 0.700; P = 0.03), %Chl a/SPM (rs = 0.767; P = 0.01) and %PON/SPM (rs = 0.833;

P = 0.002).

The percentage of the attached bacteria to total 3H-Leu uptake was positively correlated

with temperature (rs = 0.933, P < 0.001), DOC (rs = 0.717, P = 0.03), POC (rs = 0.667,

P = 0.04), PON (rs = 0.678, P = 0.04) and picophytoplankton abundance (rs = 0.867,

P < 0.001) (Annex 1.6). It was negatively correlated with salinity (rs = -0.900, P < 0.001),

< 3 µm 3H-Leu uptake (rs = -0.867, P < 0.001) and the percentage of < 3 µm fraction to

total Chl a (rs = -0.857, P = 0.002). It was also positively correlated with %POC/SPM

(rs = 0.750; P = 0.02), % POC/POM (rs = 0.700; P < 0.03) and NO3 (Emmerton 2006)

(rs = -0.787; P = 0.009), which may simply reflect the cross-correlations with salinity.

Discussion This study was designed to evaluate the gradients in microbial community structure and

heterotrophic picoplankton production across the freshwater-saltwater transition zone of the

Mackenzie River and Beaufort Sea system. Although our sampling lacked seasonal

resolution, it was done during the most active period of the system in terms of open water

conditions and biological processes. In addition, this project was one of the first combining

the analysis of the microbial community structure and bacterial productivity in the

Mackenzie River estuary, and extended previous observations by Parsons et al. (1988;

1989) who were the firsts to realize the potential importance of microbiota in this system.

59

We observed large spatial gradients in environmental characteristics along our transect.

Surface salinity was about 25 psu higher and temperature about 10oC lower in the coastal

zone than in the river (Fig.1.7 and 1.9). In addition, POC and PON decreased significantly

between the Mackenzie River and the Beaufort Sea (Fig.1.11). Mackenzie River had a well

mixed water column which began to stratify in the transition zone (Fig.1.6), but it was only

in the coastal zone that temperature and salinity became significantly different between the

surface and the bottom layer (Fig. 1.9). The surface layer associated with the fluvial plume

of the Mackenzie River was warm and brackish while the bottom layer was saline and cold

water coming from the Arctic Ocean.

Microbial community structure across the freshwater-saltwater transition zone showed

marked changes in terms of picophytoplankton (Fig. 1.12 and 1.13; Table 1.1) and protist

(Fig. 1.14 and 1.16) abundance, and protist species (Table 1.2). Contrary to our hypothesis,

picocyanobacteria were more abundant in the river and the TZ than in the Beaufort Sea

(Table 1.1) and autotrophs were dominating the protist community in the three zones

(Table 1.3). The percentage of microbial biomass as heterotrophs did not change

significantly along the transect contrary to our predictions (Table 1.4), but in terms of

abundance, heterotrophic picoplankton dominated at all stations (Fig. 1.16).

Heterotrophic picoplankton metabolism (Fig. 1.17 and 1.18) did not increase towards the

marine zone nor did it vary significantly among zones. Consistent with our hypothesis,

particle-attached bacteria were a major component of total bacterial metabolism accounting

for 16.4 to 99.6% of total production and their importance increased with the concentration

of POC (Annex 1.6). The percentage of the production due to particle-bound bacteria was

significantly higher in the river than in the coastal zone (Fig. 1.18). It was also higher at the

bottom of the Beaufort Sea than at the surface, probably because of the higher particulate

load in this layer (Fig. 1.11).

Spatial gradients The Mackenzie River is the major inflow to the Canadian arctic coastal environment,

providing 90% of the water and 99% of the suspended particulate material discharge into

the Beaufort Sea region (Rachold et al. 2004). The Mackenzie River is unique among all

arctic rivers because it discharges annually more POC than DOC to the Arctic Ocean

60

(Dittmar and Kattner 2003). However, at the time of our sampling, DOC concentration was

higher than POC concentration (Fig. 1.7 and 1.8; Annex 1.1).

We observed DOC concentrations varying between 2.2 and 4.9 mg L-1 along our transect

(Fig. 1.7 and 1.8; Annex 1.1) which are within the range of measurements by Garneau et al.

(2006) in the same region in October 2002 (0.9 to 5.7 mg L-1). Meon and Amon (2004)

found in the Ob and Yenisei Rivers surface waters DOC concentrations ranging between

8.2 and 14.0 mg L-1 in August-September 2001, which are higher than what we measured

in the Mackenzie River (3.2 to 4.8 mg L-1). They also measured higher DOC concentrations

in the Ob and Yenisei Estuaries (4.4 to 6.6 mg L-1), than what we observed in the

Mackenzie transition zone. In the Kara Sea, that receives the inputs from the Ob and

Yenisei, DOC varied from 1.7 to 4.4 mg L-1 which is in the same range that we measured in

the Beaufort Sea. Meon and Amon (2004) used precombusted GF/F filters to separate DOC

from POC which is different from our method. However, Nayar and Chou (2003) found

that precombusted GF/F filters performed as well as 0.2 µm membrane filters in retaining

particles.

During our sampling, surface waters of the Mackenzie River contained an average of

54 mg L-1 of SPM (Fig. 1.7; Annex 1.1), which was higher than the value of 35.2 mg L-1

measured in October 2002 by Garneau et al. (2006). This difference may be explained by

the higher water discharge during summer than during fall in the Mackenzie River. The

SPM content of the riverine runoff entering the Arctic Ocean varies from 8 to 207 mg L-1,

with an average of 36 mg L-1 for the Russian Arctic rivers and 63 mg L-1 for all arctic rivers

(Rachold et al. 2004). The major part of the particulate matter of the riverine runoff

precipitates in estuaries and adjacent shelf area (60-90%) (Vetrov and Romankevich 2004),

i.e. the marginal filter between the river and seawater mixing zones (Lisitsyn 1995). We

observed that this precipitation of SPM for the Mackenzie occurred between the first station

of the transition zone (R5a) and the first station of the coastal zone (R7), SPM

concentration decreasing by 65% (Fig. 1.7; Annex 1.1). The proportion of organic carbon

in SPM composition averaged 12.5% in the Ob River estuary and 6.5% in the Yenisei River

estuary in September 1997 (Vetrov and Romankevich 2004) and increased toward the sea

stations. These values differ from ours for the Mackenzie estuary (TZ) where the

61

%POC/SPM was on average only 2.1% and decreased further toward the coastal stations.

In the freshwater part of the Mackenzie River, we found surface %POC/SPM of 2.3%.

Lobbes et al. (2000) found that the contribution of organic carbon to SPM ranged between

0.4 to 5.7 % with an average of 2.0% (note that they use GF/F filters) in twelve Siberian

Arctic rivers in summer 1994-1995. This last ratio is in accordance with our results.

Our Chl a concentrations in surface waters of the ARDEX transect (0.2 to 3.9 µg L-1) (Fig.

1.7; Annex 1.2) were higher than in October 2002 when Chl a biomass ranged between 0.2

and 1.4 µg L-1 (Garneau et al. 2006) in the Mackenzie River and Beaufort Sea systems. In

July-August 1995, total Chl a concentration in the freshwater part of the Great Whale River

range between 1.2 and 1.6 µg L-1 (Rae and Vincent 1998b) being smaller than what we

observed in the Mackenzie River. Our values fell within the range for Siberian systems. For

example, in August-September 2001, measurements in the Ob and Yenisei Rivers, estuaries

and Kara Sea showed that Chl a concentrations varied between ≤ 0.9 µg L-1 and 4.2 µg L-1,

with the highest biomass in the rivers and the lower in the sea (Meon and Amon 2004).

Microbial community structure Surface heterotrophic picoplankton abundances measured in the Mackenzie River during

the ARDEX cruise were smaller than those measured by Garneau et al. (2006) in

September-October 2002 which were of 1.4 to 1.8 x 106 cells ml-1. Because sediments can

interfere in the evaluation of cell concentration (Kepner and Pratt 1993), this difference

could be due to lower sediment concentration in September-October (35.2 mg L-1) than in

July-August (53.6 mg L-1). The difference could also be due to the long storage of our

preserved water samples before filtration. The effect of long-term storage of preserved

seawater samples has been shown to affect the concentration of bacteria such that counts

from stored, preserved seawater samples may results in large underestimates of bacterial

numbers and importance in the marine ecosystem (Turley 1993). The loss of bacterial cells

in preserved seawater samples was earlier believed to be mostly due to attachment to the

inner surfaces of sample bottles (Turley 1993), but more recent studies indicate that

protease activity may be a major cause of bacterial loss in preserved samples (Gundersen et

al. 1996). Bacterial concentrations measured in other arctic rivers are also larger than our

values. For example, Meon and Amon (2004) measured bacterial concentrations of 1.2 to

62

2.5 x 106 cells ml-1 in the Ob and Yenisei Rivers and estuaries. In the Lena River and delta,

heterotrophic picoplankton abundance ranged between 6.0 x 105 and 8.3 x 106 cells ml-1

(Saliot et al. 1996). In the TZ, Garneau et al. (2006) found heterotrophic prokaryote

concentrations of 3.9 to 5.7 x 105 cells ml-1 and of 3.6 x 105 cells ml-1 in the coastal zone

which are consistent with our measurements in those zones. Meon and Amon (2004)

evaluated bacterial concentrations of 2.3 to 4.7 x 105 cells ml-1 in the Kara Sea and Saliot et

al. (1996) found bacterial abundance varying between 2.0 x 105 and 2.0 x 106 cells ml-1 in

the Laptev Sea. Compared to our TZ data, Parsons et al. (1988; 1989) found smaller

concentrations of heterotrophic picoplankton in summer 1986 (104 cells ml-1) and higher

concentrations in 1987 (> 106 cells ml-1) that could be explained by advective processes

caused by on-shore winds in 1987. An interesting study from Rae and Vincent (1998a)

showed in the Great Whale River heterotrophic averaged bacterial concentration of 106

cells ml-1, but with less than 10% of these cells that were metabolically active as measured

with CTC (5-cyano-2,3-ditolyl tetrazolium chloride). Zweifel and Hagström (1995) showed

that only a fraction of the particles that are enumerated as bacteria by traditional

epifluorescence techniques are truly nucleoid-containing cells, the remaining cells are

considered nonfunctional “bacterial ghosts”. The occurrence of these non-functional cells

in the environment could explain in part the lack of relationship between our bacterial

production and abundance measurements.

Cyanobacteria are a major component of the microbiota in arctic lakes and streams

(Vincent and Hobbie 2000). In our study, we considered picoplankton as both single and

colonial picocyanobacteria forms. Picophytoplankton abundance dropped by two orders of

magnitude between R7 and R9 and was strongly correlated with salinity and temperature.

In autumn 2002, sampling in the Mackenzie River showed that picocyanobacteria

populations passing through 3 µm pore size filters (thus excluding colonial forms) were one

order of magnitude more abundant in the Mackenzie River and estuary than in the Beaufort

Sea (Garneau et al. 2006; Waleron et al. 2007). They found picocyanobacteria

concentration one order of magnitude smaller than ours in the surface waters of the

Mackenzie River what was probably due to the exclusion of colonial forms caused by the

prefiltration through 3 µm. During this sampling period, phylogenetic analysis found that

the picocyanobacteria met were in majority composed of species of the Synechococcus

63

group and that the offshore picocyanobacteria were probably originating from the

Mackenzie River (Waleron et al. 2007). Even though our picoeukaryote results are limited

in scope, we observed higher proportions of picoeukaryotic cells in total picophytoplankton

abundance in the Beaufort Sea (26% at R8, 0 m) than in the Mackenzie River (less than

1%). We also observed that the proportion of picoeukaryotes increased towards the bottom

in the Beaufort Sea to reach 58% at R9, 21m. Our results are supported by Waleron et al.

(2007) who also found higher proportions of picoeukaryotes in the marine zone than in the

freshwater Mackenzie River, and who also observed a greater representation of

picoeukaryotes in deep relative to surface waters of the Beaufort Sea. Rae and Vincent

(1998a) found in August 1995, concentration of photosynthetic picoplankton of

103 cells ml-1 with a dominance of picocyanobacteria and a very small representation of

picoeukaryotes (< 1% total picophytoplankton abundance) in the Great Whale River.

Sorokin and Sorokin (1996) observed the presence of picocyanobacteria in the freshwater

part of the Lena River and their complete disappearance in the mixing zone of the river

with the Laptev Sea. Consistent with our findings, a study by Bertrand and Vincent (1994)

in the St. Lawrence River estuary showed that cell pairs or larger cell aggregates made a

sizeable contribution to the total picophytoplankton cell counts. They also found higher

picophytoplankton concentrations in the freshwater section of the estuary. The presence of

loose or tightly clustered colonies may have an adaptive function, both in increasing the

efficiency of nutrient recycling and in providing protection against predators (Callieri and

Stockner 2002). Heterotrophic and mixotrophic nanoflagellates and small ciliates are the

most important autotrophic picoplankton grazers (Callieri and Stockner 2002).

Picocyanobacteria are in a size range suitable for utilization by nauplii and early copepodite

stages of copepods as well as rotifers. Nevertheless, protozoa are the predominant

organisms of microbial food webs feeding directly on autotrophic picoplankton (Callieri

and Stockner 2002). We observed a significant fraction of heterotrophic and mixotrophic

protists in the Mackenzie River and Beaufort Sea systems (Fig. 1.14 and 1.15; Table 1.2

and 1.3).

Our picoplankton counting was on filtered samples that had been frozen without fixation.

This method may not be the best way to evaluate the picoeukaryote concentration because

many eukaryotic picoplankton such as flagellates that lack rigid cell wall cannot survive

64

filtration or freezing without previous fixation (MacIsaac and Stockner 1993). This could

explain why picoeukaryotes have not been observed in every sample and why we should

consider our picoeukaryote results with care. In addition, degradation of chlorophylls to

weaker fluorescing phaeophytins that absorb at shorter wavelengths can occur during

fixation, freezing or storage that further decreases the facility to detect picoeukaryote cells

(MacIsaac and Stockner 1993). In contrast, picocyanobacteria are the easiest types of cells

to preserve since these prokaryotic cells can often withstand freezing, filtration and even

desiccation on filters, without prefixation (MacIsaac and Stockner 1993). Also,

phycobillins may increase in fluorescence immediately after preservation (chemical

fixatives or freezing) due to the uncoupling of energy transfer to chlorophyll a or alteration

of their protein complexes (MacIsaac and Stockner 1993).

Few studies have been done on protist community structures in large arctic rivers. We

found protist cell abundance of 1238 cells L-1 in the Mackenzie River dominated by

Bacillariophyceae, Cryptophyceae, heterotrophic groups and Chrysophyceae (Fig. 1.14). In

the freshwater zone of the Great Whale River, in August 1995, the cell abundance of

plankton > 2 µm was of 977 cells ml-1. The most abundant protist taxa were

Chlorophyceae, Bacillariophyceae and Chrysophyceae (Rae and Vincent 1998a). In the

Lena River, the major phytoplanktonic groups observed by Sorokin and Sorokin (1996)

were Bacillariophyceae, nanoplanktonic phytoflagellates and coenobial cyanobacteria. This

latter group may include colonies of picocyanobacteria that we observed in the Mackenzie

River.

In the Beaufort Sea, we observed dominance of the protist community by Dinophyceae,

Prymnesiophyceae, Chrysophyceae, Prasinophyceae, Bacillariophyceae and heterotrophic

taxa (Fig. 1.14). Consistent with our observations, a review of Sakshaug (2004) concluded

that the most common algal groups in the arctic and subarctic seas are Bacillariophyceae,

Chrysophyceae, Dinophyceae, Prymnesiophyceae and green flagellates, and that most

phytoplankton are nanoplankton. However, HPLC analysis in the Chukchi and Eastern

Beaufort Seas showed that low productivity and biomass are observed at the surface and

Prasinophyceae, Haptophyceae (syn. Prymnesiophyceae) and Bacillariophyceae are

identified as major contributors to the shelf community (Hill et al. 2005).

65

Some uncertainties may have affected our protists community analysis. First, the long term

storage may have caused cell loss and fluorescence fading of the photosynthetic pigments.

The high concentration of sediments in riverine and TZ samples have complicated the

identification of the organisms to taxonomic level. Also, the large concentration of

sediments prevented us to sediment a large volume of water, what cause the small quantity

of cells observed in river samples. When dealing with marine phytoplankton, non-random

(aggregated) distribution of the settled organisms are often encountered, particularly when

chain forming species are present (Hasle 1978). We observed such aggregates during our

analysis. Those aggregates contained mainly diatom cells from the genus Chaetoceros and

Thalassiosira. The use of tall cylinder should be done with caution because of attachment

of organisms to the chamber wall, especially of chain-forming species of the setae-bearing

genus Chaetoceros (Hasle 1978).

Heterotrophic picoplankton 3H-leucine uptake We observed large difference in the total bacterial activity of riverine stations sampled at

the beginning (R1 and R4) and at the end of the cruise (R2 and R3). R1 and R4 gave

production rates 3 times higher than R2 and R3. The difference was mainly due to changes

in the attached bacterial production (Fig. 1.17; Annex 1.3). This difference may be due to

changes in the water mass characteristics between the sampling periods (almost 6 days).

There were high quantities of precipitations in the watershed during the sampling mission

that induced a significant increase in the discharge of the Mackenzie River (Fig. 1.5).

Heavy rains bringing large quantities of water may dilute the leachate and what is often

observed is an increase in DOC at the beginning of the flood and a decrease, due to

dilution, in the second part of the flood (Cauwet 2002). In the Mackenzie River, DOC was

higher at R1 and slightly higher at R4 than at R2 and R3, but it is possible that DOC

composition might have varied despite the relatively constant values of bulk DOC.

Meon and Amon (2004) measured bacterial 3H-Leu uptakes in the Ob and Yenisei Rivers,

and the Kara Sea in August and September 2001. Mean bacterial production in surface

waters was highest in the rivers Ob (295 pM h-1) and Yenisei (197 pM h-1), decreased

towards stations in the estuaries (105 pM h-1) and was lowest in the Kara Sea (45 pM h-1).

Their fluvial and estuarine measurements are higher than what we found in the Mackenzie

66

system. However, this could be due to the fact that our estimates of bacterial production are

conservative since we might have not saturated the uptake with 10 nM of leucine (Fig. 1.3).

However, their production data at the surface of the Kara Sea are similar to our estimate in

the coastal Beaufort Sea. Meon and Amon (2004) also found that bacterial production

decreased with increasing depth at stations in the estuaries and the Kara Sea. It was also the

case with our measurements in the Beaufort Sea, but not in the TZ where bottom

production was higher than surface production. Interestingly, Meon and Amon (2004)

found a significant correlation between total bacterial production and Chl a concentrations

in the rivers and the Kara Sea. In our case, there was no correlation between these two

parameters. However, we found a significant positive correlation between particle-bound

bacterial production and total Chl a which can be considered as a proxy of newly produced

organic matter.

Attached bacteria (>3 µm) were a major component of total bacterial production in the

Mackenzie River and its estuarine freshwater-saltwater TZ. In the river, particle-attached

bacteria accounted for 84% to 99% of total bacterial activity and for 72% (surface) to

almost 100% (bottom) in the estuary. Previous sampling in October 2002 in the TZ of the

Mackenzie River and the Beaufort Sea (salinity 25.4 psu) showed that 68% of surface 3H-Leu uptake was due to particle-bound bacteria (>3 µm) (Garneau et al. 2006), a value

similar to what we measured. Droppo et al. (1998) found that bacteria were an important

constituent of Mackenzie River Delta flocs. This is consistent with studies on turbid rivers

and estuaries elsewhere. In the Columbia River and estuary, particle-attached bacterial

carbon production (>3 µm) represented on average 90% of total bacterial production and

was positively correlated with SPM and POC, but free living bacterial carbon production

was not (Crump et al. 1998). Vincent et al. (1996) found that there was a very large

increase in the contribution of bacteria attached to particles >2 µm in the frontal zone of

increasing salinity and turbidity in the St. Lawrence River estuary, passing from a non

significant contribution to an average of 46% (range 40-60%) of the total activity. In the

Tamar estuary, the activity of the attached bacteria fraction followed the concentration of

suspended particles and was a major proportion of the total bacterial production in the

maximum turbidity zone (Plummer et al. 1987). The relative proportion of aggregate-

associated bacteria to total bacterial numbers varies greatly, mainly depending on the

67

abundance of aggregates. In general, in riverine and estuarine systems, they may contribute

between 14 to 90% to total bacterial abundance or production (reviewed by Zimmermann-

Timm 2002).

We also observed that the contribution of particle-bound bacteria to total bacterial

production was significantly higher at the bottom of the coastal zone than at the surface.

This difference may be due to the higher concentration of particles in the bottom waters of

the Beaufort Sea, even if this difference what not statistically significant.. Turbidity, SPM,

POM, POC, PON, %POC/SPM, DOC and Chl a were all higher at the bottom (Fig. 1.10

and 1.11; Annex 1.1 and 1.2), with only DOC showing a significant difference. A study on

fine particles (< 2 to 10 µm) in the Lena River-Laptev Sea system showed that in coastal

waters, bottom samples yielded higher particle concentrations than surface samples due to

particle resuspension and the presence of a marked halocline, which prevented vertical

mixing (Moreira-Turcq and Martin 1998).

Particles are hot spots of microbial activity (Azam et al. 1993). Growth in aggregates

provides many benefits to microbes. Aggregates form microhabitats where microorganisms

can interact (commensalism, paratism, and the exchange of genetic material), provide

protection from some bacterivores, and can increase nutrient uptake through attachment to

organic materials, increased availability of organic materials at surfaces and improved mass

transfer (Logan and Hunt 1987). Another advantage of living on aggregates is that riverine

POM is considered to be less degraded than DOM, although both are derived

predominantly from refractory compounds of vascular plants (Dittmar and Kattner 2003).

This lesser degree of diagenesis suggests that the riverine POM may have a higher lability

(i.e., bioreactivity) than DOM.

Bacterial carbon cycling

The saturation curve measured in the river showed that 10 nM was below required for

maximum leucine uptake by bacteria, which began around 20 nM of 3H-Leu (Fig. 1.3).

Although we did not measure saturation values in the TZ and in the Sea, we can make a

conservative estimation of net bacterial C production with the following theoretical formula

(Simon and Azam 1989):

68

where mol Leu = moles of exogenous leucine incorporated; 100/7.3 = 100/mol% of leucine

in protein; 131.2 = formula weight of leucine (grams per mole of leucine); 2 = internal

isotope dilution; and 0.86 is the factor used to transform protein production in carbon

production (grams of carbon per grams of protein). This gives a conversion factor of

3091.3 grams of carbon per mole of leucine, a value often used in literature (see Annex 1.8

for bacterial carbon production (BCP) data).

We calculated depth integrated net BCP for each station by using a rectangular integration

formula (see results in Annex 1.8) which gave an estimation of the production occurring in

the entire water column per unit area. Integrated BCP allows comparisons of different

ecosystems in terms of carbon fluxes. It also allows estimation of carbon fluxes in the

Mackenzie River between Inuvik and the estuary. For example, if we consider only the East

Channel over a distance of 125 km, a mean width of 711 m (mean of 6 widths separated by

25 km on Google Earth) and a mean BCP of 3753 ± 182.1 µg C m-2 h-1 for this section of

the Mackenzie River (note that R1 and R4 are in the East Channel and that R2 and R3 are

in the Middle Channel), the BCP for one summer month would be 243 ± 11.6 Mg C.

From this value, we can estimate the bacterial respiration (BR) for the same river section

and the same period. To do this, we need a bacterial growth efficiency (BGE) value. BGE

is the amount of new bacterial biomass produced (BP) per unit of organic carbon substrate

assimilated (BP+BR). Meon and Amon (2004) have calculated a BGE of 25% in another

large arctic river, the Siberian Ob River. Using this BGE, the bacterial carbon respiration

for this section of the Mackenzie East Channel for one summer month would be of

729 ± 34.8 Mg C.

With these integrated BCP rates, we can compare the importance of bacterial metabolism

with other processes implicated in the carbon cycling of this system. Bélanger et al. (2006)

estimated the photoproduction of dissolved inorganic carbon (DIC) in the Mackenzie Shelf.

For late July, they found rates of DIC photoproduction ranging from 3.5 to 5.0 mg C m-2 d-1

(mean 4.25) for the first 25 m of the water column. For the same period of the year in

69

coastal Beaufort Sea, we measured net BCP rates of 44.2 ± 4.5 mg C m-2 d-1 at R8 (16 m)

and 68.0 ± 5.9 mg C m-2 d-1 at R9 (32 m), for a mean of 56.1 ± 5.22 mg C m-2 d-1.

Considering a BGE of 27% measured by Meon and Amon (2004) for coastal arctic

bacteria, the associated bacterial respiration rate would be of 151.6 ± 14.11 mg C m-2 d-1.

Thus, for this region and this period, DIC photoproduction was equivalent to 2.8% of the

bacterial respiration.

Metabolic balance of the system To answer to the overall question “Are northern waters a source or a sink of greenhouse

gases?”, we need to know the metabolic state of the ecosystem or the net ecosystem

production (NEP). NEP is the balance between gross primary production (GPP) and total

respiration (R) of the ecosystem. In autotrophic systems, GPP > R and autochthonous

carbon is sufficient to sustain the ecosystem metabolism. When R > GPP, the ecosystem is

said to be heterotrophic and allochthonous carbon must complement the autochthonous

carbon pool to fill the ecosystem carbon demand. Net heterotrophic aquatic communities

are net producers of CO2 to the atmosphere whereas net autotrophic communities act as net

CO2 sinks (Duarte and Prairie 2005). During the ARDEX cruise, L. Retamal (unpublished

data) measured net algal primary production (NPP) in parallel to our bacterial production

(BP) measurements. Thus, we can combine our data to estimate the metabolic balance of

the system (Table 1.5). Obviously, planktonic organisms other than bacteria (zooplankton,

flagellates and phytoplankton itself) also respire, but bacteria are generally responsible for

most aerobic respiration (Cole and Caraco 2001). Thus, since bacterial respiration (BR) is

only one component of total community respiration, when BR > NPP, these system must be

heterotrophic (Cole 1999). In addition to BR, we have considered that we know the

phytoplankton respiration, because with NPP, phytoplankton respiration is already

subtracted from GPP. We have to remember in these calculations that our bacterial

production measurements (and thus BR estimates), may be considered as underestimates of

real production rates, especially in the river and the TZ.

In the river and the TZ, depth integrated BR was 1.5 (R2) to 22 (R4) times higher than

depth integrated NPP (Table 1.5) which implies that the Mackenzie River and upper TZ

(R5d) are net heterotrophic ecosystems. In contrast, coastal Beaufort Sea station R8 showed

70

NPP 3.8 times higher than BR rates being a net autotrophic zone (Table 1.5). At R9, BR

was equivalent to 1.4 times the NPP value (Table 1.5) indicating a net heterotrophic station.

To have an independent evaluation of the metabolic balance of an ecosystem, we can use

measurements of CO2 concentrations in surface waters and the overlaying atmosphere.

Unbalanced aquatic metabolic processes will generate gaseous disequilibria with respect to

the atmosphere, which can therefore indicate the prevalence of auto- or heterotrophy

(Duarte and Prairie 2005). During the ARDEX cruise, P. Ramlal (unpublished data) did

such measurements using a continuous flow of surface water passing through a gas

equilibrator that splits the gas flow between an oxygen sensor and a LiCor GasHound

infrared CO2 analyzer. Her results were consistent with our biological data (Annex 1.9). At

the station R4 of the Mackenzie River, Ramlal found that surface waters were

supersaturated in CO2 (mean 26.8 mmol m-3; 694.2 ppm) with a CO2 concentration

equivalent to 146.0% of the atmospheric value (18.4 mmol m-3; 475.5 ppm). In the coastal

zone, at R8, dissolved CO2 concentration (21.2 mmol m-3; 358.5 ppm) was undersaturated

(89.5%) compared to the atmospheric value (23.7 mmol m-3; 400.8 ppm). At R9, water CO2

concentration (22.1 mmol m-3; 418.4 ppm) was close to the atmospheric level with a slight

supersaturation equivalent to 103.1% the atmospheric concentration (21.4 mmol m-3;

405.7 ppm).

Considering these two independently obtained data sets, we can conclude that the

Mackenzie River and upper TZ are net heterotrophic ecosystems that require an external

supply of organic carbon to sustain their metabolism. The degradation of this allochthonous

organic carbon produced CO2 in excess in the water, making them sources of CO2 to the

atmosphere. For the coastal Beaufort Sea, the interpretations were more complex. At R8,

the net autotrophy was supported by the microbial and gaseous data, meaning that

autochthonous organic carbon was sufficient to support the community metabolism of this

zone. However, at R9 which is located more offshore than R8, metabolic ecosystem

balance seemed to tend towards net heterotrophy according to both gaseous and microbial

data. One explanation for this variability may be the highly variable spatial and temporal

extent of riverine plumes. On event scales, weather patterns contribute significantly to

variations in direction and strength of plume flow and mixing (Dagg et al. 2004). Thus,

71

plume morphology is dictated by many variables. Garneau et al. (2006) explained the

presence of higher salinities near shore and lower salinities offshore in the Beaufort Sea, in

September 2002, by the morphology of the Mackenzie plume that deflected back eastwards

into the study region, and was cut off from the coast by wind-induced upwelling. This

situation probably occurred during our sampling given the diminution in salinity and rise in

temperature at R9 compared to R8.

Heterotrophy is prevalent in many aquatic ecosystems meaning that they respire more

organic carbon than they produce by autochthonous photosynthesis (Cole 1999; Cole and

Caraco 2001; Duarte and Prairie 2005). Data compilations by Duarte and Prairie (2005) for

streams, lake and rivers around the world showed that the vast majority of freshwater

ecosystems investigated (almost 90%) emit CO2 to the atmosphere. In arctic Alaska,

measurements of PCO2 in 29 aquatic ecosystems showed that in most cases (27 of 29) CO2

was released to the atmosphere (Kling et al. 1991). A study of Raymond et al. (1997) in the

Hudson River showed that, throughout the year, water PCO2 was always supersaturated

(mean 1147 µatm) relatively to the atmosphere (mean 416 µatm). Even though organic

matter in arctic rivers is considerer to be largely composed of recalcitrant soil-derived

material instead of material released from algae (Dittmar and Kattner 2003), the observed

ecosystem heterotrophy implies that at least some of this allochthonous matter is consumed

in the system. In contrast, marginal seas at high and temperate latitudes have been

identified as net annual sinks of atmospheric CO2 (Borges et al. 2005).

Implications for the future of the Arctic According to many global climate models, the circumpolar Arctic will suffer large

temperature increases in the next century (ACIA 2005). ACIA (2005) predicts decreases in

sea-ice and terrestrial snow extent during the 21st century, as well as general increases both

in precipitation minus evaporation over the marine Arctic and in river discharge to the

Arctic Ocean from the surrounding terrestrial watersheds. Also, permafrost temperatures

increased markedly during the latter half of the 20th century in the northernmost permafrost

regions of North America and Eurasia and this trend appears to be accelerating (Nelson

2003). Considering that more than half of global soil organic carbon is stored in the Arctic

Ocean basin (Dixon et al. 1994), we can expect that a large quantity of organic carbon

72

would be release by this melting. Frey and Smith (2005) showed that the -2oC mean annual

air temperature (MAAT) isotherm is a critical temperature threshold, above which

watersheds produce increasing DOC as a function of peatland abundance. A warming arctic

climate could lead to increased release of currently sequestered peat carbon through

introduction of a new source of DOC from older and deeper areas of the peat column,

caused by permafrost degradation (Frey and Smith 2005). Climate model simulations

predict a major northward advance of the -2oC MAAT isotherm by 2100 that would nearly

double the West Siberian land surface with air temperature exceeding this threshold (Frey

and Smith 2005). DOC concentrations in freshwater draining from upland catchments in

the United Kingdom have already begun to increase (Evans et al. 2005; Freeman et al.

2001). Sediment load in arctic rivers is also predicted to increase by 22% for every 2oC

warming of the averaged drainage basin temperature and of 32% if this warming is

combined with a 20% increase in runoff (Syvitski 2002). Some studies have shown that

very old organic matter can support a significant fraction of bacterial metabolism in rivers.

Highly 14C-depleted carbon of ancient terrestrial origin (1000-5000 years old) may be an

important source of labile organic matter to the Hudson River system and support the

excess respiration (Cole and Caraco 2001). Another study in the Hudson River found that

bacterial production was partly (up to 25%) supported by old (24 000 yr) allochthonous

organic matter, which was presumably derived from soil (McCallister et al. 2004).

Our study has shown that the Mackenzie River, its estuary and adjacent coastal Beaufort

Sea support a well developed microbial food web with a high percentage of heterotrophic

organisms. Heterotrophic prokaryote activity was high across the system with an important

percentage associated to particle-bound bacteria. Our evaluations of the metabolic state of

the system reveal that net heterotrophy occurs in the Mackenzie River and its transition

zone, meaning that heterotrophic processes are fueled by allochthonous organic carbon

from the watershed in addition of autochthonous organic carbon. Considering the large

amount of organic carbon stored in the Arctic Ocean catchment area and the predicted

warming of the Arctic, we can postulated that the runoff of carbon-rich waters towards the

Mackenzie River associated with permafrost melting could further increase the net

heterotrophy of this large river ecosystem, and thereby create an important positive

feedback on greenhouse gas production and warming.

73

Table 1.1 Autotrophic and heterotrophic picoplankton community abundance along the ARDEX transect.

Station Depth Autotrophic picoplankton

Heterotrophic prokaryotes

(m) (104 cells ml-1) (104 cells ml-1) R1 0 3.84 - R2 0 4.51 96.2 R2 20 5.13 91.6 R3 0 2.38 63.8 R3 26 2.77 83.3 R4 0 1.88 - R4 6 1.50 -

R5d 0 0.94 54.3 R5d 3.5 0.96 89.2 R5b 0 1.55 - R5a 0 1.16 69.4 R5a 2.5 0.35 82.1 R7 0 0.81 89.9 R7 6.5 0.11 35.7 R8 0 0.07 47.0 R8 12 0.004 69.9 R9 0 0.003 42.5 R9 15 0.002 17.3 R9 21 0.007 21.0

74

Table 1.2 Observed protist taxa at each station at the surface along ARDEX transect.

Station Group Species

R3 Cryptophyceae Cryptophyceae spp.

Chlorophyceae Ankistrodesmus spp. and/or Monoraphidium spp. cf. Chlamydomonas sp. Monoraphidium cf. minutum (Naegeli) Komarkova Tetraedron sp.

Bacillariophyceae Cyclotella spp. Diatoma sp. Synedra spp. Nitzschia spp.

Dinophyceae Unidentified autotrophic Dinophyceae

Chrysophyceae Dinobryon bavaricum Imhof cf. Pseudokephyrion sp.

Heterotrophic organisms Amoeboid sp. Bicoeca sp. Choanoflagellate sp. Unidentified heterotrophic flagellates

Autotrophic flagellates Unidentified autotrophic flagellates


Bacillariophyceae Cyclotella spp. Unidentified pennate diatoms

Chrysophyceae Dinobryon bavaricum Imhof cf. Pseudokephyrion spp.

Ciliates Ciliate sp.

Heterotrophic organisms Amoeboid sp. Aulomonas sp. Bicoeca sp. Choanoflagellate sp. Stelexomonas dichotoma Lackey Unidentified heterotrophic flagellates

R5d Cryptophyceae Cryptophyceae spp.

Chlorophyceae Ankistrodesmus spp. and/or Monoraphidium spp. Colonial coccoid Chlorophyceae Oocystis sp. Tetraedron cf. minimum (A. Braun) Hansgirg

Bacillariophyceae Cyclotella spp. Chaetoceros sp. Nitzschia spp. Unidentified centric diatoms

Dinophyceae Unidentified heterotrophic Dinophyceae

Chrysophyceae Dinobryon bavaricum Imhof cf. Synura sp.

Heterotrophic organisms cf. Telonema subtile Griessmann

R5a Cryptophyceae Cryptophyceae spp.

Chlorophyceae Ankistrodesmus spp. and/or Monoraphidium spp. Crucigenia cf. tetrapedia (Kirchner) W. and G. S. West Oocystis sp. Scenedesmus sp.

Bacillariophyceae Asterionella sp. Chaetoceros spp. Diatoma cf. elongatum Agardh Diatoma sp. Nitzschia sp. Pseudo-nitzschia sp. Unidentified centric diatoms Unidentified pennate diatoms

Dinophyceae Heterocapsa rotundata (Lohmann) Hansen

75

Raphidophyceae Raphidophyceae sp.

Ciliates Ciliate spp.


Heterotrophic organisms Ebria tripartita (Schumann) Lemmermann cf. Telonema subtile Griessmann


Chlorophyceae Flagellated Chlorophyceae (cf. Dunaliella sp.)

Bacillariophyceae Chaetoceros spp. Leptocylindrus danicus Cleve Pseudo-nitzschia spp. Thalassiosira spp.

Dinophyceae Gymnodinium spp. Gyrodinium spp. Heterocapsa rotundata (Lohmann) Hansen cf. Prorocentrum sp. Unidentified autotrophic Dinophyceae Unidentified heterotrophic Dinophyceae

Chrysophyceae Dinobryon balticum (Schütt) Lemmermann Ollicola vangoorii (Conrad) Vørs

Raphidophyceae Raphidophyceae sp.

Prymnesiophyceae Chrysochromulina spp. cf. Prymnesium sp. Coccolithophorid sp.

Prasinophyceae Pachysphaera pelagica Ostenfeld Pyramimonas spp.

Ciliates Strombidiids Ciliate spp.


Heterotrophic organisms cf. Calliacantha sp. cf. Salpingoeca sp. cf. Stephanoeca sp. Choanoflagellates spp. Heterotrophic flagellates


Chlorophyceae Flagellated Chlorophyceae

Bacillariophyceae Thalassiosira spp. Chaetoceros spp. Pseudo-nitzschia spp. Thalassionema sp.

Dinophyceae Gymnodinium spp. Gyrodinium spp. Heterocapsa rotundata (Lohmann) Hansen Ceratium sp.

Chrysophyceae Dinobryon balticum (Schütt) Lemmermann Ollicola vangoorii (Conrad) Vørs

Euglenophyceae Euglenophyceae sp.

Prymnesiophyceae Chrysochromulina spp. Coccolithophorid sp.

Prasinophyceae Pyramimonas spp.

Ciliates Strombidiids Ciliate spp.


Heterotrophic organisms Amoeboid sp. cf. Calliacantha sp. Choanoflagellates spp. cf. Telonema subtile Griessmann

76

Table 1.3 Percentage of autotrophs and heterotrophs in the protist community in terms of abundance and carbon biomass. Numbers in parenthesis are the percentages obtained after the exclusion of the unknown group.

Abundance Biomass Autotrophic Heterotrophic Unknown Autotrophic Heterotrophic Unknown (%) (%) (%) (%) (%) (%)

R3 69.9 (70.6) 29.0 (29.4) 1.4 (0) 79.4 (79.9) 20.0 (20.1) 0.6 (0) R4 64.4 (67.9) 30.5 (32.1) 5.1 (0) 48.1 (49.6) 48.8 (50.4) 3.1 (0) R5d 87.1 (92.0) 7.5 (8.0) 5.4 (0) 76.5 (77.9) 21.7 (22.1) 1.9 (0) R5a 84.4 (95.4) 4.1 (4.6) 11.6 (0) 70.4 (72.2) 27.0 (27.8) 2.6 (0) R8 82.2 (87.0) 12.3 (13.0) 5.5 (0) 72.5 (85.6) 12.2 (14.4) 15.3 (0) R9 89.6 (89.6) 10.4 (10.4) 0 (0) 89.8 (89.8) 10.2 (10.2) 0 (0)

Table 1.4 Percentages of autotrophs and heterotrophs in the microbial community (picoplankton and protists) in terms of carbon biomass. Numbers in parenthesis are the percentages obtained after the exclusion of the unknown group.

Biomass Autotrophic Heterotrophic Unknown (%) (%) (%)

R3 63.1 (63.3) 36.6 (36.7) 0.4 (0) R4 - - - R5d 58.1 (58.8) 40.6 (41.1) 1.3 (0) R5a 63.7 (65.2) 34.0 (34.8) 2.3 (0) R8 53.5 (60.2) 35.3 (39.8) 11.2 (0) R9 68.2 (68.2) 31.8 (31.8) 0 (0)

77

Table 1.5 Comparisons between bacterial metabolism (net bacterial production plus bacterial respiration) and net primary production (L. Retamal, unpublished). BGE are from Meon and Amon (2004).

Station Integration depths

BGE used

Bacterial production

Bacterial respiration

Primary production BR vs PP

(m) (%) (mg C m-2 d-1) (mg C m-2 d-1) (mg C m-2 d-1) R1 0 - 3 25 27.0 ± 1.3 81.0 ± 3.9 11.72 BR >> PP R2 0 - 21.4 25 70.9 ± 5.4 212.7 ± 16.2 142.06 BR > PP R3 0 - 30 25 110.5 ± 7.8 331.5 ± 23.4 56.95 BR >> PP R4 0 - 19 25 152.0 ± 2.9 456.0 ± 8.7 20.51 BR >> PP R5d 0 - 3.7 27 16.0 ± 1.5 43.1 ± 3.9 28.33 BR > PP R8 0 - 16 27 44.19 ± 4.5 119.48 ± 12.2 454.56 BR << PP R9 0 -32 27 67.96 ± 5.9 183.7 ± 16.0 133.80 BR > PP

78

Figure 1.1 Sampling site and stations. White dots: Mackenzie River stations. Black dots: transition zone stations. Gray dots: Beaufort Sea stations. The picture shows the 20 m and 50 m isobaths.

79

Figure 1.2 Tests for the determination of the sonication length which optimized the bacterial counts at two different stations.

Figure 1.3 Saturation curve and temporal series for the calibration of the 3H-leucine uptake measurements. The measurements have been made at the riverine station R3. The error bars represent the standard deviation for each measurements (n = 3). The dotted lines show the 3H-leucine concentration and the incubation length used during the present study.

80

Figure 1.4 Mean air temperature and precipitations that occurred during the ARDEX cruise sampling period. Data are shown for three sites. Norman Wells is located upstream of Inuvik and is shown in order to indicate the weather in the southern part of the watershed. The dashed lines show the beginning of the cruise. Data are from the National Climate Data and Information Archives of Environment Canada (Online-a).

81

Figure 1.5 Mackenzie River discharge between July 26th and August 5th 2004 at two different stations. Data are from the Water survey of Canada of Environment Canada (Online-b).

82

Figure 1.6 Water column characteristics of the stations. The dashed lines show the sampling depths. S: Salinity curve; C: Chl a curve; T: temperature curve. **Note that the Chl a fluorometer was not calibrated and that the data are given in relative units.

83

Figure 1.7 Surface water properties at each station along ARDEX transect.

84

Figure 1.8 Bottom water properties at each station along ARDEX transect.

85

Figure 1.9 Salinity and temperature data separated according to sampling zones (River, TZ and Sea) and depths (Surface, Bottom). Plots entitled SURFACE and BOTTOM compare the average value of each zone for one depth. Those entitled RIVER, TRANSITION ZONE and SEA compare the average value of each depth for one zone. The errors bars represent the standard deviations of the data. The letters show the results of the multiple comparison tests. n.s. = no significant difference. * = significant difference (P ≤ 0.05). ** = highly significant difference (P ≤ 0.01).

86

Figure 1.10 Turbidity, SPM, POM and <3 µm Chl a / total Chl a ratio data separated according to sampling zones (River, TZ and Sea) and depths (Surface, Bottom). Plots entitled SURFACE and BOTTOM compare the average value of each zone for one depth. Those entitled RIVER, TRANSITION ZONE and SEA compare the average value of each depth for one zone. The errors bars represent the standard deviations of the data. The letters show the results of the multiple comparison tests. n.s. = no significant difference. * = significant difference (P ≤ 0.05). ** = highly significant difference (P ≤ 0.01).

87

Figure 1.11 DOC, POC, PON, Total Chl a and <3 µm Chl a data separated according to sampling zones (River, TZ and Sea) and depths (Surface, Bottom). Plots entitled SURFACE and BOTTOM compare the average value of each zone for one depth. Those entitled RIVER, TRANSITION ZONE and SEA compare the average value of each depth for one zone. The errors bars represent the standard deviations of the data. The letters show the results of the multiple comparison tests. n.s. = no significant difference. * = significant difference (P ≤ 0.05). ** = highly significant difference (P ≤ 0.01).

88

Figure 1.12 Picophytoplankton abundance along the ARDEX transect between the Mackenzie River and the Arctic Ocean.

89

Figure 1.13 Bacteria and picophytoplankton data separated according to sampling zones (River, TZ and Sea) and depths (Surface, Bottom). Plots entitled SURFACE and BOTTOM compare the average value of each zone for one depth. Those entitled RIVER, TRANSITION ZONE and SEA compare the average value of each depth for one zone. The errors bars represent the standard deviations of the data. The letters show the results of the multiple comparison tests. n.s. = no significant difference. * = significant difference (P ≤ 0.05). ** = highly significant difference (P ≤ 0.01).

90

Figure 1.14 Protist community structure and abundance across the freshwater-saltwater transition zone of the Mackenzie River and Beaufort Sea. Crypt.: Cryptophyceae; Chloro.: Chlorophyceae; Bacillario.: Bacillariophyceae; Dino.: Dinophyceae; Chryso.: Chrysophyceae; Raphido.: Raphidophyceae; Eugleno.: Euglenophyceae; Prymnesio.: Prymnesiophyceae; Prasino.: Prasinophyceae; Choano.: Choanoflagellates; Flag.: Flagellates; Auto.: Autotroph or autotrophic; Hete.: Heterotroph or heterotrophic; Ident.: Identified; and Unident.: Unidentified.

91

Figure 1.15 Protist community structure and biomass across the freshwater-saltwater transition zone of the Mackenzie River and Beaufort Sea. Crypt.: Cryptophyceae; Chloro.: Chlorophyceae; Bacillario.: Bacillariophyceae; Dino.: Dinophyceae; Chryso.: Chrysophyceae; Raphido.: Raphidophyceae; Eugleno.: Euglenophyceae; Prymnesio.: Prymnesiophyceae; Prasino.: Prasinophyceae; Choano.: Choanoflagellates; Flag.: Flagellates; Auto.: Autotroph or autotrophic; Hete.: Heterotroph or heterotrophic; Ident.: Identified; and Unident.: Unidentified.

92

Figure 1.16 Protist community and total microbial community (protists and picoplankton) structure in terms of abundance and carbon biomass. Nano. : Nanoplankton. Micro. : Microplankton. Picophyto. : Picophytoplankton. Pico. : Picoplankton. * Heterotrophic picoplankton data missing.

93

Figure 1.17 Heterotrophic picoplankton production as 3H-leucine uptake rates. Top panels: Total and <3 µm 3H-leucine uptakes (pmol L-1 h-1). Bottom panels: Percentage of the total 3H-leucine uptake that is due to particulate-attached bacteria (>3 µm). The error bars represent the standard deviation for each measurement (n = 3).

94

Figure 1.18 Total 3H-leucine uptake, <3 µm 3H-leucine uptake and >3 µm / total 3H-leucine uptake ratio data separated according to sampling zones (River, TZ and Sea) and depths (Surface, Bottom). Plots entitled SURFACE and BOTTOM compare the average value of each zone for one depth. Those entitled RIVER, TRANSITION ZONE and SEA compare the average value of each depth for one zone. The errors bars represent the standard deviations of the data. The letters show the results of the multiple comparison tests. n.s. = no significant difference. * = significant difference (P ≤ 0.05). ** = highly significant difference (P ≤ 0.01).

CHAPTER 2: FACTORS CONTROLLING BACTERIAL PRODUCTION IN TWO HIGH LATITUDE RIVERS

Résumé Les rivières sont des écosystèmes importants puisqu’elles forment un lien direct dans le

transfert du carbone organique terrestre vers le milieu océanique. À l’intérieur même des

rivières, le carbone organique subit des transformations, mais l’ampleur de celles-ci dépend

de plusieurs facteurs. Dans la présente étude, nous voulions étudier les facteurs limitant la

dégradation du carbone organique par les bactéries dans le système du fleuve Mackenzie et

de la Grande Rivière de la Baleine (GRB). Nous avons également étudié l’effet de la

photochimie sur la labilité du carbone organique dissous pour les bactéries dans le système

du Mackenzie. Les résultats de nos expérimentations ont montré que le métabolisme

bactérien est limité par la disponibilité du carbone organique dans le fleuve Mackenzie et

par le phosphore dans la GRB. Nous avons constaté que la photochimie augmente la labilité

du carbone dans le fleuve Mackenzie, mais la diminue dans la mer de Beaufort.

Abstract Rivers are important conduits for the transport of organic carbon from terrestrial

ecosystems to the sea. Along the riverine path, organic carbon is transformed, but the

degree of transformation depends on many factors. This study aimed to investigate the

limiting factors for organic carbon degradation by bacteria in the Mackenzie River and

Great Whale River systems. It also aimed to study the impact of photochemical processes

on organic carbon lability to bacteria in the Mackenzie system. The results of enrichment

experiments showed that bacterial metabolism was carbon limited in the Mackenzie River

but phosphorus limited in the Great Whale River. Photochemical reactions increased the

biolability of organic carbon in the Mackenzie River but decreased biolability in the

Beaufort Sea.

96

Introduction Rivers are major freshwater ecosystems of the Arctic. They discharge annually 3299 km3 of

freshwater to the Arctic Ocean, accounting for approximately 11% of the global runoff

(Rachold et al. 2004). Dixon et al. (1994) have estimated that more than half of the global

organic carbon pool is stored in the catchment areas of the Arctic Ocean. Global circulation

models predict that global warming will be more important in high nordic latitudes (Moritz

et al. 2002). Permafrost responds to climate warming by a decline of its spatial extent

causing rapid carbon loss (Zimov et al. 2006). The organic matter concentrations of arctic

rivers are among the largest (Dittmar and Kattner 2003) and with permafrost melting, more

organic matter will potentially be released in those ecosystems. However, little is known

about the microbial ecology of arctic rivers despite the globally significant quantities of

DOC that they carry.

Heterotrophic picoplankton plays the dominant role in the degradation of organic matter

and several factors limit their efficiency in aquatic ecosystems (del Giorgio and Davis

2003). These factors can be classified as intrinsic and extrinsic factors (del Giorgio and

Davis 2003). The former includes the chemical characteristics of the DOM affecting its

availability to bacteria, such as the molecular weight distribution and the nutrient content,

which are determined by the source and the diagenetic state of the matter. Extrinsic factors

are those regulating the metabolism of bacteria, and, therefore, the utilization of organic

matter by the bacterial community. These include temperature, the availability of inorganic

and trace nutrients, trophic interactions within microbial food webs, and the phylogenetic

composition of the bacterial assemblage.

Another factor that influences carbon cycling in the aquatic environment is the

photodegradation of chromophoric dissolved organic matter (CDOM). Photochemical

transformations of CDOM by solar radiations have been revealed by the photobleaching of

CDOM and the appearance of photoproducts including dissolved inorganic carbon

(Bertilsson and Tranvik 2000; Lean 1998; Moran and Covert 2003; Moran and Zepp 1997).

The photodegration of CDOM affects its biolability and can increase or decrease its

degradation by the bacterial community. The resulting effect of photochemistry on CDOM

97

lability depends on the original characteristics of CDOM itself with fresh phytoplankton

molecules being transformed into more refractory products and old refractory humic

material being transformed into more available products (Brisco and Ziegler 2004;

Obernosterer et al. 1999; Tranvik and Bertilsson 2001).

In the previous chapter, we investigated microbial community structure and heterotrophic

picoplankton production in the Mackenzie River and Estuary (Nunavut, Canada). The

present study focused on controls of the bacterial metabolism. We aimed to evaluate the

limitation of bacterial metabolism by the availability of organic carbon in the Mackenzie

River and estuary systems and the effect of UV-photochemical reactions on its lability.

In order to compare the Mackenzie River with another high latitude region, we undertook

additional experiments in the Great Whale River and estuary (Nunavik, Canada) to

investigate the abiotic factors controlling bacterial metabolism. In the latter system, we

tested the carbon and phosphorus limitation of bacterial activity.

The Mackenzie River is the fourth largest arctic river in term of freshwater discharge in the

Arctic Ocean, with an annual export of 330 km3 (Macdonald et al. 1998). It flows 1500 km

from Great Slave Lake to the sea and its watershed is the largest in Canada, with a total

area 1.8 x 106 km2 (Macdonald et al. 1998). The Great Whale River is a subarctic river

originating from Bienville Lake and flowing across the granitic Precambrian Shield. Its

annual discharge is approximately 22 km3 (Ingram 1981).

Methods

Mackenzie River and Beaufort Sea Water samples were collected along a 300 km transect aboard the CCGS Nahidik in the

Mackenzie River and the Beaufort Sea within the framework of ARDEX (Arctic River

Delta Experiment), a multidisciplinary satellite program of CASES (Canadian Arctic Shelf

Exchange Study). The transect crossed the freshwater-saltwater transition zone (TZ) of this

system beginning at Inuvik, NWT, Canada, and ending at a station 50 km offshore in the

Beaufort Sea (see Fig. 1.1 in Chapter 1). Surface samples were collected using a clean

plastic bucket. Environmental and microbial variables were measured along this transect. A

complete description of this sampling can be found in the Methods section of Chapter 1.

98

Great Whale River and Hudson Bay In the Great Whale River and Hudson Bay, sampling was carried out between July 2nd and

5th 2005 from a small boat. Six stations were sampled at the surface (Fig. 2.1) including

three stations in the freshwater part of the Great Whale River (GRB1, 3 and 2), one at the

river mouth (EMB) and two in the riverine plume extending in the Hudson Bay (EST and

BAY). EST and BAY were separated by a salinity front. This front was clearly visible

because of the presence of foam and debris forming a line at the surface, and from evident

differences in water color. Samples were taken with a clean bucket and transferred in soft

plastic cubitainers covered with black plastic bags to prevent exposure to sunlight. Samples

were processed within 2 to 3 hours of sampling. Three stations were sampled for the

experiments: GRB1, EST and BAY.

Physical and chemical characteristics In the Great Whale River and Hudson Bay, surface temperature and salinity were measured

using a YSI 600R sonde (YSI Environmental).

For dissolved organic carbon (DOC) samples, water was filtered through 0.2 µm pore size

cellulose acetate filters (47 mm diameter) (the first 50 ml were discarded in order to prevent

potential acetate contamination from the filter) and stored at 4oC in acid washed brown

glass bottles rinsed with milli-Q water and with the sample. DOC samples were analyzed

by high combustion-direct injection in a gas analyzer. The samples were bubbled with

CO2-free nitrogen for 7 min to ensure the removal of all the dissolved inorganic carbon

(DIC). Analyses were done using a Shimadzu TOC Analyzer 5000A (detection limits of

0.05 mg l-1) at the Institut National de la recherche scientifique, Centre Eau, Terre et

Environnement (INRS-ETE, Québec City, Canada).

Soluble reactive phosphorus (SRP) samples were filtered as above for DOC, acidified with

H2SO4 (0.2% final conc.) and stored at 4oC in acid washed clear glass bottles previously

rinsed with milli-Q water and the sample. Samples were analyzed at the INRS-ETE on a

Lachat autoanalyzer (detection limit of 1 µg L-1).

Samples for suspended particulate matter (SPM) were filtered onto pre-combusted and pre-

weighed GF/F glass fiber filters (Whatman, 0.7 µm, 47 mm), folded, wrapped in aluminum

99

foil and stored at -20oC. The filters were then dry for 24 h at 60oC and re-weighed for

determination of SPM mass. The filters were subsequently combusted at 500oC for 1.5 h

and weighed to obtain the mass of the particulate inorganic matter (PIM). The particulate

organic matter (POM) was obtained by subtraction of PIM from SPM.

Particulate organic carbon and nitrogen (POC and PON) samples were filtered on GF75

glass fiber filters (Millipore, 0.7µm, 25 mm) and stored in aluminum foil at -20oC while on

the field (up to one month), and then at -80oC. POC and PON concentrations were analyzed

by high temperature oxidation using an elemental analyzer LECO CHNS-932 (INRS-ETE)

with a detection limit of 0.03 mg L-1 for carbon and of 0.005 mg L-1 for nitrogen. Filters

were acidified with HCl fumes overnight and allowed to dry at 65°C previous to analysis in

tin or silver sleeves.

Chlorophyll a (Chl a) samples were filtered onto GF75 glass fiber filters (Millipore, 0.7µm,

47 mm), stored at -20oC and then at -80oC. The filters were extracted with hot ethanol

(95 %) and Chl a concentrations were determined by fluorometry before and after

acidification using a Cary Eclipse spectrofluorometer (Varian) against a calibration curve

made with Chl a standard quantified by spectrophotometry (Cary 300 Bio U.V., Varian).

Bacteria were preserved with glutaraldehyde in acid washed clear glass bottles and stored at

4oC until further processing (within 2 months of sampling). The samples were filtered

through Nucleopore black polycarbonate membranes (25 mm, 0.22 µm) placed onto

cellulose acetate backing filters (25 mm, 0.8 µm) under low pressure. The DNA marker

DAPI (Porter and Feig 1980) (5 µg L-1 final conc.) was added when 2 ml of sample were

remaining and left to stain for 15 min. The filters were then mounted between microscope

slides and cover slips with non fluorescent immersion oil and stored at -20oC until

counting. Counting was made on a Zeiss Axioskop 2 epifluorescence microscope, under

UV light and 1000x magnification with immersion oil. A minimum of 15 field and 400

cells were counted where possible.

Bacterial production The 3H-leucine (3H-Leu) incorporation method was used as a measurement of protein

synthesis by heterotrophic picoplankton. The complete protocol used for bacterial

100

production measurements is described in Chapter 1. In the Great Whale River, 3H-Leu

specific activity was of 167 Ci mmol-1. Saturation curve and temporal series experiments

showed that 3H-Leu incorporation was linear for at least 300 min and that 10 nM below that

for saturating bacterial uptake of 3H-Leu (Annex 2.3). Thus, our bacterial production

calculations should be considered as conservative estimates of actual production rates.

Mackenzie River experimental design

Carbon limitation At stations R4, R5b and R9, three polypropylene bottles (Nalgene, 1L) were filled with

non-filtered water after being thoroughly rinsed with the sample. Two bottles received

5 µM (final conc.) of glucose and one was kept unamended to serve as a control. Bottles

were incubated for 24 h in the dark and at simulated in situ temperature in an improvised

incubator (a cooler filled with water from the sampled stations). At the end of the

incubation, fractionated bacterial production was measured in each bottle as described in

the bacterial production section (Chapter 1, Methods). The stimulation of bacterial

production was expressed as the ratio between treatment and control leucine incorporation

rates (a value of 1 corresponding to no stimulation) and these ratios were used in statistical

tests. To test the difference between the calculated ratios and 1, we used unilateral

hypothesis test according to which:

and

S being the standard deviation of the ratios, n = 2 and α = 0.05. H0 was rejected if

t0 > + t α, n-1. The t value found in tables was: t 0.05, 1 = 6.314.

Also, an ANOVA was run to test the difference between the ratios found at the three

studied stations.

101

Photodegradation effect on carbon biolability At stations R4 and R9, water sterilized by filtration through 0.2 µm Gelman PALL filters

was exposed to sunlight for 3 days in quartz bottles in shipboard experiment (1 dark and 1

light treatment, no duplicate) (Osburn, unpublished data). After sunlight exposure, 90 ml of

the exposed water were placed in 125 ml clear glass bottles and inoculated with 10 ml of

bacterial inoculum (0.8 µm filtered) coming from the same station. Bacteria were allowed

to grow for 24 h at simulated in situ temperatures and in the dark. Bacterial production was

then measured in each bottle.

Great Whale River experimental design

Carbon and phosphorus limitation At three stations (GRB1, EST and BAY), twelve polycarbonate bottles (Nalgene, 1L) were

filled with bulk water after being thoroughly rinsed with the sample. Three bottles were

kept unenriched, three received 5 µM of glucose (as the organic carbon substrate, final

conc.), three received 5 µg L-1 of phosphorus (as K2HPO4; final conc.) and 3 received both

glucose and phosphorus. Bottles were incubated in the dark at simulated in situ temperature

for 24 h. At the end of incubation, bacterial production was measured in each bottle. An

ANOVA was run to test the difference between the treatments (control, glucose,

phosphorus, glucose+phosphorus) at each station. If ANOVA detected a significant

difference between the treatments, a Fischer-LSD multiple comparison test was used to

assess which treatments were different from the others. The stimulation of bacterial

production was expressed as the ratio between treatment and control leucine incorporation

rates (a value of 1 corresponding to no stimulation).

Results

Mackenzie River and Beaufort Sea Sampling conditions and environmental characteristics during the ARDEX cruise can be

found in the Results section of Chapter 1. Table 2.1 presents the initial environmental and

biological characteristics of the stations R4, R5b and R9 before the experiments were

undertaken.

102

Great Whale River and Hudson Bay

Sampling conditions The day prior GRB1 sampling (July 1st), Kuujjuarapik received 34 mm of rain. On July 2nd,

mean air temperature was of 11oC and the weather was cloudy. July 3rd was cloudy, rainy

(8 mm of rain) and cold. On July 4th, EMB, GRB2 and GRB3 were sampled. It was a very

cold day with mean air temperature of 3oC and strong winds. The last stations were

sampled in the evening of July 5th when the wind allowed offshore sampling. It was a cold

day without rain. Climate data were obtained from the National Climate Data and

Information Archive of Environment Canada website (Online-a).

Environmental characteristics and gradients Great Whale River salinity was 0.01 psu. Salinity rose after the EMB station to reach 18.5

psu at the BAY station (Annex 2.1). The temperature was relatively constant in the river

(around 14oC) and dropped at the EST and BAY stations (Annex 2.1).

SPM concentration was variable along the transect with a mean of 15.6 ± 4.9 mg L-1

(Annex 2.1). The %POM/SPM increased towards the Hudson Bay. POC load was higher in

the freshwater zone than at the brackish stations EST and BAY (Annex 2.1). PON averaged

0.025 ± 0.007 mg L-1 and did not vary significantly along the transect (Annex 2.1). DOC

concentration was >4 mg L-1 at all stations except BAY where the concentration was at

least 1.5 mg L-1 smaller than at the other stations (Annex 2.1). The SRP concentration was

0.011 ± 0.015 mg L-1 with the highest concentration at EST station with 0.040 mg L-1

(Annex 2.1). Chl a biomass was highest in the freshwater stations and dropped of more

than 3 times at the more saline stations EST and BAY (Annex 2.1).

Bacterial abundance in the Great Whale River varied between 1.5 and 2.3 x 109 cells L-1

(Annex 2.2). The concentration decreased in saltwater zone to a mean of 0.9 x 109 cells L-1.

Bacterial 3H-Leu uptake rates was highly variable in the freshwater part of the system

ranging from 28.3 ± 1.5 to 132.9 ± 1.6 pmol L-1 h-1 (Annex 2.4). It dropped to 7.5 ± 0.4

pmol L-1 h-1 at EST station and then increased to 40.1 ± 3.4 at the BAY station located the

other side of the salinity front. The percentage contribution of particle-attached bacteria to

total 3H-Leu uptake averaged 91% in the river and 9% in the saltwater stations (Annex 2.4).

103

The initial environmental and biological characteristics of each station sampled for the

experiments are given in Table 2.2.

Response of bacterial activity to glucose addition in the Mackenzie River / Beaufort Sea ecosystems Total heterotrophic picoplankton community responded differently to glucose addition in

the river, the estuary and the sea (Fig. 2.2 and Annex 2.1), with greater stimulation in the

river than in the sea, and intermediate effects in the TZ. Statistically, there was no

difference between the stations in terms of the response level of bacterial activity to glucose

addition (Kruskal-Wallis, P = 0.067). However, this lack of significance could be

attributable to the low power of the statistical test due to the small sample size (n = 2) for

each treatment. In the river station R4, bacterial activity was almost 4 times higher in the

glucose enriched bottles than in the control (ratio = 3.9 ± 0.4, significantly different from 1,

t = 9.78, P < 0.05). This stimulation decreased in the TZ (R5b) and the bacterial production

was 2.5 ± 0.03 times higher in the glucose amended bottles than in the control (highly

significantly different from 1, t = 80.86, P < 0.01). Bacterial production responded slightly

to glucose addition in the sea, but the stimulation was not significantly different from 1

(t = 5.80, P > 0.05).

For the <3 µm community fraction, response to glucose addition was not clear (Fig. 2.2 and

Annex 2.1) and there was no significant difference between the stations. In the river,

glucose addition seemed to have induced a stimulation of bacterial activity, but the large

difference in bacterial activity between replicates prevented any clear interpretation. In the

transition zone and the sea, the stimulation was weak.

Response of bacterial activity to the exposure of DOC to sunlight in the Mackenzie River / Beaufort Sea ecosystems Although there was no replication in this experiment, there were large differences observed

between control and treatment bottles can provide insight into the role of UV in this

system. At R4, riverine DOC exposition to sunlight induced an almost 2 fold increase in

bacterial metabolism (Fig. 2.3). However, in the coastal station, the inverse response was

observed: the sunlight treatment showed almost 3 times less activity than the dark control.

104

Response of bacterial activity to glucose and phosphorus addition in the Great Whale River / Hudson Bay ecosystems Great Whale River bacterial production was limited by the availability of phosphorus. In

fact, at GRB1 the phosphorus amended treatment showed a significant 1.4-fold increase in

bacterial activity compared to the unenriched control (ANOVA, F = 53.363, P < 0.001;

LSD, P = 0.001) (Figure 2.4). At the same station, bacterial production was secondarily

limited by carbon availability as showed by the significant, 1.9-fold higher production

(ANOVA, F = 53.363, P < 0.001; LSD, P < 0.001) in the glucose + phosphorus treatment

than in the control. Glucose addition alone did not stimulate bacterial metabolism. In EST

and BAY stations, none of the added nutrients stimulated heterotrophic picoplankton

activity.

Discussion Both the Mackenzie and Great Whale Rivers occur at high latitudes where climate change

effects are likely to be pronounced in the future (Moritz et al. 2002). The large stores of

carbon in permafrost soils (Dixon et al. 1994) make it critically important to understand the

microbial processes that may mobilize it to greenhouse gases. In Chapter 1, we described

the environmental and microbial gradients occurring along the saltwater-freshwater

transition zone of the Mackenzie River and the Beaufort Sea. In the present chapter, we

used experimental approaches to study the controlling factors of bacterial metabolism.

In the Mackenzie River, we observed that glucose addition increased total bacterial

production (Fig. 2.2), suggesting that bacterial metabolism was limited by the lability of

available organic carbon. The stimulation of total bacterial production by the addition of

glucose diminished towards the coastal zone, with no significant stimulation at the station

R9 (Fig. 2.2). Free living bacterial metabolism was not significantly stimulated by glucose

addition (Fig.2.2). The exposure of DOC to sunlight induced a contrasting response of

bacterial production in the Mackenzie River and in the Beaufort Sea (Fig.2.3). In the

Mackenzie River, sunlight exposed DOC stimulated bacterial production (Fig.2.3).

However, Beaufort Sea DOC decreased bacterial metabolism after being exposed to

sunlight (Fig.2.3).

105

Bacterial production in the Great Whale River was primarily limited by phosphorus and

secondarily by carbon (Fig. 2.3). However, in the freshwater-saltwater mixing zone and in

the Hudson Bay, phosphorus and carbon additions did not stimulate bacterial metabolism

(Fig. 2.3).

Effect of carbon on bacterial metabolism Addition of glucose as a carbon source in the Ob and Yenisei Rivers and estuary, and in the

Kara Sea significantly increased bacterial production relatively to control treatments

indicating a carbon limitation of bacterial growth in the rivers and throughout the Kara Sea

(Meon and Amon 2004). In the Amazon River, bacterial respiration and production were

carbon limited, indicating that the bulk of the relatively abundant particulate and dissolved

organic matter was of limited bioavailability (Amon and Benner 1996; Benner et al. 1995).

In Raunefjorden on the western coast of Norway, bacterial production was carbon limited

and this response was consistent with low phytoplankton growth, low light conditions and

high nutrient availability occurring in November at this latitude (Flaten et al. 2003).

Our evidence of carbon limitation indicates that bacteria will be responsive to increases in

DOC concentration. Even if the majority of DOC is refractory, bacteria are ready to use the

labile fraction of this organic carbon pool. The stimulatory effect of carbon addition will

cease when another nutrient becomes limiting for bacterial metabolism (for example P or

N).

Effect of phosphorus addition on bacterial metabolism Previous work on arctic waters, specifically lakes and tundra ponds on Banks, Ellef-

Ringnes, Ellesmere and Devon Island, showed that bacterial growth was significantly

enhanced in phosphorus-amended cultures. Addition of carbon further increased BP in

lakes but not in ponds (Granéli et al. 2004). This contrasting response was due to in situ

concentrations of dissolved nutrients: total phosphorus was low in all systems, whereas

total dissolved nitrogen and organic carbon was many times higher in ponds than in lakes.

Similarly, phosphorus limitation of bacterial production in the Great Whale River indicates

that an increase in organic carbon concentration will not stimulate bacterial metabolism.

However, if phosphorus concentration increases, for example due to anthropogenic

106

activities or increased soil mineralization through climate change, the degradation of carbon

will rise and organic carbon may become the primary limiting factor for bacterial

production.

Effect of photochemistry on bacterial metabolism Photochemical reactions are of major interest in the context of climate warming because

UV-dependent processes are likely to accelerate as a result of shrinking sea ice and

decreasing ice thickness (Bélanger et al. 2006). Under an ice-free scenario in southeastern

Beaufort Sea, the photodegradation of DOC to DIC could pass from the present value of

2.8% to 6.2% of the DOC input from the Mackenzie (Bélanger et al. 2006).

In the York River estuary, photobleaching increased bacterial DOC decomposition by 27 to

200% (McCallister et al. 2005). However, exposure of surface water DOC to sunlight in the

Gulf of Mexico resulted in a 75% reduction in bacterial production (Benner and Biddanda

1998). Tranvik and Bertilsson (2001) revealed the contrasting effects of solar UV radiation

on dissolved organic sources for bacterial growth. They explain that the net effect of

radiation on DOC bioavailability can be modeled on the basis of the characteristics of DOC

and older humic material. Recently produced algal DOC is mainly transformed into

compounds of lower microbial substrate quality (Benner and Biddanda 1998; Tranvik and

Bertilsson 2001). However, old humic material is converted into products supporting

increased bacterial growth after exposition to UV (McCallister et al. 2005; Smith and

Benner 2005; Tranvik and Bertilsson 2001). Brisco and Ziegler (2004) and Obernosterer et

al. (1999) found similar conclusions by showing that refractory DOC is transformed in

more labile molecules, and that labile DOC becomes more refractory after exposure to solar

radiation. The increase in bacterial productivity following photodegradation of refractory

DOM is explained by the formation of a variety of biolabile photoproducts as low

molecular weight organic compounds, carbon gases, unidentified bleached organic matter

and nitrogen and phosphorus-rich compounds (Moran and Zepp 1997). A possible

mechanism for a reduction in bioavailability is photochemical condensation into refractory

macromolecules, for example the formation of humic substances from fatty acids (Moran

and Covert 2003).

107

The DOM in the Yenisei and Ob Rivers is potentially photoreactive as demonstrated by

photooxidation experiments made by Amon and Meon (2004). However, as in the

Mackenzie River, strong light attenuation in riverine and estuarine waters restricted

photooxidation effects to surface waters. In wetland and highly colored rivers and lakes, the

rapid attenuation of sunlight in water constrains the production of most biologically

available DOM photoproducts to the top meter of the water column (Moran and Zepp

1997). Photochemistry, even though limited to a thin surface layer, may provide readily

available organic carbon from the refractory organic matter fraction especially if the system

is physically dynamic (mixing).

Smith and Benner (2005) (St. Helena Sound, South Carolina) showed that sunlight

irradiation of terrigenous DOM significantly alters its ultimate fate (CO2 vs. bacterial

biomass) and that bacterial carbon metabolism of photoaltered DOM is coupled to an

enhanced demand for inorganic nutrients, which may considerably influence ecosystem-

scale carbon and nutrient interactions in the coastal zone. They found that photobleached

DOM decreased the BGE of bacteria growing on it, increasing the proportion of C that is

respired. In other words, photodegradation increased biolability, but a larger fraction of

what was consumed was respired instead of fixed in biomass. McCallister et al. (2005)

observed similar results in the York River estuary. Their findings suggested that the

combination of photochemical and microbial alteration of DOM may increase bacterial

demand for inorganic nutrients, alter BGE, and thereby influence the partitioning of C

between bacterial biomass and respiration.

In our study, the exposure of DOC to sunlight induced a twofold increase of bacterial

production rates in the Mackenzie River, but a factor of three decrease in the Beaufort Sea.

However, we did not measure bacterial respiration rates in the present study. Such

measurements will be crucial in the future to determine the fate of organic carbon after

photochemical modifications, and in particular the proportions that are ultimately used for

bacterial biomass production versus respiration.

Comparison of Mackenzie River and Great Whale River The fluvial part of the Mackenzie River ecosystem was more concentrated in phosphate

than the Great Whale River (Table 2.1 and 2.2). Also, DOC was higher in the Great Whale

108

River ecosystem than in the Mackenzie River ecosystem (Table 2.1 and 2.2). This could

explain the difference in response between the two ecosystems to nutrient and carbon

additions. Water chemistry of these two high latitude rivers likely differ because of

difference in watershed characteristics, the drainage basin strongly regulating the

characteristics of waters within it (Wetzel 2001).

The Mackenzie River drainage basin (1.8 x 106 km2) can be separated in four

physicogeographical regions with diverse rock and soil types (Woo and Thorne 2003). This

is reflected in the high conductivity of its surface waters. The western zone is located in the

Western Cordillera which is the major contributor of the total Mackenzie flow. The central

section is the Interior Plains with wetlands, lakes, and vegetation (from grassland prairies in

the south through boreal forest to tundra in the north). The eastern part is composed of the

Precambrian Canadian Shield. Finally, the northernmost part is the Mackenzie Delta, an

assemblage of distributaries, levees, wetlands and lakes. The Mackenzie River discharges

annually 330 km3 of freshwater in the Arctic Ocean (Macdonald and Yu 2006).

The Great Whale River watershed (4.2 x 104 km2) is smaller than the Mackenzie basin and

is underlain by the Precambrian Canadian Shield. This bedrock is of granite that is

extremely low in phosphorus content, leading to P-deficient waters. It is covered by the

boreal forest dominated by spruce (Picea spp.), tamarix (Larix laricina (Du Roi) K. Koch)

and quaking aspen (Populus tremuloides Michx.) with zones of lichens and of peat bogs

(Hudon 1994). Great Whale River discharges in the Hudson Bay about 22 km3 a-1 (Ingram

1981). Compared to the Mackenzie River, the Great Whale River flows across a watershed

that is less diversified in habitat types.

In summary, the different patterns of nutrient limitation in the Mackenzie River and the

Great Whale River will likely affect their response to a potential increase in organic carbon

runoff from the watershed. In the Mackenzie River, bacterial production should be

stimulated by this new organic carbon source and the amplitude of this increase will depend

on the lability of organic carbon inputs. For the Great Whale River, the likely response will

be different, because an increase in organic carbon concentration would not stimulate

bacterial production given the strong phosphorus limitation. If this limitation is eliminated

by an increase in phosphorus concentration, then organic carbon will become the limiting

109

factor. Potential future phosphorus input may be from anthropogenic sources such as from

human wastes, or from increased soil mineralization associated with climate change. In

brief, the response of an ecosystem to new organic carbon inputs is likely to depend of the

lability of this new carbon and of nutrients limiting its bacterial production. In the

Mackenzie River, high nutrient concentrations appear to result in strong C-limitation, while

P-limitation appears to operate in the more dilute Great Whale River waters.

110

Table 2.1 Initial environmental and biological characteristics of ARDEX stations. SRP data are from Emmerton (2006).

R4 R5b R9 Temperature (oC) 17.9 14.0 9.2 Salinity (psu) 0.13 4.5 25.2 DOC (mg L-1) 3.8 3.6 2.2 SPM (mg L-1) 59.0 38.0 20.7 POM (mg L-1) 4.3 5.2 3.3 POC (mg L-1) 1.38 0.74 ≤ 0.05 PON (mg L-1) 0.22 0.10 < 0.01 SRP (mg L-1) 0.024 0.038 0.271 Chla (µg L-1) 2.5 n.a. 0.22 Bacterial abundance (105 cells L-1) n.a. n.a. 4.3

Bacterial production ± S.D. (pM 3H-Leu h-1) 117.6 ± 9.8 24.7 ± 1.7 53.8 ± 0.9

Table 2.2 Initial environmental and biological characteristics of GRW stations.

GRB1 EST BAY Temperature (oC) 14.7 9.3 8.0 Salinity (psu) 0.01 ≈ 14 (12 to 16) 18.5 DOC (mg L-1) 4.3 4.3 2.9 SPM (mg L-1) 7.9 20.1 16.0 POM (mg L-1) 1.5 5.1 3.9 POC (mg L-1) 0.2 0.20 0.18 PON (mg L-1) 0.026 0.014 0.028 SRP (mg L-1) 0.002 0.040 0.007 Chla (µg L-1) 0.69 0.33 0.09 Bacterial abundance (109 cells L-1) 1.7 0.9 0.9

Bacterial production ± S.D. (pM 3H-Leu h-1) 28.3 ± 1.5 7.5 ± 0.4 40.1 ± 3.4

111

Figu

re 2

.1 S

ampl

ing

site

s in

the

Gre

at W

hale

Riv

er. B

etw

een

EST

and

BA

Y th

ere

was

a s

alin

ity fr

ont.

The

BA

Y s

ite w

as m

ore

salin

e th

an th

e ES

T st

atio

n.

112

Figure 2.2 Response of bacterial activity to glucose addition as the ratio of treatment 3H-leucine uptake to control in the Mackenzie River, estuary and the Beaufort Sea. The error bars represent the standard deviation of the two ratios. * Significantly >1 with P < 0.05. ** Significantly >1 with P < 0.01.

113

Figure 2.3 Bacterial 3H-leucine uptake in filtered water (0.2 µm) that had been exposed to sunlight for 3 days and subsequently inoculated with a bacterial community. Each bar represents one bottle. Error bars represent the analytical standard deviation of 3 replicates for 3H-leucine uptake measurements in each bottle.

114

Figure 2.4 Response of bacterial activity to glucose addition (C), phosphorus addition (P) and simultaneous glucose and phosphorus addition (C+P) as the ratio of treatment to control 3H-leucine uptake in the Great Whale River and estuary. The letters a, b and c are the results of ANOVA and Fisher-LSD tests run to compare the different treatments (control, glucose, phosphorus, glucose + phosphorus) at each station. n.s. means that no significant difference was detected between the 4 treatments. Different letters correspond to significant difference between the treatments (P < 0.05). In GRB1 the glucose addition treatment was not significantly different from the control.

Conclusion générale Les modèles de circulation globale prévoient une forte augmentation de la température de

surface annuelle moyenne pour l’Arctique. Ce changement de climat sera accompagné

d’une fonte du pergélisol qui amènera à son tour une augmentation du ruissellement de

carbone organique et de particules vers les fleuves. Mes résultats montrent que le fleuve

Mackenzie soutient déjà une communauté microbienne complexe qui change le long de sa

zone de transition vers la mer de Beaufort. Les écosystèmes du fleuve Mackenzie et de la

mer de Beaufort supportent une forte activité bactérienne dans laquelle les bactéries

associées aux particules jouent un grand rôle. Le fleuve Mackenzie et sa zone de transition

sont des producteurs nets de CO2 puisque la respiration bactérienne dépasse la production

primaire. Pour supporter cet état métabolique hétérotrophe du système, des apports de

carbone organique allochtone sont nécessaires. Ainsi, une partie du carbone organique

provenant du bassin versant du Mackenzie est respiré à l’intérieur même du fleuve. Nous

avons vu que le métabolisme bactérien du fleuve Mackenzie est limité par la disponibilité

du carbone, ce qui suggère qu’une fraction du carbone organique présent dans le milieu est

réfractaire à la dégradation bactérienne et que la communauté bactérienne est prête à

répondre à de nouvelles entrées de carbone organique. De plus, notre expérience de

photodégradation a montré que le carbone organique dissous du fleuve Mackenzie est rendu

plus labile à la suite d’une exposition au rayonnement solaire. Même si ces réactions sont

limitées à une mince couche superficielle dans le fleuve due à la forte atténuation de

l’éclairement dans celui-ci, la photodégradation est une entrée de carbone labile pour les

bactéries à partir du carbone organique allochtone réfractaire. Ces résultats indiquent qu’il

y aura potentiellement une forte réponse du métabolisme microbien aux augmentations de

carbone organique et de particules allochtones, ce qui augmentera l’importance des grands

fleuves arctiques comme conduits de dioxyde de carbone vers l’atmosphère.

Ce ne sont cependant pas tous les écosystèmes fluviaux qui réagiront de la sorte à une

augmentation de l’apport en carbone organique allochtone. Les fleuves dont la production

bactérienne est limitée par la disponibilité du carbone montreront une telle réponse. Il est

possible que les rivières comme la Grande Rivière de la Baleine ne réagissent pas de façon

similaire à un nouvel apport en carbone organique puisque la production bactérienne est

116

limitée primairement par le phosphore. Pour que ces rivières respirent plus de carbone

organique, elles devront premièrement recevoir plus de phosphore. Or, avec l’augmentation

actuelle de l’activité anthropogénique dans les bassins versants des rivières et des fleuves

arctiques, il est probable que le ruissellement de phosphore augmentera, en provenance, par

exemple, des eaux usées des villages et associé avec la fonte du pergélisol. Cette

augmentation de concentration de phosphore permettra une plus grande minéralisation du

carbone organique aquatique fluvial par les bactéries. Une fois cette limitation par le

phosphore levée, nos résultats suggèrent que la production bactérienne sera contrôlée par la

disponibilité du carbone organique comme c’est actuellement le cas dans le fleuve

Mackenzie.

En conclusion, dans les milieux fluviaux nordiques, le carbone organique peut être fixé

dans la biomasse microbienne, respiré par les microorganismes ou bien exporté hors du

système. Son destin dépend de plusieurs facteurs autant biologiques qu’environnementaux

(Fig. C.1). La labilité du carbone organique, qui dépend de son origine et de son état de

diagenèse, dicte la facilité avec laquelle il sera minéralisé par la communauté microbienne

du milieu. La structure du réseau alimentaire microbien et les interactions trophiques entre

les organismes le composant influencent également les taux de dégradation du carbone

organique. Finalement, nos résultats préliminaires montrent que des processus

physicochimiques, spécifiquement la photodégradation par le rayonnement solaire,

affectent considérablement la minéralisation du carbone organique par les microorganismes

en le rendant plus labile ou plus réfractaire. Ainsi, pour mieux prévoir les impacts des

changements de climat et de la fonte du pergélisol sur les environnements aquatiques

arctiques et sur les rejets de CO2 dans l’atmosphère, nous devons poursuivre nos études sur

l’écologie microbienne et la photochimie dans les rivières, les fleuves et les lacs des hautes

latitudes nordiques. Nous devrons surtout focaliser nos études sur les facteurs affectant le

métabolisme microbien aquatique et conséquemment la conversion du carbone organique

allochtone en gaz à effet de serre.

117

Figure C.1 Schéma récapitulatif illustrant le destin du carbone organique dissous (DOC) dans le milieu aquatique.

Références

ACIA. 2005. Arctic Climate Impact Assessment. Cambridge University Press. Amon, R. M. W., and R. Benner. 1996. Photochemical and microbial consumption of

dissolved organic carbon and dissolved oxygen in the Amazon River system. Geochim. Cosmochim. Acta 60: 1783-1792.

Amon, R. M. W., and B. Meon. 2004. The biogeochemistry of dissolved organic matter and nutrients in two large Arctic estuaries and potential implications for our understanding of the Arctic Ocean system. Mar. Chem. 92: 311-330.

Azam, F., D. C. Smith, G. F. Steward, and A. Hagstrom. 1993. Bacteria-organic-matter coupling and its significance for oceanic carbon cycling. Microb. Ecol. 28: 167-179.

Bélanger, S., H. Xie, N. Krotkov, P. Larouche, W. F. Vincent, and M. Babin. 2006. Photomineralization of terrigenous dissolved organic matter in Arctic coastal waters from 1979 to 2003: Interannual variability and implications of climate change. Global Biogeochem. Cycles 20: GB4005, doi:10.1029/2006gb002708.

Benner, R., and B. Biddanda. 1998. Photochemical transformations of surface and deep marine dissolved organic matter: Effects on bacterial growth. Limnol. Oceanogr. 43: 1373-1378.

Benner, R., S. Opsahl, G. ChinLeo, J. E. Richey, and B. R. Forsberg. 1995. Bacterial carbon metabolism in the Amazon River system. Limnol. Oceanogr. 40: 1262-1270.

Bérard-Therriault, L., M. Poulin, and L. Bossé. 1999. Guide d'identification du phytoplancton marin de l'estuaire et du golfe du Saint-Laurent: incluant également certains protozoaires. Conseil national de recherches du Canada (Publ. spéc. can. sci. halieut. aquat., 128).

Bertilsson, S., and L. J. Tranvik. 2000. Photochemical transformation of dissolved organic matter in lakes. Limnol. Oceanogr. 45: 753-762.

Bertrand, N., and W. F. Vincent. 1994. Structure and dynamics of photosynthetic picoplankton across the saltwater transition zone of the St.-Lawrence River. Can. J. Fish. Aquat. Sci. 51: 161-171.

Borges, A. V., B. Delille, and M. Frankignoulle. 2005. Budgeting sinks and sources of CO2 in the coastal ocean: Diversity of ecosystems counts. Geophys. Res. Lett. 32.

Bourrelly, P. 1968. Les algues d'eau douce: Initiation à la systématique, Tome II: Les algues jaunes et brunes; Chrysophycées, Phéophycées, Xanthophycées et Diatomées, N. Boubée & Cie ed.

---. 1970. Les algues d'eau douce: Initiation à la systématique, Tome III: Les Algues bleues et rouges; les Eugléniens, Péridiniens et Cryptomonadines, N. Boubée & Cie ed.

Bouvier, T. C., and P. A. del Giorgio. 2002. Compositional changes in free-living bacterial communities along a salinity gradient in two temperate estuaries. Limnol. Oceanogr. 47: 453-470.

Brisco, S., and S. Ziegler. 2004. Effects of solar radiation on the utilization of dissolved organic matter (DOM) from two headwater streams. Aquat. Microb. Ecol. 37: 197-208.

Callieri, C., and J. G. Stockner. 2002. Freshwater autotrophic picoplankton: a review. J. Limnol. 61: 1-14.

Canfield, D. E., K. Erik, and T. Bo. 2005. Heterotrophic carbon metabolism, p. 129. Adv. Mar. Biol. Academic Press.

119

Carmack, E. C., and R. W. Macdonald. 2002. Oceanography of the Canadian Shelf of the Beaufort Sea: A setting for marine Life. Arctic 55: 29-45.

Carson, M. A., J. N. Jasper, and F. M. Conly. 1998. Magnitude and sources of sediment input to the Mackenzie Delta, Northwest Territories, 1974-94. Arctic 51: 116-124.

Cauwet, G. 2002. DOM in the coastal zone, p. 579-609. In D. A. Hansell and C. A. Carlson [eds.], Biogeochemistry of Marine Dissolved Organic Matter. Academic Press.

Chrétiennot-Dinet, M.-J. 1990. Atlas du phytoplancton marin, CNRS ed. Cole, J. J. 1999. Aquatic microbiology for ecosystem scientists: new and recycled

paradigms in ecological microbiology. Ecosystems 2: 215-225. Cole, J. J., and N. F. Caraco. 2001. Carbon in catchments: connecting terrestrial carbon

losses with aquatic metabolism. Mar. Freshw. Res. 52: 101-110. Crump, B. C., and J. A. Baross. 2000. Characterization of the bacterially-active particle

fraction in the Columbia River estuary. Mar. Ecol. Prog. Ser. 206: 13-22. Crump, B. C., J. A. Baross, and C. A. Simenstad. 1998. Dominance of particle-attached

bacteria in the Columbia River estuary, USA. Aquat. Microb. Ecol. 14: 7-18. Crump, B. C., C. S. Hopkinson, M. L. Sogin, and J. E. Hobbie. 2004. Microbial

biogeography along an estuarine salinity gradient: Combined influences of bacterial growth and residence time. Appl. Environ. Microbiol. 70: 1494-1505.

Dagg, M., R. Benner, S. Lohrenz, and D. Lawrence. 2004. Transformation of dissolved and particulate materials on continental shelves influenced by large rivers: plume processes. Cont. Shelf Res. 24: 833-858.

del Giorgio, P. A., and J. Davis. 2003. Patterns in dissolved organic matter lability and consumption across aquatic ecosystems, p. 399-424. In S. E. G. Findlay and R. L. Sinsabaugh [eds.], Aquatic Ecosystems: Interactivity of Dissolved Organic Matter. Aquatic Ecology. Academic Press.

Dittmar, T., and G. Kattner. 2003. The biogeochemistry of the river and shelf ecosystem of the Arctic Ocean: a review. Mar. Chem. 83: 103-120.

Dixon, R. K., S. Brown, R. A. Houghton, A. M. Solomon, M. C. Trexler, and J. Wisniewski. 1994. Carbon pools and flux of global forest ecosystems. Science 263: 185-190.

Droppo, I. G., D. Jeffries, C. Jaskot, and S. Backus. 1998. The prevalence of freshwater flocculation in cold regions: A case study from the Mackenzie River Delta, northwest territories, Canada. Arctic 51: 155-164.

Duarte, C. M., and Y. T. Prairie. 2005. Prevalence of heterotrophy and atmospheric CO2 emissions from aquatic ecosystems. Ecosystems 8: 862-870.

Eisma, D., and G. C. Cadée. 1991. Particulate matter processes in estuaries. In E. T. Degens, S. Kempe and J. Richey [eds.], Biogeochemistry of Major World Rivers. John Wiley.

Emmerton, C. A. 2006. Downstream nutrient changes through the Mackenzie River delta and estuary, Western Canadian Arctic. Master Thesis. Simon Fraser University.

Environment Canada. Online-a. National Climate Data and Information Archive. http://www.climate.weatheroffice.ec.gc.ca/climateData/canada_f.html

---. Online-b. Water Survey of Canada. http://www.wsc.ec.gc.ca/ Evans, C. D., C. Freeman, D. T. Monteith, B. Reynolds, and N. Fenner. 2002. Climate

change - Terrestrial export of organic carbon - Reply. Nature 415: 862.

120

Evans, C. D., D. T. Monteith, and D. M. Cooper. 2005. Long-term increases in surface water dissolved organic carbon: observations, possible causes and environmental impacts. Environ. Pollut. 137: 55-71.

Findlay, D. L., and H. J. Kling. 1979. A species list and pictorial reference to the phytoplankton of central and northern Canada. Report 1503. Fisheries and Marine Service, Department of Fisheries and the Environment. 619 p.

Flaten, G. A. F., T. Castberg, T. Tanaka, and T. F. Thingstad. 2003. Interpretation of nutrient-enrichment bioassays by looking at sub-populations in a marine bacterial community. Aquat. Microb. Ecol. 33: 11-18.

Fletcher, M. 1991. The physiological activity of bacteria attached to solid surfaces. Adv. Microb. Physiol. 32: 53-85.

Freeman, C., C. D. Evans, D. T. Monteith, B. Reynolds, and N. Fenner. 2001. Export of organic carbon from peat soils. Nature 412: 785-785.

Frenette, J. J., W. F. Vincent, J. J. Dodson, and C. Lovejoy. 1995. Size-dependent variations in phytoplankton and protozoan community structure across the St-Lawrence River transition regions. Mar. Ecol. Prog. Ser. 120: 99-110.

Frey, K. E., and L. C. Smith. 2005. Amplified carbon release from vast West Siberian peatlands by 2100. Geophys. Res. Lett. 32.

Fuhrman, J. A., and F. Azam. 1980. Bacterioplankton secondary production estimates for coastal waters of British Columbia, Antarctica, and California. Appl. Environ. Microbiol. 39: 1085-1095.

Galand, P. E., C. Lovejoy, and W. F. Vincent. 2006. Remarkably diverse and contrasting Archaea communities in a large arctic river and the coastal Arctic Ocean. Aquat. Microb. Ecol. 44: 115-126.

Garneau, M.-E., W. F. Vincent, L. Alonso-Sáez, Y. Gratton, and C. Lovejoy. 2006. Prokaryotic community structure and heterotrophic production in a river-influenced coastal arctic ecosystem. Aquat. Microb. Ecol. 42: 27-40.

Granéli, W., S. Bertilsson, and A. Philibert. 2004. Phosphorus limitation of bacterial growth in high Arctic lakes and ponds. Aquat. Sci. 66: 430-439.

Gundersen, K., G. Bratbak, and M. Heldal. 1996. Factors influencing the loss of bacteria in preserved seawater samples. Mar. Ecol. Prog. Ser. 137: 305-310.

Hansell, D. A., D. Kadko, and N. R. Bates. 2004. Degradation of terrigenous dissolved organic carbon in the Western Arctic Ocean. Science 304: 858-861.

Hasle, G. R. 1978. Concentrating phytoplankton: Settling: The inverted-microscope method, p. 88-96. In A. Sournia [ed.], Phytoplankton Manual. UNESCO.

Hedges, J. I. 2002. Why dissolved organics matter? p. 1-33. In D. A. Hansell and C. A. Carlson [eds.], Biogeochemistry of Marine Dissolved Organic Matter. Academic Press.

Hill, V., G. Cota, and D. Stockwell. 2005. Spring and summer phytoplankton communities in the Chukchi and Eastern Beaufort Seas. Deep-Sea Res. Pt. II 52: 3369-3385.

Hillebrand, H., C. D. Durselen, D. Kirschtel, U. Pollingher, and T. Zohary. 1999. Biovolume calculation for pelagic and benthic microalgae. J. Phycol. 35: 403-424.

Hudon, C. 1994. Biological events during ice breakup in the Great-Whale River (Hudson-Bay). Can. J. Fish. Aquat. Sci. 51: 2467-2481.

Ingram, R. G. 1981. Characteristics of the Great Whale River plume. J. Geophys. Res. 86: 2017-2023.

121

Ittekkot, V., and R. W. P. M. Laane. 1991. Fate of riverine particulate organic matter. In E. T. Degens, S. Kempe and J. Richey [eds.], Biogeochemistry of Major World Rivers. John Wiley.

Kepner, R. L., and J. R. Pratt. 1993. Effects of sediments on estimates of bacterial density. Trans. Am. Microsc. Soc. 112: 316-330.

Kirchman, D., E. K'nees, and R. Hodson. 1985. Leucine incorporation and its potential as a measure of protein synthesis by bacteria in natural aquatic systems. Appl. Environ. Microbiol. 49: 599-607.

Kirchman, D. L. 1993. Leucine incorporation as a measure of biomass production by heterotrophic bacteria, p. 509-517. In P. F. Kemp, B. F. Sherr, E. B. Sherr and J. J. Cole [eds.], Handbook of Methods in Aquatic Microbial Ecology. Lewis.

Kling, G. W., G. W. Kipphut, and M. C. Miller. 1991. Arctic lakes and streams as gas conduits to the atmosphere - implications for tundra carbon budgets. Science 251: 298-301.

Lean, D. 1998. Attenuation of solar radiation in humic waters. In D. O. Hessen and L. J. Tranvik [eds.], Aquatic Humic Substances: Ecology and Biogeochemistry. Ecological Studies. Springer-Verlag.

Lebour, M. V. 1925. The Dinoflagellates of Northern Seas. Marine Biological Association of the United Kingdom.

Lee, S., and J. A. Fuhrman. 1987. Relationships between biovolume and biomass of naturally derived marine bacterioplankton. Appl. Environ. Microbiol. 53: 1298-1303.

Lisitsyn, A. P. 1995. The marginal filter of the ocean. Oceanology 34: 671-682. Lobbes, J. M., H. P. Fitznard, and G. Kattner. 2000. Biogeochemical characteristics of

dissolved and particulate organic matter in Russian Rivers entering the Arctic Ocean. Geochim. Cosmochim. Acta 64: 2973-2983.

Logan, B. E., and J. R. Hunt. 1987. Advantages to microbes of growth in permeable aggregates in marine systems. Limnol. Oceanogr. 32: 1034-1048.

Lovejoy, C., W. F. Vincent, J. J. Frenette, and J. J. Dodson. 1993. Microbial gradients in a turbid estuary - application of a new method for protozoan community analysis. Limnol. Oceanogr. 38: 1295-1303.

Lovejoy, C. and others. 2007. Distribution, phylogeny and growth of cold-adapted picoprasinophytes in Arctic Seas. J. Phycol. 43: 78-89.

Macdonald, R. W., S. M. Solomon, R. E. Cranston, H. E. Welch, M. B. Yunker, and C. Gobeil. 1998. A sediment and organic carbon budget for the Canadian Beaufort shelf. Mar. Geol. 144: 255-273.

Macdonald, R. W., and Y. Yu. 2006. The Mackenzie estuary of the Arctic Ocean. In P. J. Wangersky [ed.], Water Pollution: Estuaries. The Handbook of Environmental Chemistry. Springer Verlag.

MacIsaac, E. A., and J. G. Stockner. 1993. Enumeration of phototrophic picoplankton by autofluorescence microscopy, p. 187-197. In P. F. Kemp, B. F. Sherr, E. B. Sherr and J. J. Cole [eds.], Handbook of methods in aquatic microbial ecology. Lewis.

Mackenzie River Basin Board. Online. Mackenzie River Basin Board. http://www.mrbb.ca/ Mann, K. H. 2000. Ecology of Coastal Waters with Implications for Management, Second

ed. Blackwell Science. McCallister, S. L., J. E. Bauer, J. E. Cherrier, and H. W. Ducklow. 2004. Assessing sources

and ages of organic matter supporting river and estuarine bacterial production: A

122

multiple-isotope (∆14C, δ13C, and δ15N) approach. Limnol. Oceanogr. 49: 1687-1702.

McCallister, S. L., J. E. Bauer, J. Kelly, and H. W. Ducklow. 2005. Effects of sunlight on decomposition of estuarine dissolved organic C, N and P and bacterial metabolism. Aquat. Microb. Ecol. 40: 25-35.

Menden-Deuer, S., and E. J. Lessard. 2000. Carbon to volume relationships for dinoflagellates, diatoms, and other protist plankton. Limnol. Oceanogr. 45: 569-579.

Meon, B., and R. M. W. Amon. 2004. Heterotrophic bacterial activity and fluxes of dissolved free amino acids and glucose in the Arctic rivers Ob, Yenisei and the adjacent Kara Sea. Aquat. Microb. Ecol. 37: 121-135.

Moran, M. A., and J. S. Covert. 2003. Photochemically mediated linkages between dissolved organic matter and bacterioplankton, p. 246-262. In S. E. G. Findlay and R. L. Sinsabaugh [eds.], Aquatic Ecosystems: Interactivity of Dissolved Organic Matter. Aquatic Ecology. Academic Press.

Moran, M. A., and R. G. Zepp. 1997. Role of photoreactions in the formation of biologically labile compounds from dissolved organic matter. Limnol. Oceanogr. 42: 1307-1316.

Moreira-Turcq, P. F., and J. M. Martin. 1998. Characterisation of fine particles by flow cytometry in estuarine and coastal Arctic waters. J. Sea Res. 39: 217-226.

Moritz, R. E., C. M. Bitz, and E. J. Steig. 2002. Dynamics of recent climate change in the Arctic. Science 297: 1497-1502.

Mueller, D. R., W. F. Vincent, and M. O. Jeffries. 2003. Break-up of the largest Arctic ice shelf and associated loss of an epishelf lake. Geophys. Res. Lett. 30.

Nayar, S., and L. M. Chou. 2003. Relative efficiencies of different filters in retaining phytoplankton for pigment and productivity studies. Estuarine, Coastal and Shelf Science 58: 241-248.

Nelson, F. E. 2003. (Un)frozen in time. Science 299: 1673-1675. Obernosterer, I., B. Reitner, and G. J. Herndl. 1999. Contrasting effects of solar radiation

on dissolved organic matter and its bioavailability to marine bacterioplankton. Limnol. Oceanogr. 44: 1645-1654.

Painchaud, J., D. Lefaivre, J. C. Therriault, and L. Legendre. 1995. Physical processes controlling bacterial distribution and variability in the upper St-Lawrence estuary. Estuaries 18: 433-444.

Parsons, T. R., D. G. Webb, H. Dovey, R. Haigh, M. Lawrence, and G. E. Hopky. 1988. Production studies in the Mackenzie River - Beaufort Sea estuary. Polar Biol. 8: 235-239.

Parsons, T. R., D. G. Webb, B. E. Rokeby, M. Lawrence, G. E. Hopky, and D. B. Chiperzak. 1989. Autotrophic and heterotrophic production in the Mackenzie River - Beaufort Sea estuary. Polar Biol. 9: 261-266.

Peterson, B. J. and others. 2002. Increasing river discharge to the Arctic Ocean. Science 298: 2171-2173.

Plummer, D. H., N. J. P. Owens, and R. A. Herbert. 1987. Bacteria particle interactions in turbid estuarine environments. Cont. Shelf Res. 7: 1429-1433.

Porter, K. G., and Y. S. Feig. 1980. The use of DAPI for identifying and counting aquatic microflora. Limnol. Oceanogr. 25: 943-948.

123

Rachold, V. and others. 2004. Modern terrigenous organic carbon input to the Arctic Ocean. In R. Stein and R. W. Macdonald [eds.], The Organic Carbon Cycle in the Arctic Ocean. Edition Springer-Verlag.

Rae, R., and W. F. Vincent. 1998a. Effects of temperature and ultraviolet radiation on microbial foodweb structure: potential responses to global change. Freshwater Biol. 40: 747-758.

---. 1998b. Phytoplankton production in subarctic lake and river ecosystems: development of a photosynthesis-temperature-irradiance model. J. Plankton Res. 20: 1293-1312.

Raymond, P. A., N. F. Caraco, and J. J. Cole. 1997. Carbon dioxide concentration and atmospheric flux in the Hudson River. Estuaries 20: 381-390.

Retamal, L., W. F. Vincent, C. Martineau, and C. L. Osburn. 2007. Comparison of the optical properties of dissolved organic matter in two river-influenced coastal regions of the Canadian Arctic. Estuar. Coast. Shelf Sci. 72: 261-272.

Ricard, M. 1987. Atlas du phytoplancton marin, CNRS ed. Sakshaug, E. 2004. Primary and secondary production in the Arctic seas, p. 57-81. In R.

Stein and R. W. Macdonald [eds.], The Organic Carbon Cycle in the Arctic Ocean. Edition Springer-Verlag.

Saliot, A., G. Cauwet, G. Cahet, D. Mazaudier, and R. Daumas. 1996. Microbial activities in the Lena River delta and Laptev Sea. Mar. Chem. 53: 247-254.

Selje, N., and M. Simon. 2003. Composition and dynamics of particle-associated and free-living bacterial communities in the Weser estuary, Germany. Aquat. Microb. Ecol. 30: 221-237.

Serreze, M. C. and others. 2000. Observational evidence of recent change in the northern high-latitude environment. Clim. Change 46: 159-207.

Sigee, D. C. 2005. Freshwater Microbiology: Biodiversity and Dynamic Interactions of Microorganisms in the Aquatic Environment. John Wiley & Sons Ltd.

Simon, M., and F. Azam. 1989. Protein-content and protein-synthesis rates of planktonic marine-bacteria. Mar. Ecol. Prog. Ser. 51: 201-213.

Smith, D. C., and F. Azam. 1992. A simple, economical method for measuring bacterial protein synthesis rates in seawater using 3H-leucine. Marine Microbial Food Webs 6: 107-114.

Smith, E. M., and R. Benner. 2005. Photochemical transformations of riverine dissolved organic matter: effects on estuarine bacterial metabolism and nutrient demand. Aquat. Microb. Ecol. 40: 37-50.

Smith, G. M. 1950. The Freshwater Algae of the United States, Second Edition ed. McGraw-Hill book company, inc.

Sorokin, Y. I., and P. Y. Sorokin. 1996. Plankton and primary production in the Lena River estuary and in the south-eastern Laptev Sea. Estuarine, Coastal and Shelf Science 43: 399-418.

Sournia, A. 1986. Atlas du phytoplancton marin, Vols 1, 2, 3, CNRS ed. Spears, B. M., and L. F. W. Lesack. 2006. Bacterioplankton production, abundance, and

nutrient limitation among lakes of the Mackenzie Delta (western Canadian arctic). Can. J. Fish. Aquat. Sci. 63: 845-857.

Stroeve, J. C. and others. 2005. Tracking the Arctic's shrinking ice cover: Another extreme September minimum in 2004. Geophys. Res. Lett. 32.

Syvitski, J. P. M. 2002. Sediment discharge variability in Arctic rivers: implications for a warmer future. Polar Res. 21: 323-330.

124

Tomas, C. R. [ed.]. 1997. Identifying Marine Phytoplankton. Academic Press. Tranvik, L. J., and S. Bertilsson. 2001. Contrasting effects of solar radiation on dissolved

organic sources for bacterial growth. Ecol. Lett. 4: 458-463. Tranvik, L. J., and M. Jansson. 2002. Climate change - terrestrial export of organic carbon.

Nature 415: 861-862. Tsuji, T., and T. Yanagita. 1981. Improved fluorescent microscopy for measuring the

standing stock of phytoplankton including fragile components. Mar. Biol. 64: 207-211.

Turley, C. M. 1993. Direct estimates of bacterial numbers in seawater samples without incurring cell loss due to sample storage, p. 143-147. In P. F. Kemp, B. F. Sherr, E. B. Sherr and J. J. Cole [eds.], Handbook of methods in aquatic microbial ecology. Lewis.

Vetrov, A. A., and E. A. Romankevich. 2004. Carbon Cycle in the Russian Arctic Seas. Springer-Verlag.

Vincent, W. F., and J. J. Dodson. 1999. The St. Lawrence River, Canada-USA: the need for an ecosystem-level understanding of large rivers. The Japanese Journal of Limnology 60: 29-50.

Vincent, W. F., J. J. Dodson, N. Bertrand, and J. J. Frenette. 1996. Photosynthetic and bacterial production gradients in a larval fish nursery: The St Lawrence River transition zone. Mar. Ecol. Prog. Ser. 139: 227-238.

Vincent, W. F., and J. E. Hobbie. 2000. Ecology of arctic lakes and rivers, p. 197-232. In M. Nuttall and T. V. Callaghan [eds.], The Arctic: Environment, People, Policy. Harwood Academic Publishers.

Waleron, M., K. Waleron, W. F. Vincent, and A. Wilmotte. 2006. Allochthonous inputs of riverine picocyanobacteria to coastal waters in the Arctic Ocean. FEMS Microbiol. Ecol. 59: 356-365; doi: 310.1111/j.1574-6941.2006.00236.x.

Wehr, J. D., and R. G. Sheath [eds.]. 2003. Freshwater Algae of North America: Ecology and Classification, Academic Press ed. Elsevier Science.

Wells, L. E., M. Cordray, S. Bowerman, L. A. Miller, W. F. Vincent, and J. W. Deming. 2006. Archaea in particle-rich waters of the Beaufort Shelf and Franklin Bay, Canadian Arctic: Clues to an allochthonous origin? Limnol. Oceanogr. 51: 47-59.

Wetzel, R. G. 2001. Limnology: Lake and River Ecosystems. Academic Press. Wetzel, R. G., and G. E. Likens. 2000. Limnological Analysis, Third ed. Springer-Verlag. Williamson, C. E., D. P. Morris, M. L. Pace, and A. G. Olson. 1999. Dissolved organic

carbon and nutrients as regulators of lake ecosystems: Resurrection of a more integrated paradigm. Limnol. Oceanogr. 44: 795-803.

Winkler, G., J. J. Dodson, N. Bertrand, D. Thivierge, and W. F. Vincent. 2003. Trophic coupling across the St. Lawrence River estuarine transition zone. Mar. Ecol. Prog. Ser. 251: 59-73.

Woo, M. K., and R. Thorne. 2003. Streamflow in the Mackenzie Basin, Canada. Arctic 56: 328-340.

Zimmermann-Timm, H. 2002. Characteristics, dynamics and importance of aggregates in rivers - an invited review. Int. Rev. Hydrobiol. 87: 197-240.

Zimov, S. A., E. A. G. Schuur, and F. S. Chapin. 2006. Permafrost and the global carbon budget. Science 312: 1612-1613.

125

Zweifel, U. L., and A. Hagström. 1995. Total counts of marine bacteria include a large fraction of non-nucleoid-containing bacteria (ghosts). Appl. Environ. Microbiol. 61: 2180-2185.

Ann

exes

A

nnex

1.1

Env

ironm

enta

l cha

ract

eris

tics o

f the

stat

ions

. SPM

: Sus

pend

ed p

artic

ulat

e m

atte

r; PO

C: P

artic

ulat

e or

gani

c ca

rbon

; PO

N:

Parti

cula

te o

rgan

ic c

arbo

n; D

OC

: Dis

solv

ed o

rgan

ic c

arbo

n; -:

una

vaila

ble

data

. (±

rang

e; n

= 2

).

Stat

ion

Dep

th

Dat

e Te

mp.

Salin

ity

Turb

idity

SP

M

POM

PO

C

PON

D

OC

(m)

dd/m

m

(o C)

(psu

) (N

TU)

(mg

L-1)

(mg

L-1)

(mg

L-1)

(mg

L-1)

(mg

L-1)

R1

0 26

/07

18.8

1 0.

13

54.6

46

.3 ±

0.9

4.

6 ±

0.8

1.28

± 0

.023

0.

12 ±

0.0

27

4.8

R2

0 01

/08

17.6

2 0.

13

76.2

42

.3 ±

5.1

10

.2 ±

9.7

1.

03 ±

0.1

52

0.07

± 0

.032

3.

2 R

2 20

01

/08

17.6

0.

13

64.3

39

.5 ±

3.4

6.

2 ±

2.5

1.31

± 0

.076

0.

17 ±

0.0

04

3.2

R3

0 01

/08

17.1

0.

13

75.1

67

.0 ±

3.8

5.

7 ±

3.0

1.19

± 0

.151

0.

16 ±

0.0

06

3.4

R3

26

01/0

8 17

.17

0.13

75

49

.2 ±

6.9

5.

6 ±

4.5

1.69

± 0

.083

0.

12 ±

0.0

31

3.7

R4

0 27

/07

17.8

7 0.

13

76.8

59

.0 ±

2.7

4.

3 ±

2.3

1.38

± 0

.049

0.

22 ±

0.0

26

3.8

R4

6 27

/07

17.7

5 0.

13

74.5

76

.7 ±

0.7

7.

1 ±

0.3

1.90

± 0

.017

0.

30 ±

0.1

32

3.6

R5d

0

31/0

7 14

.39

1.57

97

.1

66.8

± 0

.7

5.6

± 2.

3 1.

73 ±

0.4

81

0.13

± 0

.067

3.

4 R

5d

3.5

31/0

7 13

.98

6.53

48

.6

85.5

± 6

.1

6.0

± 4.

5 1.

37 ±

0.1

48

0.11

± 0

.080

3.

1 R

5b

0 31

/07

14.0

1 4.

54

55.9

38

.0 ±

11.

7 5.

2 ±

0.2

0.74

± 0

.021

0.

10 ±

0.0

11

3.6

R5a

0

31/0

7 12

.21

8.19

29

.7

30.0

± 9

.6

4.3

± 0.

4 0.

54 ±

0.1

00

0.09

± 0

.003

4.

9 R

5a

2.5

31/0

7 11

.25

15.2

8 -

41.6

± 3

.8

5.4

± 0.

1 0.

46 ±

0.0

17

0.14

± 0

.002

2.

8 R

7 0

30/0

7 8.

73

21.8

1 8.

6 35

.6 ±

2.0

4.

2 ±

2.1

0.32

± 0

.021

<

0.03

2.

7 R

7 6.

5 30

/07

2.03

29

.83

125.

3 16

6.8

± 9.

5 16

.1 ±

0.2

1.

98 ±

0.1

68

0.17

± 0

.018

1.

8 R

8 0

30/0

7 7.

35

27.0

4 10

.4

23.6

± 0

.8

3.9

± 0.

1 0.

10 ±

0.0

23

≤ 0.

01

2.8

R8

12

30/0

7 -0

.36

31.5

5 39

.2

72.9

± 9

.5

9.9

± 1.

7 0.

50 ±

0.0

01

0.06

± 0

.015

1.

3 R

9 0

28/0

7 9.

23

25.2

2 7.

5 20

.7 ±

0.8

3.

3 ±

0.2

≤ 0.

05

< 0.

01

2.2

R9

15

28/0

7 -0

.47

31.7

2 23

.9

20.5

± 8

.0

3.4

± 0.

8 0.

25 ±

0.0

01

≤ 0.

01

2.3

R9

21

28/0

7 -1

.36

32.1

9 31

.3

26.6

± 0

.5

4.0

± 0.

3 0.

22 ±

0.0

01

≤ 0.

01

1.5

Annex 1.2 Chlorophyll a biomass measured during the ARDEX cruise. Chla: total chlorophyll a; Chla<3µm: <3 µm chlorophyll a fraction. Each value is the mean of duplicates ± SD.

Station Depth Chla Chla<3µm Chla<3µm / Chla

(m) (µg L-1) (µg L-1) (%) R1 0 2.829 ± 0.242 - - R2 0 3.249 ± 0.055 0.068 ± 0.029 2.1 R2 20 3.342 ± 0.072 0.124 ± 0.076 3.7 R3 0 3.160 ± 0.270 0.061 ± 0.020 1.9 R3 26 3.341 ± 0.165 0.167 ± 0.003 5.0 R4 0 2.498 ± 0.400 0.152 ± 0.020 6.1 R4 6 1.982 ± 0.053 0.053 ± 0.030 2.7

R5d 0 2.149 ± 0.123 0.219 ± 0.002 10.2 R5d 3.5 2.734 ± 0.059 0.090 ± 0.016 3.3 R5b 0 - - - R5a 0 3.983 ± 0.026 0.644 ± 0.083 16.2 R5a 2.5 3.241 ± 0.044 0.147 ± 0.005 4.5 R7 0 0.923 ± 0.079 0.516 ± 0.003 56.0 R7 6.5 1.636 ± 0.084 0.237 ± 0.069 14.5 R8 0 0.443 ± 0.001 0.191 ± 0.001 43.0 R8 12 6.636 ± 0.129 0.108 ± 0.007 1.6 R9 0 0.223 ± 0.006 0.152 ± 0.001 68.0 R9 15 1.433 ± 0.018 0.474 ± 0.008 33.1 R9 21 1.355 ± 0.158 0.151 ± 0.018 11.2

Annex 1.3 Heterotrophic picoplankton production as 3H-leucine uptake rates measured during the ARDEX cruise. Each value is the mean of duplicates ± SD.

Station Depth Incubation temperature

Total 3H-Leu uptake

<3 µm 3H-Leu uptake

>3 µm / total 3H-Leu uptake

(m) (oC) (pmol L-1 h-1) (pmol L-1 h-1) (%) R1 0 19 134.21 ± 6.73 3.61 ± 0.12 97.3 R2 0 15 40.87 ± 3.92 1.83 ± 0.16 95.5 R2 20 15 47.78 ± 2.92 5.05 ± 0.04 89.4 R3 0 17 36.12 ± 0.95 4.63 ± 0.32 87.2 R3 26 17 61.19 ± 5.56 9.90 ± 0.53 83.8 R4 0 19 117.59 ± 9.79 5.14 ± 0.39 95.6 R4 6 19 105.62 ± 0.63 0.73 ± 0.16 99.3

R5d 0 15 39.72 ± 2.56 11.20 ± 0.56 71.8 R5d 3.5 15 124.03 ± 15.16 11.01 ± 2.19 91.1 R5b 0 10 24.65 ± 1.76 - - R5a 0 10 106.15 ± 5.40 26.41 ± 5.16 75.1 R5a 2.5 10 131.48 ± 14.34 0.52 ± 0.03 99.6 R7 0 9 74.27 ± 7.49 62.05 ± 5.40 16.4 R7 6.5 0 46.59 ± 3.98 2.24 ± 0.49 95.2 R8 0 8 43.82 ± 6.72 23.03 ± 2.72 47.4 R8 12 0 31.92 ± 1.52 0.86 ± 0.24 97.3 R9 0 10 53.75 ± 0.91 38.00 ± 2.68 29.3 R9 15 0 23.35 ± 4.11 18.48 ± 0.35 20.9 R9 21 0 24.53 ± 1.12 3.73 ± 0.59 84.8

Annex 1.4 Surface variables mean and S.D. for each zone. River: n = 4. TZ: n = 3. Sea: n = 3. Exceptions are mentioned in the table.

Variable River TZ Sea

Salinity (psu) 0.13 ± 0.00 4.77 ± 3.32 24.69 ± 2.66

Temperature (oC) 17.85 ± 0.72 13.53 ± 1.17 8.44 ± 0.98

Bacteria (108 cells L-1) 8.00 ± 2.29 6.19 ± 1.07 5.98 ± 2.62

Picophytoplankton (106 cells L-1) 31.53 ± 12.27 12.17 ± 3.05 2.92 ± 4.48

Total 3H-Leu uptake (pmol L-1 h-1) 82.20 ± 50.95 56.84 ± 43.36 57.28 ± 15.53

< 3 µm 3H-Leu uptake (pmol L-1 h-1) 3.80 ± 1.46 18.80 ± 10.76 41.03 ± 19.69

> 3 µm / total 3H-Leu uptake (%) 93.91 ± 4.56 73.47 ± 2.34 31.06 ± 15.58

Turbidity (NTU) 70.66 ± 10.76 60.92 ± 33.95 8.83 ± 1.46

SPM (mg L-1) 53.62 ± 11.38 44.93 ± 19.39 26.59 ± 7.89

POM (mg L-1) 6.20 ± 2.74 5.03 ± 0.68 3.79 ± 0.45

DOC (mg L-1) 3.79 ± 0.69 3.97 ± 0.78 2.57 ± 0.29

POC (mg L-1) 1.22 ± 0.15 1.00 ± 0.64 0.16 ± 0.14

PON (mg L-1) 0.14 ± 0.06 0.11 ± 0.02 0.02 ± 0.01

Chla (µg L-1) 2.93 ± 0.34 3.07 ± 1.30 (n = 2) 0.53 ± 0.36

Chla<3µm (µg L-1) 0.09 ± 0.05 (n = 3)

0.43 ± 0.30 (n = 2) 0.29 ± 0.20

Chla<3µm / Chla (%) 3.36 ± 2.35 (n = 2)

13.19 ± 4.23 (n = 2) 55.66 ± 12.52

Annex 1.5 Bottom variables mean and S.D. for each zone. River: n = 3. TZ: n = 2. Sea: n = 3. Exceptions are mentioned in the table.

Variable River TZ Sea

Salinity (psu) 0.13 ± 0.00 10.90 ± 6.19 31.19 ± 1.22

Temperature (oC) 17.51 ± 0.30 12.62 ± 1.93 0.11 ± 1.74

Bacteria (108 cells L-1) 8.74 ± 0.59 8.57 ± 0.50 4.22 ± 2.51

Picophytoplankton (106 cells L-1) 31.34 ± 18.44 6.54 ± 4.25 0.40 ± 0.59

Total 3H-Leu uptake (pmol L-1 h-1) 71.53 ± 30.27 127.76 ± 5.27 34.35 ± 11.23

< 3 µm 3H-Leu uptake (pmol L-1 h-1) 5.22 ± 4.59 5.77 ± 7.42 2.28 ± 1.44

> 3 µm / total 3H-Leu uptake (%) 90.86 ± 7.84 95.36 ± 6.00 92.43 ± 6.70

Turbidity (NTU) 71.27 ± 6.05 48.63 (n = 1) 65.27 ± 52.14

SPM (mg L-1) 55.12 ± 19.31 63.55 ± 31.09 88.76 ± 71.48

POM (mg L-1) 6.28 ± 0.77 5.71 ± 0.47 9.97 ± 6.06

DOC (mg L-1) 3.49 ± 0.27 2.96 ± 0.16 1.57 ± 0.25

POC (mg L-1) 1.63 ± 0.30 0.91 ± 0.64 0.90 ± 0.94

PON (mg L-1) 0.19 ± 0.09 0.12 ± 0.01 0.08 ± 0.08

Chla (µg L-1) 2.89 ± 0.78 2.99 ± 0.36 3.21 ± 2.97

Chla<3µm (µg L-1) 0.11 ± 0.06 0.12 ± 0.04 0.17 ± 0.07

Chla<3µm / Chla (%) 3.79 ±1.18 3.92 ± 0.88 9.10 ± 6.68

Annex 1.6 Spearman correlation coefficients. Bold: P ≤ 0.01. x: no significant correlation.

In

situ

Te

mpe

ratu

re

SPM

POM

Turb

idity

DO

C

POC

PON

Tota

l Chl

a

<3 µ

m C

hl a

<3 µ

m /t

otal

C

hl a

ratio

Bac

teria

Pico

phyt

o-pl

ankt

on

Tota

l 3 H-L

eu

upta

ke

<3 µ

m 3 H

-Leu

up

take

>3 µ

m /t

otal

3 H-

Leu

upta

ke ra

tio

-0.9

76

-0.8

30

-0.6

85

-0.6

97

x

-0.8

55

-0.8

39

x X

0.85

7

x

-0.8

67

x

0.76

7

-0.9

00

Salinity

0.74

5

0.67

3

0.69

7

x

0.81

8

0.76

0

x X

-0.8

10

x

0.86

7

x

-0.8

17

0.93

3 In situ tempera-

ture

0.79

4

0.84

2

x

0.91

5

0.90

0

x X

-0.8

33

x

0.66

1

x x x SPM

0.80

6

x

0.70

9

x

0.70

0

X

-0.9

29

x

0.80

6

x

-0.8

33

x POM

x

0.90

3

0.79

6

x X

-0.7

62

x x x

-0.6

83

x Turbidity

x

0.65

0

0.71

7

X x x x x x

0.71

7

DOC

0.90

0

x X

-0.7

62

x

0.63

6

x x

0.66

7

POC

x X

-0.7

66

x x x x

0.67

8

PON

X x x

0.81

7

x x x Total Chl a

x x x x

0.76

2

x <3 µm Chl a

x

-0.9

29

x

0.92

9

-0.8

57 <3 µm /

total Chl a ratio

0.75

0

x x x Bacteria

x

-0.8

83

0.86

7 Picophyto-plankton

x x

Total 3H-Leu

uptake

-0.8

67 < 3 µm

3H-Leu uptake

Annex 1.7 Protist and cyanobacteria photomicrographs. Station numbers are indicated.

Annex 1.8 Volumetric and integrated net heterotrophic picoplankton carbon production.

Volumetric Depth integrated

Station Depth Total production <3 µm production Integration depths

Bacterial carbon production

(m) (µg C m-3 h-1) (µg C m-3 h-1) (m) (µg C m-2 h-1) R1 0 416 ± 21 11.2 ± 0.4 0 - 2.7 1123 ± 56 R2 0 127 ± 12 5.7 ± 0.5 0 - 21.4 2956 ± 225 R2 20 148 ± 9 15.7 ± 0.1 - - R3 0 112 ± 3 14 ± 1 0 - 29.6 4604 ± 325 R3 26 190 ± 17 31 ± 2 - - R4 0 365 ± 30 16 ± 1 0 - 19 6332 ±123 R4 6 327 ± 2 2.3 ± 0.5 -

R5d 0 123 ± 8 34.7 ± 1.7 0 - 3.7 665 ± 61 R5d 3.5 384 ± 47 34.1 ± 6.8 - - R5b 0 76 ± 6 - - - R5a 0 329 ± 17 82 ± 16 0 - 2.9 1064 ± 87 R5a 2.5 408 ± 44 1.6 ± 0.1 - - R7 0 230 ± 23 192 ± 17 0 - 7.5 1598 ± 158 R7 6.5 144 ± 12 7 ± 2 - - R8 0 136 ± 21 71 ± 9 0 - 16 1841 ± 188 R8 12 99 ± 5 2.7 ± 0.7 - - R9 0 167 ± 3 118 ± 8 0 - 32 2832 ± 247 R9 15 72 ± 13 57 ± 1 - - R9 21 76 ± 4 12 ± 2 - -

Annex 1.9 Percentage saturation of CO2 in water compared to atmospheric value. % saturation = [CO2]water / [CO2]air * 100%. Concentrations were in mmol m-3 corrected for solubility and temperature. (Ramlal, unpublished data).

Annex 2.1 Surface water properties of the Great Whale River stations.

Annex 2.2 Bacterial abundance in surface waters of the Great Whale River stations.

Annex 2.3 Saturation curve and temporal series for the calibration of the 3H-leucine uptake measurements. The measurements were made on water from the Great Whale River sampled on July 19th from the shore upstream of the Cris dock. The error bars represent the analytical standard deviation of the method (n = 3). The dotted lines show the 3H-leucine concentration and the incubation length used during the present study.

Annex 2.4 Heterotrophic picoplankton uptake of 3H-leucine and percentage of the activity due to particle-attached bacteria in the Great Whale River (2005). The error bars represent the analytical error (S.D. of triplicate measurements of the same water sample).

Annex 2.5 Response to glucose addition at three stations in the ARDEX cruise 2004. The graphs show measurements done in each bottle. C: Control bottle. A and B: treatment bottles that have received 5 µM of glucose. Dashed line: Level of control 3H-Leu uptake for total community. Dotted line: Level of control 3H-Leu uptake for < 3 µm fraction. The error bars represent the analytical standard deviation of the method (n = 3).

Annex 2.6 Nutrient addition experiments in the Great Whale River system. The error bars represent the SD of the treatment (n = 3). The letters show the results of the Fisher-LSD multiple comparison tests. Different letters mean significant differences (P < 0.05).

Documents

PRODUCTION BACTÉRIENNE ET STRUCTURE DU RÉSEAU … · Mme Patricia Ramlal (Freshwater Institute, Department of Fisheries and Oceans, Winnipeg, Canada) ... Merci aussi à M. Julian