AQUATIC MICROBIAL ECOLOGYAquat Microb Ecol

Vol. 75: 187–205, 2015doi: 10.3354/ame01758

Published online July 6

INTRODUCTION

Riverine ecosystems comprise a variety of ecologi-cally important habitats and support high levels oforganism diversity (Arthington et al. 2006, Dudgeonet al. 2006). Underpinning the function, stability andproductivity of flowing fresh water (or lotic) eco -systems are the microorganisms that form the base

of the food web (Stahl et al. 2006, Blum & Mills2012). Microbes including bacteria, phytoplanktonand microzooplankton are important drivers of loticchemical and energy cycles and are abundant anddynamic members of the lotic biota (Havens et al.2000)

Lotic bacterial communities are biogeographicallydiverse and are typically characterised by local habi-

© Inter-Research 2015 · www.int-res.com*Corresponding author: richard.carney@student.uts.edu.au

River bacterioplankton community responses to a high inflow event

R. L. Carney1,*, S. M. Mitrovic2, 3, T. Jeffries1, D. Westhorpe3, N. Curlevski1, J. R. Seymour1

1Plant Functional Biology and Climate Change Cluster (C3), University of Technology, Sydney, PO Box 123, Broadway, New South Wales 2007, Australia

2Centre for Environmental Sustainability (CENS), University of Technology, Sydney, PO Box 123, Broadway, New South Wales 2007, Australia

3Department of Primary Industries, Office of Water, PO Box 3720, Sydney, New South Wales 2024, Australia

ABSTRACT: Microbes drive chemical cycling and productivity within river ecosystems, but theirinfluence may shift when intense allochthonous inputs accompany high freshwater inflow (flood)events. Investigating how floods influence microbial processes is fundamentally important for ourunderstanding of river ecology, but is generally overlooked. We analysed bacterioplankton com-munity composition (BCC) and abundance over 4 mo following an enormous flood event in theHunter River, Australia, that resulted in a major fish kill. Concentrations of dissolved organic car-bon (DOC) and inorganic nutrients (N and P) were up to 3 times higher during the flood eventcompared to prior and subsequent months. Bacterial cell abundances were up to 10 times higherat impacted sites during the flood event. Using Automated Ribosomal Intergenic Spacer Analysiswe found significant shifts in BCC between the flood impacted month and subsequent months(p < 0.05). Distance linear modelling indicated that DOC and dissolved N and P correlated moststrongly with BCC patterns during the high inflow, whereas community dynamics correlated moststrongly with nitrogen oxides and ammonium during the river’s recovery phase. 16S rRNA ampli-con pyrosequencing revealed that common soil-associated and facultative anaerobic genera ofProteobacteria were most dominant during the flood period, suggesting that a proportion of thebacterial community observed during this event were potentially inactive soil microbes trans-ported into the river via terrestrial runoff. During the recovery period, Cyanobacteria and fresh-water-associated genera of Actinobacteria and Proteobacteria became dominant in 16S rRNApyrosequencing profiles. These observations indicate that allochthonous nutrients delivered viafloods can significantly stimulate bacterial growth, underpinning substrate-controlled successionof bacterial communities and ultimately shaping the ecology within river ecosystems.

KEY WORDS: Bacterioplankton · ARISA · Lotic ecology · Network analyses

Resale or republication not permitted without written consent of the publisher

This authors' personal copy may not be publicly or systematically copied or distributed, or posted on the Open Web, except with written permission of the copyright holder(s). It may be distributed to interested individuals on request.

Aquat Microb Ecol 75: 187–205, 2015

tat-specific endemism, yet are more phylogeneticallysimilar to other freshwater habitats than marine andsoil groups (Zwart et al. 2002). Longitudinal gradi-ents in microbial community composition and func-tional groups are common in rivers and estuaries, asa consequence of changing hydrology and salinityconcentrations (Crump et al. 2004, Maranger et al.2005). Similarly, lotic bacterial community assem-blages can shift seasonally, with patterns in bacterialcomposition typically driven by algal (auto chtho -nous) carbon during low precipitation seasons and byallochthonous carbon during wetter seasons (Al -meida et al. 2005, Pinhassi et al. 2006, Hitchcock &Mitrovic 2013). Compared to marine environments,bacterial abundance, productivity and respirationrates can be significantly higher in lotic systems (DelGiorgio & Cole 1998, Del Giorgio & Williams 2005)due to higher nutrient inputs from terrestrial inputs(Cole & Caraco 2001, Pollard & Ducklow 2011), andas a consequence riverine ecosystems are often netheterotrophic (Hadwen et al. 2010). Our knowledgeof the ecology of lotic microbes is much less devel-oped than our understanding of marine and lakemicrobial communities. In particular, there is cur-rently limited knowledge of how microbial communi-ties respond to the constantly shifting environmentalconditions that are characteristic of river ecosystems(Curtis & Sloan 2004, Schultz et al. 2013).

Physical and chemical conditions within lotic sys-tems can shift slowly, such as between seasons, orvery abruptly, particularly following precipitationevents that lead to high inflows and flooding (West-horpe & Mitrovic 2012). As the foundation of loticfood webs, microbes are among the first organisms toexperience and respond to changes in the dynamicphysicochemical environment of rivers and theirestuaries (Kirchman et al. 2004, Wear et al. 2013).Understanding how the abundance, composition andfunction of microbial assemblages change over dif-ferent temporal scales is fundamental to our under-standing of riverine ecology (Blum & Mills 2012).

Medium to large river inflows, where precipitationevents lead to a subsequent increase in flow volumeand velocity, are a pivotal regulator of physical habi-tat and the primary transporter of nutrients in loticsystems (Arthington & Pusey 2003). They couple abi-otic components of terrestrial environments (suchas flood plains) with biotic processes in aquatic sys-tems (Junk et al. 1989, Tockner et al. 2000, Sieczko& Peduzzi 2014). The frequency, variability, volumeand quality of river inflows ultimately determinethe ecological character of rivers and their associ-ated estuaries (Poff & Zimmerman 2010). Notably, in

many parts of the world, major rivers and estuariesare subject to flow regulation, and receive reducedflow volumes that differ greatly from historical pat-terns (Vörösmarty et al. 2010). In determining theecological significance of fresh water inflows, anunderstanding of natural biological responses tophysical changes at the base of lotic food webs isimperative, but currently lacking.

In many aquatic and marine environments, hetero-trophic bacterioplankton abundance and activity istightly regulated by the availability of DOC obtainedprimarily from autochthonous sources including phy -toplankton extracellular exudates and metazoanwastes (Azam et al. 1983, Wilson & Devlin 2013).However, rivers and estuaries periodically receivelarge inputs of particulate organic carbon (POC) andDOC from terrestrial (allochthonous) sources duringseasonal inflows and floods (Farjalla et al. 2009), andthis may temporarily uncouple the reliance of bacte-rioplankton on autochthonously derived C (Almeidaet al. 2005, Ameryk et al. 2005). Such a shift may sig-nificantly alter chemical cycling and trophic dynam-ics within lotic ecosystems over a variety of temporalscales, as has been observed in lake environments(Kritzberg et al. 2004).

When large seasonal floods (that are induced byheavy precipitation events) breach river embank-ments, floodwaters inundate catchment and flood-plains. Organic material from flood-impacted landwashes into the river as the water levels recede. Car-bon and nutrients then leach from the inundatedarea, often turning the water a dark tea colour,termed ‘black water’. In river systems where con -siderable land modification has taken place (e.g.cleared pastures), organic carbon and inorganicnutrients accumulate in high concentrations in theflood waters, which contributes strongly to microbialactivity and decomposition processes (Carvalho et al.2003, Almeida et al. 2005, Farjalla et al. 2009). This,in combination with increased turbidity levels thatoften accompany inflows, can shift the base of thelotic system to a more heterotrophic state (Westhorpeet al. 2010), and in some cases leads to the formationof hypoxic or anoxic zones (Paerl et al. 1998, Zhang etal. 2010). In extreme cases, these low O2 black waterscan cause mass mortality in the food web and candetrimentally impact fishery stocks and river ecolog-ical health (Salles et al. 2006). The extent of theseevents will be influenced by changes in compositionand activity of microbial populations that followinflow events.

In marine and lake ecosystems, pulses of organicand inorganic nutrients, derived from a variety of

188A

utho

r cop

y

Carney et al.: Bacterioplankton response to a river inflow

biotic (e.g. algal blooms) and abiotic (e.g. nutrientupwelling) events can lead to substrate-controlledsuccession of bacterial populations (Fawcett & Ward2011, Teeling et al. 2012). It is probable that alloch-thonous resource pulses related to inflow eventswithin rivers will have a similar effect on lotic micro-bial communities (Wear et al. 2013). Resource vari-ability could then select for copiotrophic microbeswhich normally persist in low abundances but arecapable of very efficient resource assimilation duringnutrient pulses. However, to date very few studieshave observed these processes in lotic systems.

Previous studies have performed nutrient manipu-lation experiments using microcosms to investigateheterotrophic bacterial growth response to simulatedinflows, floods and their associated nutrient inputs(Hitchcock et al. 2010, Hitchcock & Mitrovic 2013,Mitrovic et al. 2014). These studies have indicatedthat heterotrophic bacterial growth in rivers andestuaries is often limited or co-limited by availableDOC, N and P during periods of low river discharge(Jansson et al. 2006, Hadwen et al. 2010, Hitchcocket al. 2010, Hitchcock & Mitrovic 2013), but very littleis known about how these events influence the com-position and diversity of bacterial assemblages, orthe implications for lotic chemical cycling and trophicdynamics. This knowledge is of growing importancebecause there is a global pattern of high flow eventsbeing returned to many major coastal rivers andestuaries as a result of flow management (Dudgeon2010). Here we examined the compositional dynam-ics of a lotic microbial community during and after amajor black water event, with the objective of deter-mining to what extent ephemeral inputs of allochtho-nous organic material alter the microbial ecology ofriver ecosystems.

MATERIALS AND METHODS

Study site

Sampling was conducted in the HunterRiver estuary system, located on the mid-eastern coastline of New South Wales(NSW), Australia (63.34° S, 37.26° E to63.76° S, 37.91° E). Hunter River dischargeto its estuary is regulated by dams andweirs, and 2 of the Hunter River system’smain tributaries — the Williams and Pater-son rivers — are also regulated. The HunterRiver estuary was chosen as a suitablesample site for this study as it is represen-

tative of many regulated estuaries of the temperateeast Australian coast, and has recently had new flowregulation rules assigned to it (NSWDPI 2003).

Sampling regime

Sampling was conducted at high tide at 7 sites,across the tidal zone of the Hunter River estuary, aswell as within its 2 main tributaries (Paterson Riverand Williams River). Sites 1, 3, 4, 6 and 7 werelocated in the Hunter River, Site 2 was located withinthe Paterson River and Site 5 was located in theWilliams River (Fig. 1, Table 1). Biological samplingbegan in March 2013, in conjunction with a monthlywater quality sampling regime which had begun in2012. The first bacterial sampling occasion con-ducted in March 2013 coincided with an intenseinflow where depletion in dissolved oxygen (DO)resulted in a mass fish kill. Following this black waterevent, sampling continued on a monthly basis untilJune 2013.

Water sampling protocols

Hydrographic data were obtained from gaugingstations operated by the NSW Office of Water throug -hout the Hunter River estuary and its catchment.Flow discharge, expressed as mean ml d−1, was re -corded every 15 min over each 24 h period and con-verted to mean daily flow by averaging flow dis-charge across 24 h intervals. In situ water qualitymeasurements including temperature, pH, conduc-tivity and DO were taken at all sites from 20 cmbelow the water surface, using a calibrated Hydrolab

189

Location Site Distance from Salinityestuary mouth (km) Mar Apr May Jun

Hunter River 1 49.1

Aquat Microb Ecol 75: 187–205, 2015

Surveyor (Hydrolab) and MS5 Sonde probe (Hydro-lab). Turbidity measurements were taken using aHach turbidity Meter (Hach Company).

Surface samples for dissolved organic carbon(DOC) were collected in triplicate and immediatelyfiltered through 0.45 µm cellulose acetate filters (Sar-torius Stedim Biotech) into pre-combusted 200 mlglass bottles, refrigerated and then acidified with 2Nhydrochloric acid before analysis on a ShimadzuTOC-VCSH analyser using the High TemperatureCombustion Method (Eaton et al. 1995, Westhorpe& Mitrovic 2012). Nutrient samples (50 ml) were separately analysed for total nitrogen (TN), dissolvedorganic nitrogen (DON), oxidised nitrogen (NOx),ammoniacal nitrogen (NH4), total phosphorus (TP)and filtered reactive phosphorus (FRP). Nutrientsamples were filtered through 0.45 µm filters, andstored in triple-rinsed 50 ml PET bottles and refriger-ated at 4°C. Analysis was conducted using a seg-mented flow analyser (OI Analytical Model FS3100)in accordance with standard methods (Eaton et al.1995).

Chlorophyll a (chl a) concentrations were deter-mined by filtering 200 ml of surface water onto

0.75 µm pore-size glass microfiber filters and wereanalysed using the methods of Eaton & Franson(2005).

Bacterial cell enumeration

Triplicate 1 ml surface samples were fixed withglutaraldehyde (2% final concentration) and snapfrozen in liquid nitrogen. Samples were stored at−80°C prior to flow cytometric (FCM) analysis. Inpreparation for FCM, samples were quickly thawedin hot water and stained with SYBR Green I nucleicacid stain (1:10000 final dilution; Molecular Probes)(Marie et al. 1997, Gasol & Del Giorgio 2000, Sey-mour et al. 2005). Fluorescent reference beads (1 µmdiameter, yellow/green; Molecular Probes) wereadded to each sample immediately prior to analysesat a final concentration of 105 ml−1. Samples wereanalysed using an LSRII flow cytometer (BectonDickinson) and bacterial populations were discrimi-nated according to cell side scatter (SSC) and SYBRGreen fluorescence. Data were analysed using Win-MDI v. 2.9 software (WinMDI; Joseph Trotter, Salk

190

Fig. 1. Sampling sites within the tidal pool of the Hunter River and its 2 main tributaries, the Paterson River and the Williams River, New South Wales, Australia

Aut

hor c

opy

Carney et al.: Bacterioplankton response to a river inflow

Institute for Biological Studies). Cell concentrationsare reported in bacterial cells ml−1.

DNA extraction

Triplicate 1 l surface water samples were collectedin rinsed plastic bottles, and filtered using a peri-staltic pump onto 5 µm and 0.2 µm membrane filtersarranged in an in-line filtration set up. The 5 µmpore size membrane filter was employed to capturebacteria attached to suspended particles (Nair &Simidu 1987, Gram et al. 2010, Tang et al. 2011,Rösel & Grossart 2012), allowing discriminationfrom the free-living bacterial community capturedon 0.2 µm membrane filters. In the absence of a uni-versally agreed filter size standard appropriate fordiscriminating attached and free-living bacterialcommunities, we conservatively selected a 5 µmmembrane pore size which will allow for thecapture of medium to large particles colonised bybacteria (Azam & Hodson 1977, Acinas et al. 1999,Riemann & Winding 2001). Filters were immediatelysnap-frozen in liquid nitrogen, and stored at −80°Cuntil DNA extraction was performed. Extraction ofmicrobial genomic DNA was performed using abead beating and chemical lysis kit (MoBio Power-Water) according to the manufacturer’s directions.Extracted DNA was quantified using a Qubit Fluo-rometer (Life Technologies).

Bacterial community fingerprinting

To investigate bacterial community diversity weused Automated Ribosomal Intergenic Spacer Analy-ses (ARISA) (Brown et al. 2005). PCR was performedusing 16S and 23S Internal Transcribed Spacer (ITS)specific primers. These included the 6-carboxyfluo -rescein (FAM)-labelled forward primer 1392f (5’-GYACAC ACC GCC CGT-3’) and the reverse primer 23Sr(5’-GGG TTB CCC CAT TCR G-3’) (Sigma-Aldrich).Each 20 µl PCR reaction mixture consisted of 2 µl ofextracted DNA, 10 µM of each primer, 10 µl ofGoTaq® Colorless Master Mix, (2×: GoTaq® DNAPolymerase supplied in 2× Colorless GoTaq® Reac-tion Buffer at pH 8.5), 400 µM dATP, 400 µM dGTP,400 µM dCTP, 400 µM dTTP and 3 mM MgCl2. Thisreaction mixture was subjected to a PCR cycle of94°C for 5 min, and then 35 cycles (94°C for 40 s,56°C for 40 s, 72°C for 90 s), followed by 72°C for5 min (Brown et al. 2005). Fragments were sizedusing a 3730xl DNA Analyzer (Applied Biosystems)

at the Australian Genome Research Facility (AGRF),using the LIZ1200 internal size.

ARISA profiles were analysed using Peakscannersoftware (v. 1, Applied Biosystems). Fragments with arelative fluorescence intensity (RFI) 5 µm) and free-liv-ing (>0.2 µm,

Aquat Microb Ecol 75: 187–205, 2015

Statistical methods

Linear correlations between environmental vari-ables and bacterial cell counts were measured byapplying Pearson’s Correlation Coefficient to data.Determination of variability between flood affectedand post-flood months, sites and bacterial filter sizeclasses was achieved by applying a suite of permu -tational multivariate analysis of variance (PERM-ANOVA) statistical methods in PRIMER (Primer-E) toenvironmental and biological (ARISA and pyrose-quencing) data. To determine whether environmen-tal conditions and bacterial community composition(BCC) at river sites varied significantly between ‘dur-ing flood’ and ‘post flood’ periods, PERMANOVAwas used to compare the variance in the means ofenvironmental and biological data from the floodimpact month of March 2013, to the post floodmonths of April, May and June 2013 (Anderson2001). An assumption of PERMANOVA is that dataare equally dispersed (Anderson et al. 2007). Un -equal dispersion in environmental variable data wastested using draftsman plots. Where unequal disper-sion of data was detected, it was resolved by LOG+1transformation, therefore reducing the skewness ofthe data while maintaining proportionality and satis-fying the assumptions of PERMANOVA.

PERMANOVA with PRIMER + PERMANOVA soft-ware v. 6 (Anderson et al. 2007) was used to de -termine significant dissimilarity within biological(ARISA) and environmental data, between the floodimpacted month and subsequent months. Data werethen graphically represented using non-parametricmulti-dimensional scaling (nMDS) plots. Distancebased linear modelling (DistLM), using a stepwiseprocedure for adjusted R2, was then used to select theenvironmental variables most likely to explain pat-terns in the biological data. This was graphically rep-resented by a distance-based redundancy analysis(dbRDA) plot.

A detailed analysis of relationships between envi-ronmental variables and specific bacterial taxonomicunits was also made using network analysis (Steeleet al. 2011). Statistical associations between variableswere determined using the maximal information co-efficient (Reshef et al. 2011) by selecting the 50 mostdominant bacterial genera identified from 28 sam-ples across March and April, using the 16S rRNAamplicon pyrosequencing data and visualized inCytoscape (Fuhrman & Steele 2008). All environ-mental variable and bacterial taxa interactions thatwere insignificant (p > 0.05) or of correlative strengthbelow 0.8 were excluded. The resultant network was

created using an edge-weighted force-directed lay-out which allowed visualisation of the correlationstrength between individual bacterial genera and themeasured environmental variables (Zhou et al. 2010,Steele et al. 2011). Relationships between variableshave been represented by edges, in our network pos-itive associations are represented by black lines andred lines indicate negative associations. The lengthof the edge relates the strength of correlation be -tween variables, (e.g. shorter edges relate to strongerassociations between variables). Variables (includingbacterial groups and water chemistry measurements)have been represented by nodes, which have beencolour-coded according to the bacterial phyla orenvironmental variable they belong to, and the sizeof nodes relates to their relative abundance.

RESULTS

Physico-chemical dynamics

We used environmental variable data (including allphysical and nutrient measurements) collected dur-ing sampling occasions from January 30 until June12, 2013, which encompassed the physicochemicalconditions within the Hunter River estuary under alow flow state (January), during a major inflow event(March), and during the 3 mo recovery phase there-after (April to June).

During the high flow period in March, flow rates atsome sites increased massively from baseline levelsof below 300 ml d−1 to in excess of 90000 ml d−1.Since flow rate recordings began in this region in1968, on average there have only been 1 to 2 eventsof this magnitude occurring each decade (http://realtimedata.water.nsw.gov.au/water). Strong linearcorrelations were identified between mean river dis-charge rates and water turbidity (r = 0.66) and DOC(r = 0.84), with DOC concentrations peaking inMarch during the high inflow event (Fig. 2). Flow dis-charge also correlated strongly with FRP (r = 0.79)and DON (r = 0.77), which were both also elevated inMarch (Fig. 2). Conversely, NOx concentrations werelowest in March at Hunter River Sites 3, 4, 6 and 7and Site 2 on the Paterson River (Fig. 2). Dissolvedoxygen (DO) concentrations fell to anoxic levels dur-ing the March inflow event with the exception of Site5, on the Williams River, which was relatively un -affected by the flood. During this period DO droppedto

Carney et al.: Bacterioplankton response to a river inflow

inflow levels of between 5 and 10 mg l−1 across allsites by the next sampling effort in April. NH4 con-centrations were highest at all sites except on theWilliams River (Site 5), in April.

Using the entire suite of water chemistry and nutri-ent parameters measured, significant variation wasobserved in physical conditions, between months(df3,27, perm998; pseudo-F = 10.98, p < 0.001) and sites(df6,27, perm996; pseudo-F = 2.05, p < 0.005), indicativeof a highly dynamic aquatic environment. Althoughsome parameters were dispersed homogenously,(PERMDISP p > 0.05), the strength of significance(p < 0.01) in variation between months and sites(Fig. 2) suggests that the major inflow event in Marchwas a strong driver of spatiotemporalheterogeneity in the physical conditionof the Hunter River ecosystem.

Chl a concentrations

Chl a varied considerably betweenmonths and sites, typically rangingbetween 2 and 40 µg l−1, with 2 excep-tions where phytoplankton blooms wereevident: Site 5 in March (83.3 µg l−1),and Site 4 in June (94.3 ± 0.88 µg l−1)(see Appendix 1). Chl a concentrationswere generally lower at impacted sitesduring the flood event in March andhigher during the recovery phase inApril through to June. Chl a concentra-tion exhibited weak, negative linear cor-relations with discharge (r = −0.26) andFRP (r = −0.31), and was positively cor-related with DO (r = 0.47) and NOx (r =0.20).

Bacterial abundance

Bacterial cell concentrations werehighest at all sites in March during theinflow event, with the exception of Site 1(Hunter River) and Site 5 (WilliamsRiver). Across the entire time-series,bacterial cell abundance was highest atSite 2 on the Paterson River in March(2.79 × 108 ± 1.12 × 107 cells ml−1) andlowest at Site 7 on the Hunter River inJune (7.57 × 106 ± 6.26 × 104 cells ml−1).Comparison of bacterial cell abun-dances during the flood period in March

to subsequent months using PERMANOVA revealedsignificant (p = 0.001) and strong (df1,27, perm997;pseudo-F = 31.504) temporal dissimilarity betweenthe flood impacted month and subsequent months,which is consistent with bacterial cell concentrationsbeing up to an order of magnitude higher in Marchcompared to the following months.

Bacterial community fingerprinting

BCC, determined using ARISA, was distinctly dif-ferent in March during the flood event compared tothe recovery period over the 3 subsequent months.

193

0

5

10

15

20

1 3 4 6 7 2 5

DO

C (m

g l-1

)

a

0

0.05

0.1

0.15

0.2

1 3 4 6 7 2 5

FRP

(mg

l -1)

b

c

0

0.1

0.20.3

0.4

0.5

0.60.70.8

0.9

1 3 4 6 7 2 5

DO

N (m

g l-1

)

Hunter River sites WilliamsRiver

FebruaryMarchAprilMay June

PatersonRiver

Fig. 2. Nutrient concentrations sampled at the Hunter River (in upstreamto downstream sequence, Sites 1, 3, 4, 6 and 7, left to right) and tributarysites (Paterson River, Site 2; Williams River, Site 5) from February to June2013. (a) Dissolved organic carbon (DOC), (b) filtered reactive phosphorus(FRP), and (c) dissolved organic nitrogen (DON). Error bars represent SE

Aut

hor c

opy

Aquat Microb Ecol 75: 187–205, 2015

PERMANOVA assumptions of normality were met(PERMDISP p < 0.05) and the patterns in ARISA datarevealed a significant (p = 0.001) month and siteinteraction (df18,78, perm999; pseudo-F = 2.76). BCCshifted very strongly between the flood-impactedmonth (March) and subsequent recovery months(df1,78, perm997; pseudo-F = 2.18, p = 0.001), sup-ported by tight clustering of months and separation

of March sites using non-metric MDS (Fig. 3). DOC,FRP and DON, which were at elevated concentra-tions during the high inflow period of March (Fig. 2),correlated most strongly with community-scale shiftsin bacterial community assemblage observed in theARISA data (Fig. 4). Diminished turbidity levels andhigher NOx concentrations were also associated withtemporal shifts in BCC (Fig. 4).

194

Fig. 3. Multi-dimensional scal-ing (MDS) plot of square roottransformed ARISA data forbacterial community composi-tion from March to June 2013.Plots are colour-coded by monthand labelled by site. March,coded in red, is tightly clus-tered and dissimilar from all

other months

Fig. 4. Distance-based Redundancy Analysis Plot usingsquare root transformed ARISA data, showing Marchbacterial communities, coded in red, strongly influencedby dissolved organic carbon (DOC) and filtered reactivephosphorus (FRP), and subsequent months influencedby low turbidity, dissolved organic nitrogen (DON) and

oxidized nitrogen (NOx)

Aut

hor c

opy

Carney et al.: Bacterioplankton response to a river inflow

Bacterial phylogenetic composition

16S rRNA amplicon pyrosequencing revealed sig-nificant shifts in the bacterial community betweenthe flood-impacted conditions in March to the recov-ery period in April (Fig. 5). In March, the Proteo -bacteria represented over 75% of bacterial se -quences across all sampling sites (Fig. 5a), and thisgroup was comprised predominantly of β-Proteo -bacteria, which made up 69% of proteobacterialsequences, followed by the ε-Proteobacteria (12%)and γ-Proteobacteria (9%) (Fig. 5c). On the otherhand, during April, Proteobacteria only comprised25% of the total bacterial community (Fig. 5b),pr marily as a consequence of strong increases inthe relative abundance of Cyanobacteria (27%) andActinobacteria (26%) (Fig. 5b). Furthermore, withinApril there was a shift in the composition of the Pro-teobacteria, with a decrease in the β-Proteobacteria(43%) relative to the γ-Proteobacteria (30%) and α-Proteobacteria (24%) (Fig. 5d).

In terms of spatial patterns, β-Proteobacteriasequences were the most abundant proteobacter-ial group across all Hunter River and tributarysites (Fig. 6). In the Hunter River, the abundanceof γ-Proteobacteria increased longitudinally inApril from the upper Hunter River sites (Sites 1and 3) to down river sites (Sites 4 and 6) (Fig. 6c),while α-Proteobacteria were most abundant atSite 7 closest to the estuarine mouth, where theycomprised half the proteobacterial abundance(Fig. 6d). The composition of Proteobacteria inPaterson and Williams rivers differed in March,whereby the ε-Proteobacteria comprised over30% of the community at Site 2 in the PatersonRiver, but was virtually absent at Site 5 in theWilliams River (Fig. 6b).

Proteobacterial profiles in the Hunter River andits tributaries also differed between free-livingand particle-attached bacterial size classes. Mostnotably, the abundance of ε-Proteobacteria wasgreatest in the free-living community in the Pater-son River in March and also at Hunter River sitesdownstream of the Paterson River (Fig. 6b). Therelative importance of γ-Proteobacteria wasgreatest in the particle-attached communities inApril where they comprised over half of thesequences across all Hunter River and tributarysites, with Site 7 (closest to the estuary mouth)being the only exception (Fig. 6c). At this time,the α-Proteobacteria comprised over 50% of thetotal proteobacterial abundance in both filter sizeclasses at Site 7 (Fig. 6c,d).

At a finer taxonomic resolution, the dominantsequences matching Proteobacteria in both filter sizeclasses in March belonged to the family Comamon-adaceae, and the genera Limnohabitans, C39, Acine-tobacter and Sulforospirillum (Fig. 7a,b). By April,the large size class was dominated by Cyanobacteria,including Planktothrix and Synechococcus (Fig. 7c),and the smaller size class was strongly representedby ACK-M1 within the Actinobacteria phylum,which replaced the Proteobacteria as the dominantmembers of the aquatic microbial community(Fig. 7d).

Network analysis revealed that the frequency ofsequences matching Proteobacteria in March wasassociated with the elevated nutrient concentrationsduring the inflow event (Fig. 8). DOC concentrationwas strongly and positively correlated with se -quences matching Dechloromonas, Acinetobacter,

195

Proteobacteria Actinobacteria Cyanobacteria Bacteroidetes Verrucomicrobia

Chloroflexi Acidobacteria Planktomycetes Firmicutes Other

Betaproteobacteria Epsilonproteobacteria Gammaproteobacteria

Alphaproteobacteria Deltaproteobacteria

a b

c d

Fig. 5. Comparison between the (a,c) flood-affected period inMarch and (b,d) recovery period in April. Bacterial communitycomposition data were produced by combining and averagingthe sequence data from the 5 and 0.2 µm filter size classes. Total bacterial community composition at the phylum level in (a)March and (b) April, and Proteobacteria at the class level in

(c) March and (d) April

Aut

hor c

opy

Aquat Microb Ecol 75: 187–205, 2015

KD1-23 and Limnohabitans (Fig. 8). Similarly,sequence matches to other Proteobacteria includingDechloromonas, Hydrogenophaga and C39 werestrongly and positively correlated with concentra-tions of DON (Fig. 8). Several sequences matchinggenera of Bacteroidetes, although in relatively lesserfrequency to Proteobacteria, shared close associationwith flood-related organic nutrients, such as DOC,DON as well as turbidity. These include Bacter oi -dales, Paludibacter, Flavobacterium and sequencesmatching the family Flavobacteriaceae. Chl a con-centration and the abundance of sequences match-ing Cyanobacteria (Planktothricoides and Synechoc-cus) were negatively correlated with environmentalparameters characteristic of the flood event in March,particularly turbidity and DON, while positively cor-related with NOx and NH4, respectively, which werein greater concentration during the recovery monthsof April to June. Taken in its entirety, the network inFig. 8 illustrates the contrast between bacterial taxo-nomic groups and their nutrient associations. Generaof Proteobacteria and some Bacteroidetes were mostcommonly and strongly associated with the elevatedorganic nutrients that characterised the river con -dition during the flood in March, and chl a con cen -

trations and Cyanobacteria were negatively associ-ated with environmental conditions typical of flood conditions.

DISCUSSION

Recent studies within Eastern Australian coastalriver and estuarine freshwater inflows have measuredpatterns in bacterial abundance and biomass, andhave revealed a common positive correlationbetween the import of allochthonous DOC and inor-ganic nutrients and these parameters (Hitchcock etal. 2010, Hitchcock & Mitrovic 2013, 2014). We haveexpanded upon these observations by identifyingshifts in BCC and taxonomic diversity in response toan inflow event and have related these changes to theinput of allochthonous C and inorganic nutrients. Ourdata revealed that shifts in community compositionwere strongly correlated with specific water chem-istry parameters including DOC, DON and FRP, andthat specific phylogenetic groups of bacteria werepossibly transported from terrestrial or sediment sub-strate into the river or promoted or suppressed by thephysicochemical conditions experienced during the

196

1 3 4 6 7 PR WR

δ-Proteobacteria

ε-Proteobacteria

α-Proteobacteria

γ-Proteobacteria

β-Proteobacteria

1 3 4 6 7 PR WR

1 3 4 6 7 PR WR

1 3 4 6 7 PR WR

March attached bacteria March free-living bacteria

April attached bacteria April free-living bacteria

b

c

a

d

Hunter River sites Hunter River sites

Fig. 6. Comparison of Proteobacteria community composition and attached and free-living communities between March andApril 2013 using 454 pyrosequencing data. Proteobacteria composition of attached bacteria in (a) March and (c) April, and Pro-teobacteria composition of free-living bacteria in (b) March and (d) April. PR: Paterson River (Site 2); WR: Williams River (Site 5)

Aut

hor c

opy

Carney et al.: Bacterioplankton response to a river inflow

flood, while others emerged as dominant communitymembers under the post-flood conditions.

High inflows following precipitation events trans-port allochthonous material into river and estuarinesystems, shifting the dominant carbon and nutrientsources for bacteria from autochthonous-derived(phytoplankton exudates and zooplankton detritus)material, to labile allochthonous DOC (Sinsabaugh &Findlay 2003, Webster & Harris 2004, Farjalla et al.2009, Petrone et al. 2009, Westhorpe & Mitrovic2013). Evidence from nutrient addition experiments

suggests that changes in resource composition andphysical conditions can temporarily uncouple bacter-ial dependence on autochthonous sources of carbon(Kritzberg et al. 2004, Pinhassi et al. 2006, Hitchcocket al. 2010, Hitchcock & Mitrovic 2013). This is poten-tially reflected in our data, where we observed in -creased bacterial abundance, and community-scaleand taxon-specific shifts within the bacterial commu-nity, which correlated with shifts in the physical con-dition of the Hunter and Paterson rivers during andfollowing a high inflow event.

197

1 3 4 6 7 1 3 4 6 7

Comamonadaceae Dechloromonas C39 Acinetobacter LimnohabitansRhodoferax Sulfurospirillum Rhodocyclaceae ACK-M1 HydrogenophagaProcabacteriaceae Flavobacterium Paludibacter Rhodobacteraceae GeothrixArcobacter Flavobacteriaceae Novosphingobium Oxalobacteraceae RamlibacterOther

March attached bacteria

PR PRWR WR*

** *

*

*

*

Unidentified OTU belonging to order or family

1 3 4 6 7 1 3 4 6 7

ACK-M1 Planktothrix Synechococcus Stramenopiles ComamonadaceaeCryptophyta Chitinophagaceae Rhodobacteraceae Actinomycetales MethylophilaceaeSphingobacteriales Opitutus Cerasicoccaceae Rhodocyclaceae FluviicolaBurkholderiales OPB56 SL56 Planktothricoides FlavobacteriaceaeOther

PR PRWR WR

*

*

* * *

* *

* *

*

Unidentified OTU belonging to family

a b

c d

March free-living bacteria

April attached bacteria April free-living bacteria

Fig. 7. Bacterial community operational taxonomic units (OTUs) clustered at 97% similarity and identified down to genus levelwhere possible, using 454 pyrosequencing data. OTU diversity of attached communities in (a) March and (c) April, and OTU

diversity of free-living communities in (b) March and (d) April. PR: Paterson River (Site 2); WR: Williams River (Site 5)

Aut

hor c

opy

Aquat Microb Ecol 75: 187–205, 2015

Changes in physical and chemical conditionsduring inflow events

We found that shifts in water chemistry at sites inthe Hunter River and its tributary (the Paterson River)were strongly influenced by the March inflow andblack water event. A significant increase in organicnutrients (DOC and DON) and FRP was observedin the Hunter and Paterson rivers. Up to 8-foldincreases in DOC concentrations have been ob -served during flood conditions in river and estuarinesystems (Westhorpe & Mitrovic 2012), and weobserved DOC concentrations in the Hunter River toincrease from less than 5 mg l−1 to above 15 mg l−1 atSites 3, 4, 6 and 7.

In contrast to all other sites, Site 5 on the WilliamsRiver received a much reduced inflow, and conse-quently did not experience the increases in nutrientsor shift in DO that were observed at the other sites.Nutrient loads and composition vary between differ-ent catchments and floodplains due to hydrologicand geographic variability (Dalzell et al. 2007), andimpoundments impair the transport of sediments andparticulates (Baldwin et al. 2010), which may explainwhy the Williams River did not receive the samenutrient loads as the Hunter and Paterson rivers. Thelack of inflow effects at Site 5 are likely a conse-quence of the presence of a weir and the discreteproperties of the Williams River watershed. Due tothe stability of chemical conditions at Site 5, this site

effectively acted as a negative control when consid-ering the shifts in microbial communities observed inthis study.

Changes in bacterial abundance associated withinflow events

Both DOC and inorganic nutrients limit microbialgrowth in rivers and estuaries during low basal flows(Pinhassi et al. 2006, Hitchcock & Mitrovic 2013). Theincrease in C, N and P concentrations (DOC, TP, FRP,TN and DON) during the high inflow period coin-cided and possibly led to a significant increase inbacterial abundance, as has been observed in sev-eral micro-mesocosm nutrient addition experiments(Jansson et al. 2006, Pinhassi et al. 2006, Hitchcock etal. 2010, Hitchcock & Mitrovic 2013). Bacterial abun-dance at impacted sites during March was up to anorder of magnitude higher (at >107 cells ml−1) com-pared to the non-impacted Site 5 or subsequentmonths (April to June).

Carbon and nutrient concentrations returned tolevels that resembled pre-flood conditions by thenext sampling occasion in April (1 mo post inflowevent). During this period, bacterial concentrationsalso declined substantially. These patterns reflect thecapacity of riverine bacteria to rapidly utilise alloch-thonous pulses of dissolved organic matter duringflood conditions.

198

Fig. 8. Edge-weighted force-directed net-work analysis generated using environ-mental variables and bacterial genus data,showing the associations between bacterialgenera from March and April communitiesand environmental variables. Environmen-tal variables and bacterial genera (nodes)are connected by lines (edges) that re -present their correlations (black = positive,red = negative). The distance between nodescorresponds with the strength of their asso-ciation. Closer nodes are strongly corre-lated while those further apart represent

weaker correlation

Aut

hor c

opy

Carney et al.: Bacterioplankton response to a river inflow

Changes in bacterial community compositionduring inflow events

Analyses of our ARISA data revealed that thechanges in bacterial abundance between inflow andpost inflow periods were accompanied by a signifi-cant shift in BCC. The dissimilarity in bacterial com-munity assemblage was far greater between Marchand April, compared to dissimilarity between sub -sequent months, indicating that the inflow event inMarch profoundly altered bacterial communitiesbeyond normal month-to-month variations. Distancelinear modelling revealed that shifts in the compo -sition of bacterial communities coincided with in -creased concentrations of organic and inorganicnutrients. This is suggestive of a community changerelated to growth responses to allochthonous nutrientinputs by copio trophic bacteria (Palijan et al. 2008).However, there is also the possibility that the linksbetween imported nutrients may reflect a coinciden-tal link between water chemistry and allochtho-nously introduced microbes advected from adjacentfloodplains, riparian zones and sediments (Miletto etal. 2008).

Notably, in addition to driving temporal variabilityin the composition of microbial communities, theMarch flood event also influenced the degree of spatial structure in microbial assemblages across thestudy region. Spatial homogeneity of bacterial com-munities in the Hunter and Paterson rivers was highest during the inflow event, and communitiesbecame more spatially heterogeneous over the 3 sub-sequent post inflow event months. This pattern indi-cates that the flood event acted to remove niche par-titioning of microbial communities by homogenizingphysical and chemical conditions across the HunterRiver ecosystem, a process that may have further,previously unconsidered, ecological consequences.In addition, the advection of microbes from riparianand adjacent terrestrial sources may have con-tributed further to the apparent spatial similarity ofBCC during the flood period. We suggest that themassive input of freshwater from the upper-rivercatchment and adjacent terrestrial zones during theflood temporarily reduced the influence of localisedfactors on water chemistry and biological composi-tion of the Hunter River, which during low inflowperiods would typically underpin spatial heterogene-ity. In addition, the impacts of high inflow rates likelydiminished tidal influence and led to reduced salinityconcentrations in the estuarine portion at Sites 6and 7 (from brackish to fresh), resulting in the morespatially homogenous physical and chemical envi-

ronment, and bacterial distributions. However byApril, the re-establishment of longitudinal physico-chemical gradients — ascending from upper river todown river sites (as seen in Fig. 2), likely led to thedevelopment of a longitudinal biological gradientresulting in a more spatially heterogeneous estuarinebacterioplankton community. This is indicated by thewider distribution of June data points in Fig. 3 asthe impacts driven by the flood event dissipatedthroughout the recovery months, which is perhapsmore typical of low inflow periods.

Although less pronounced in the ARISA profiles,16S rRNA data indicated that the bacterial commu-nity at Site 5 in the Williams River differed to flood-impacted sites during March. Specifically, the ε- Proteobacteria that were present at flood-impactedsites (Sulfurospirillum in particular) were virtuallyabsent at the Williams River site in March, which hada greater proportion of freshwater bacteria comparedto the flood-impacted tributary. This suggests that,consistent with the smaller shifts in physicochemicalconditions at this site, the bacterial community at Site5 was less influenced by the flood event than theother tested sites. Notably, the differences betweenSite 5 and the other sites were much more evident in454 pyrosequencing data than the ARISA data, andare likely a result of the greater sensitivity of 454pyrosequencing relative to ARISA.

Patterns in the 16S rRNA data were indicative of ashift in the microbial assemblage from a heterotroph-dominated community during the flood event, to acommunity with a higher proportion of autotrophicmicrobes in the recovery period. In April the relativeproportion of sequences matching cyanobacteria washigher than during the flood event in March. A likelyexplanation for this pattern is the very high turbiditylevels that occurred during the flood period, whichwill result in limited light availability to phototrophicmicrobes. The reduction in cyanobacteria sequencesin March was mirrored by a decrease in total chloro-phyll. In addition to the input of organic carbon dur-ing the flood, the turbid conditions experienced inMarch will have further assisted heterotrophic bacte-ria to outcompete phototrophs for common limitinginorganic nutrients (Almeida et al. 2005, Ameryk etal. 2005). This pattern was supported by the networkanalysis which revealed that the high levels ofchemo-lithotrophic and chemoorganotrophic organ-isms, including Sulfurospirillum, ACK-M1, Dechloro -monas and Acinetobacter, were strongly positivelycorrelated with turbidity levels.

An additional driver of bacterial community pat-terns which likely influenced BCC in March was ter-

199A

utho

r cop

y

Aquat Microb Ecol 75: 187–205, 2015

restrial run-off following the heavy rainfall, trans-porting soil and soil-associated microbes into theriver water column. Sequences matching soil- andsediment-associated, chemo-lithotrophic specialistsulphate reducers and sulphur-oxidising bacteria in -cluding Acinetobacter (Newton et al. 2011), Dechlo -romonas and Sulfurospirillum (Kelly et al. 2005) com-prised a major portion of the most abundant bacterialgenera in the water column in March, across allflood-affected sites for both filter size classes. Net-work analyses indicated that these groups of bacteriawere positively correlated with flood conditions,including increased concentrations of DOC, DONand turbidity levels. The decreases in DO, and linksto bacterial abundance and specific OTUs observedhere, could be related to either the increased abun-dance and activity of lotic microbes occurring as aconsequence of organic and inorganic nutrientinputs, or the rise in bacterial biomass associatedwith allochthonous inputs of terrestrial bacteria intothe system.

Endemic freshwater bacteria also responded to theflood conditions in March. Sequences matching freshwater chemo-organotrophic bacteria including Lim-nohabitans, ACK-M1 and C39 were abundant in the0.2 µm filter size class (representing the small cellsized and free-living bacteria) in March, which isconsistent with previous observations that theseorganisms are capable of efficient assimilation oforganic carbon and inorganic nutrients facilitatingrapid growth in response to terrestrial water intru-sions (Hahn et al. 2010, Lin et al. 2012).

In April, as the physical condition of the Hunterand Paterson river systems returned to a pre-floodstate (i.e. it transitioned from an anoxic, turbid andhighly eutrophic system during the flood to an aero-bic, low turbidity and decreased nutrient state), sub-stantial shifts in the composition of bacterial com -munities occurred. The relative importance of theProteobacteria decreased significantly from March toApril and this shift in total abundance was accompa-nied by a significant change in the composition ofthis group. The patterns observed here are consistentwith previous studies that found Proteobacteria to benumerically dominant in river and lake habitatswhen bacterial communities are supported by alloch-thonous nutrient inputs, and to be less abundantwhen autochthonous nutrients are most important(Tang et al. 2009, Schultz et al. 2013). Most notably inMarch, β-Proteobacteria dominated both filter classcommunities. This group comprised cosmopolitan,freshwater bacteria belonging to the Comamonoda -ceae family and the genus Limnohabitans (Crump &

Hobbie 2005), in addition to the Dechloromonasgenus, a group previously identified as soil-dwellingbacteria (García-Armisen et al. 2014). ε-Proteobacte-ria were the second most abundant Proteobacteria inthe free-living Hunter and Paterson river sites inMarch, consisting exclusively of bacteria belongingto the Sulfurospirillum genus, which has previouslybeen observed to thrive in micro-aerophyllic, hydro-carbon-contaminated conditions (Rossi et al. 2012).We suggest that the Proteobacteria compositionobserved in March is indicative of both a numericalincrease in endemic freshwater bacteria as a result ofallochthonous organic nutrient inputs, and the trans-port of bacteria from soil and sediment environmentsinto the river. During April, while the β-Proteobacte-ria was still abundant, a substantial increase in theproportion of γ-Proteobacteria (30%) and α-Proteo -bacteria groups was observed. The increase in γ-Proteobacteria was not represented by a singledominant group but accumulatively by several lowabundance groups. Conversely, α-Proteobacteriabelonging to the Rhodobacteraceae family was moreabundant in Hunter River sites closest to the estuarymouth. These observations are consistent with typi-cal planktonic microbial communities widely re -ported in marine and brackish estuarine environ-ments (Liao et al. 2007, Pujalte et al. 2014). Wesuggest that the change in community structure andincrease in spatial heterogeneity of bacterial assem-blages in April were influenced by reduced fresh-water inflow, allowing the reestablishment of a longi-tudinal salinity gradient across the tidal pool.

We employed 2 filter size classes to discriminatebacterial sub-communities between large and at -tached cells, which were retrieved on the 5 µm filter,and small and free-living cells, collected on the0.2 µm filter. The large filamentous cyanobacteriaPlanktothrix, and small but abundant single-celledcyanobacteria Synechococcus were among the mostabundant organisms collected on the large filter sizeduring April. The occurrence of the large filamentouscells (up to 8 µm in diameter) (Entwisle et al. 1997) inthis filter size class is as anticipated, but the occur-rence of Synechococcus in this filter size, particularlyin the estuarine Sites 6 and 7 during April, is likelyexplained by colonisation of suspended particles(Crespo et al. 2013, Thiele et al. 2014). Planktothrixand Synechococcus are obligate photoautotrophs,and their distribution and abundances across river,estuary and coastal environments are regulated bythe availability of inorganic nutrients (Partensky etal. 1999, Halstvedt et al. 2007). The post-flood condi-tions of the Hunter and Paterson rivers were con-

200A

utho

r cop

y

Carney et al.: Bacterioplankton response to a river inflow

ducive for Plankothrix and Synechococcus growth,whereby these organisms presumably exploited theelevated light levels and concentrations of reminer-alised nutrients during this period. Network analysisrevealed that phototrophic genera were negativelyassociated with the conditions during the floodperiod and likely promoted by elevated inorganicnitrogen post flood in April. In addition there was asubstantial increase in the relative importance of theactinobacterial group ACK-M1 which has been ob -served in freshwater environments to have a strongchemotactic response to ammonium (Dennis et al.2013) and be in high abundance during eutrophicconditions (Van der Gucht et al. 2005). Networkanalysis indicated that reduced turbidity levelswere also a strong contributor to ACK-M1 relativeabundance.

The shifts in BCC observed here were driven byfluctuating physical and chemical conditions asso -ciated with an inflow event and are to some extentindicative of substrate-controlled succession of micro -bial communities within this ecosystem. While suc-cessional patterns among bacterial communitiesassociated with algal blooms and nutrient upwellingshave been widely reported in marine and lake envi-ronments (Kritzberg et al. 2004, Rösel & Grossart2012, Teeling et al. 2012), investigations of largefreshwater inflow events that describe successionalchanges in bacterial communities in lotic systems,particularly of the magnitude we observed in theHunter River, are rarer. However, extrinsic, climate-related factors have previously been identified as thebest predictors of seasonal microbial community pat-terns, whereby in a 3 year temporal study of 2 inde-pendent temperate rivers temperature and flow rateswere found to be key drivers of bacterial diversity(Crump & Hobbie 2005). However, the rapid, short-term physical changes caused by large-scale floodslikely override these seasonal patterns (Junk et al.1989, Tockner et al. 2000) and lead to temporaryrestructuring of bacterial communities. A possibledriver of rapid compositional shifts is the input ofallochthonous substrate into rivers during floodevents, which subsequently lead to rapid succes-sional changes in microbial assemblages (Mašín etal. 2003, Chung et al. 2014, Santos et al. 2014).

Substrate-controlled succession occurs in bacterio-plankton communities when new pulses of resourcesbecome available and are subsequently depleted bydifferent members of the microbial assemblages. Fol-lowing pulse inputs of chemical resources, specialistcopiotrophic microbes that may normally persist atlow abundances rapidly exploit substrate pulses and

increase in abundance (Lauro et al. 2009, Patel etal. 2014). The subsequent production of secondarymetabolites and mineralised nutrients can then fuelthe rise of other groups of microbes. Alternatively, asthe original resource pulse is exhausted or physico-chemical conditions change the microbial commu-nity shifts further. We propose that major river inflowevents provide allochthonous fuel which catalysessuccessional processes within river ecosystems.

Taking our data into consideration, we suggest thata typical lotic microbial response to a major inflow orflood event would be as follows: The massive intro-duction of allochthonous DOC and inorganic nutri-ents during a flood event is followed by an increasein the activity and abundance of heterotrophic bacte-rial populations as has been reported in several pre-vious investigations of bacterial productivity incoastal aquatic environments (Pinhassi et al. 2006,Hitchcock et al. 2010, Pollard & Ducklow 2011). Theconcomitant increase in turbidity levels during theflood will restrict photoautotrophic activity, leadingto a reduction in cyanobacterial and phytoplanktonbiomass (Henley et al. 2000). This shift in communitymetabolism towards hetero trophy will support theestablishment of anoxic conditions and the subse-quent increase in anaerobic bacterial communities(Walsh et al. 2004, Diaz & Rosenberg 2008, Zhang etal. 2010). In addition, the flood event will likely leadto the transport of non-endemic bacterial groups andanaerobic bacteria including sulphate-reducing bac-teria (Miletto et al. 2008) from surrounding terrestrialsources within the catchment, into the river environ-ment (Edwards & Meyer 1986, Crump & Hobbie2005). Following the flood event, during the river’srecovery phase, the exhaustion of allochthonousDOC and the gradual reduction in turbidity levelswill allow for the re-establishment of photoauto-trophic microbes, which will likely also benefit fromthe pool of remineralised nutrients resulting from theheterotrophic bloom during the flood. These resultsreveal the importance of river inflow events andfloods as a significant driver of physico-chemical con-ditions which subsequently influence spatiotemporalpatterns in the composition of bacterial communities,ultimately shaping the ecology of river ecosystems.

CONCLUSIONS

Inflow events are likely a fundamental driver ofmicrobial community dynamics within river ecosys-tems, but to date have been largely neglected. Theresults of this study illustrate that high inflows and

201A

utho

r cop

y

Aquat Microb Ecol 75: 187–205, 2015

black water events can lead to profound changes inwater chemistry and nutrient profiles, which stimu-late sharp increases in bacterial cell abundance andphylogenetic shifts in BCC. The changes in BCCobserved here are indicative of a substrate-con-trolled succession process, providing evidence thatmedium to large freshwater inflow events are animportant driver of spatiotemporal heterogeneity inthe bacterial communities of rivers. This researchprovides fundamental insights needed to understandhow transportation of allochthonous nutrients viainflow events influence the bacterial communities atthe base of riverine food webs, and what implicationsthis may have for lotic ecological health.

Acknowledgements. We acknowledge the contributions ofthe New South Wales Office of Water and thank them forproviding nutrient data and the use of their boat for samplecollection. This project was funded by the University ofTechnology, Sydney - School of Environment (SoE), and thePlant Functionality and Climate Change Cluster (C3). Thisproject was funded by the Australian Research CouncilGrants DP120102764 and FT130100218 to J.R.S.

LITERATURE CITED

Acinas SG, Antón J, Rodríguez-Valera F (1999) Diversity offree-living and attached bacteria in offshore westernMediterranean waters as depicted by analysis of genesencoding 16S rRNA. Appl Environ Microbiol 65: 514−522

Almeida MA, Cunha MA, Alcantara F (2005) Relationship ofbacterioplankton production with primary productionand respiration in a shallow estuarine system (Ria deAveiro, NW Portugal). Microbiol Res 160: 315−328

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990)Basic local alignment search tool. J Mol Biol 215: 403−410

Ameryk A, Podgórska B, Witek Z (2005) The dependencebetween bacterial production and environmental condi-tions in the Gulf of Gdańsk. Oceanologia 47: 27−45

Anderson MJ (2001) A new method for non parametric mul-tivariate analysis of variance. Austral Ecol 26: 32−46

Anderson M, Gorley R, Clarke K (2007) PERMANOVA+ forPRIMER: a guide to software and statistical methods.Primer-e, Plymouth

Arthington AH, Pusey BJ (2003) Flow restoration and pro-tection in Australian rivers. River Res Appl 19: 377−395

Arthington AH, Bunn SE, Poff NL, Naiman RJ (2006) Thechallenge of providing environmental flow rules to sus-tain river ecosystems. Ecol Appl 16: 1311−1318

Azam F, Hodson R (1977) Size distribution and activity ofmarine microheterotrophs. Limnol Oceanogr 22: 492−501

Azam F, Fenchel T, Field J, Gray J, Meyer-Reil L, ThingstadF (1983) The ecological role of water-column microbes inthe sea. Mar Ecol Prog Ser 10: 257−263

Baldwin D, Wilson J, Gigney H, Boulding A (2010) Influenceof extreme drawdown on water quality downstream of alarge water storage reservoir. River Res Appl 26: 194−206

Blum LK, Mills AL (2012) Estuarine microbial ecology. In: Day JW, Crump BC, Kemp WM, Yáñez-Arancibia A (eds)Estuarine ecology, 2nd edn. Wiley-Blackwell, p 235–261

Brown MV, Schwalbach MS, Hewson I, Fuhrman JA (2005)Coupling 16S ITS rDNA clone libraries and automatedribosomal intergenic spacer analysis to show marinemicrobial diversity: development and application to atime series. Environ Microbiol 7: 1466−1479

Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K andothers (2010) QIIME allows analysis of high-throughputcommunity sequencing data. Nat Methods 7: 335−336

Carvalho P, Thomaz S, Bini L (2003) Effects of water level,abiotic and biotic factors on bacterioplankton abundancein lagoons of a tropical floodplain (Paraná River, Brazil).Hydrobiologia 510: 67−74

Chung CC, Huang CY, Gong GC, Lin YC (2014) Influence ofthe Changjiang River flood on Synechococcus ecology inthe surface waters of the East China Sea. Microb Ecol 67: 273−285

Cole JJ, Caraco NF (2001) Carbon in catchments: connect-ing terrestrial carbon losses with aquatic metabolism.Mar Freshw Res 52: 101−110

Crespo BG, Pommier T, Fernández Gómez B, Pedrós Alió C(2013) Taxonomic composition of the particle-attachedand free-living bacterial assemblages in the NorthwestMediterranean Sea analyzed by pyrosequencing of the16S rRNA. MicrobiologyOpen 2: 541−552

Crump BC, Hobbie JE (2005) Synchrony and seasonality inbacterioplankton communities of two temperate rivers.Limnol Oceanogr 50: 1718−1729

Crump BC, Hopkinson CS, Sogin ML, Hobbie JE (2004)Microbial biogeography along an estuarine salinity gra-dient: combined influences of bacterial growth and resi-dence time. Appl Environ Microbiol 70: 1494−1505

Curtis TP, Sloan WT (2004) Prokaryotic diversity and its lim-its: microbial community structure in nature and implica-tions for microbial ecology. Curr Opin Microbiol 7: 221−226

Dalzell BJ, Filley TR, Harbor JM (2007) The role of hydrol-ogy in annual organic carbon loads and terrestrialorganic matter export from a midwestern agriculturalwatershed. Geochim Cosmochim Acta 71: 1448−1462

Del Giorgio PA, Cole JJ (1998) Bacterial growth efficiency innatural aquatic systems. Annu Rev Ecol Syst 29: 503−541

Del Giorgio PA, Williams PJleB (2005) Respiration in aquaticecosystems. Oxford University Press, Oxford

Dennis PG, Seymour J, Kumbun K, Tyson GW (2013)Diverse populations of lake water bacteria exhibitchemotaxis towards inorganic nutrients. ISME J 7: 1661−1664

DeSantis TZ, Hugenholtz P, Larsen N, Rojas M and others(2006) Greengenes, a chimera-checked 16S rRNA genedatabase and workbench compatible with ARB. ApplEnviron Microbiol 72: 5069−5072

Diaz RJ, Rosenberg R (2008) Spreading dead zones and con-sequences for marine ecosystems. Science 321: 926−929

Dudgeon D (2010) Prospects for sustaining freshwater bio -diversity in the 21st century: linking ecosystem structureand function. Curr Opin Environ Sustain 2: 422−430

Dudgeon D, Arthington AH, Gessner MO, Kawabata ZIand others (2006) Freshwater biodiversity: importance,threats, status and conservation challenges. Biol RevCamb Philos Soc 81: 163−182

Eaton AD, Franson MAH (2005) Standard methods for theexamination of water & wastewater, 21st edn. AmericanPublic Health Association, Washington, DC

Eaton AD, Clesceri LS, Greenberg AE, Franson MAH (1995)Standard methods for the examination of water and

202A

utho

r cop

y

http://dx.doi.org/10.1017/S1464793105006950http://dx.doi.org/10.1016/j.cosust.2010.09.001http://dx.doi.org/10.1128/AEM.03006-05http://dx.doi.org/10.1038/ismej.2013.47http://dx.doi.org/10.1146/annurev.ecolsys.29.1.503http://dx.doi.org/10.1016/j.gca.2006.12.009http://dx.doi.org/10.1016/j.mib.2004.04.010http://dx.doi.org/10.1128/AEM.70.3.1494-1505.2004http://dx.doi.org/10.4319/lo.2005.50.6.1718http://dx.doi.org/10.1002/mbo3.92http://dx.doi.org/10.1071/MF00084http://dx.doi.org/10.1007/s00248-013-0299-8http://dx.doi.org/10.1023/B%3AHYDR.0000008532.71152.38http://dx.doi.org/10.1038/nmeth.f.303http://dx.doi.org/10.1111/j.1462-2920.2005.00835.xhttp://dx.doi.org/1http://dx.doi.org/10.3354/meps010257http://dx.doi.org/10.4319/lo.1977.22.3.0492http://dx.doi.org/10.1890/1051-0761(2006)016[1311%3ATCOPEF]2.0.CO%3B2http://dx.doi.org/10.1002/rra.745http://dx.doi.org/10.1016/S0022-2836(05)80360-2http://dx.doi.org/10.1016/j.micres.2005.02.005http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=9925576&dopt=Abstract

Carney et al.: Bacterioplankton response to a river inflow

wastewater, 19th edn. American Public Health Associa-tion, Washington, DC

Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R (2011)UCHIME improves sensitivity and speed of chimeradetection. Bioinformatics 27: 2194−2200

Edwards RT, Meyer JL (1986) Production and turnover ofplanktonic bacteria in two southeastern blackwaterrivers. Appl Environ Microbiol 52: 1317−1323

Entwisle TJ, Sonneman JA, Lewis SH (1997) Freshwateralgae in Australia. Sainty & Associates, Sydney

Farjalla VF, Marinho CC, Faria BM, Amado AM, EstevesFdeA, Bozelli RL, Giroldo D (2009) Synergy of fresh andaccumulated organic matter to bacterial growth. MicrobEcol 57: 657−666

Fawcett SE, Ward BB (2011) Phytoplankton succession andnitrogen utilization during the development of anupwelling bloom. Mar Ecol Prog Ser 428: 13−31

Fuhrman JA, Steele JA (2008) Community structure of mar-ine bacterioplankton: patterns, networks, and relation-ships to function. Aquat Microb Ecol 53: 69−81

García-Armisen T, İnceoğlu Ö, Ouattara NK, Anzil A, Ver-banck MA, Brion N, Servais P (2014) Seasonal variationsand resilience of bacterial communities in a sewage polluted urban river. PLoS ONE 9: e92579

Gasol JM, Del Giorgio PA (2000) Using flow cytometry forcounting natural planktonic bacteria and understandingthe structure of planktonic bacterial communities. SciMar 64: 197−224

Gibbons SM, Caporaso JG, Pirrung M, Field D, Knight R,Gilbert JA (2013) Evidence for a persistent microbialseed bank throughout the global ocean. Proc Natl AcadSci USA 110: 4651−4655

Gram L, Melchiorsen J, Bruhn JB (2010) Antibacterial activ-ity of marine culturable bacteria collected from a globalsampling of ocean surface waters and surface swabs ofmarine organisms. Mar Biotechnol 12: 439−451

Haas BJ, Gevers D, Earl AM, Feldgarden M and others(2011) Chimeric 16S rRNA sequence formation anddetection in Sanger and 454-pyrosequenced PCR amplicons. Genome Res 21: 494−504

Hadwen WL, Fellows CS, Westhorpe DP, Rees GN and others (2010) Longitudinal trends in river functioning: patterns of nutrient and carbon processing in three Aus-tralian rivers. River Res Appl 26: 1129−1152

Hahn MW, Kasalický V, Jezbera J, Brandt U, Jezberová J,Šimek K (2010) Limnohabitans curvus gen. nov., sp. nov.,a planktonic bacterium isolated from a freshwater lake.Int J Syst Evol Microbiol 60: 1358−1365

Halstvedt CB, Rohrlack T, Andersen T, Skulberg O, Edvard-sen B (2007) Seasonal dynamics and depth distribution ofPlanktothrix spp. in Lake Steinsfjorden (Norway) relatedto environmental factors. J Plankton Res 29: 471−482

Havens KE, Work KA, East TL (2000) Relative efficiencies ofcarbon transfer from bacteria and algae to zooplanktonin a subtropical lake. J Plankton Res 22: 1801−1809

Henley W, Patterson M, Neves R, Lemly AD (2000) Effects ofsedimentation and turbidity on lotic food webs: a concisereview for natural resource managers. Rev Fish Sci 8: 125−139

Hewson I, Fuhrman JA (2004) Richness and diversity of bacterioplankton species along an estuarine gradientin Moreton Bay, Australia. Appl Environ Microbiol 70:3425–3433

Hitchcock JN, Mitrovic SM (2013) Different resource limita-tion by carbon, nitrogen and phosphorus between base

flow and high flow conditions for estuarine bacteria andphytoplankton. Estuar Coast Shelf Sci 135: 106−115

Hitchcock JN, Mitrovic SM (2014) Highs and lows: the effectof differently sized freshwater inflows on estuarine car-bon, nitrogen, phosphorus, bacteria and chlorophyll adynamics. Estuar Coast Shelf Sci 156: 71−82

Hitchcock JN, Mitrovic SM, Kobayashi T, Westhorpe DP(2010) Responses of estuarine bacterioplankton, phyto-plankton and zooplankton to dissolved organic carbon(DOC) and inorganic nutrient additions. Estuaries Coasts33: 78−91

Jansson M, Bergström AK, Lymer D, Vrede K, Karlsson J(2006) Bacterioplankton growth and nutrient use effi-ciencies under variable organic carbon and inorganicphosphorus ratios. Microb Ecol 52: 358−364

Junk WJ, Bayley PB, Sparks RE (1989) The flood pulse con-cept in river-floodplain systems. Publ Spec Can SciHalieut Aquat 106: 110−127

Kelly DP, Wood AP, Stackebrandt E (2005) Genus IIThiobacillus Beijerinck 1904b, 597AL. In: Brenner DJ,Krieg NR, Staley JT, Garrity GM (eds) Bergey’s Manual®of Systematic Bacteriology, 2nd edn, Vol 2, Part C.Springer, New York, NY, p 764–769

Kim M, Morrison M, Yu Z (2011) Evaluation of different par-tial 16S rRNA gene sequence regions for phylogeneticanalysis of microbiomes. J Microbiol Methods 84: 81−87

Kirchman DL, Dittel AI, Findlay SE, Fischer D (2004)Changes in bacterial activity and community structure inresponse to dissolved organic matter in the HudsonRiver, New York. Aquat Microb Ecol 35: 243−257

Kritzberg E, Cole JJ, Pace ML, Granéli W, Bade DL (2004)Autochthonous versus allochthonous carbon sources ofbacteria: results from whole-lake C-13 addition experi-ments. Limnol Oceanogr 49: 588−596

Kumar PS, Brooker MR, Dowd SE, Camerlengo T (2011) Tar-get region selection is a critical determinant of com -munity fingerprints generated by 16S pyrosequencing.PLoS ONE 6: e20956

Lauro FM, McDougald D, Thomas T, Williams TJ and others(2009) The genomic basis of trophic strategy in marinebacteria. Proc Natl Acad Sci USA 106: 15527−15533

Liao PC, Huang BH, Huang S (2007) Microbial communitycomposition of the Danshui river estuary of Northern Tai-wan and the practicality of the phylogenetic method inmicrobial barcoding. Microb Ecol 54: 497−507

Lin X, McKinley J, Resch CT, Kaluzny R and others (2012)Spatial and temporal dynamics of the microbial commu-nity in the Hanford unconfined aquifer. ISME J 6: 1665−1676

Maranger RJ, Pace ML, del Giorgio PA, Caraco NF, Cole JJ(2005) Longitudinal spatial patterns of bacterial produc-tion and respiration in a large river−estuary: implicationsfor ecosystem carbon consumption. Ecosystems 8: 318−330

Marie D, Partensky F, Jacquet S, Vaulot D (1997) Enumera-tion and cell cycle analysis of natural populations of mar-ine picoplankton by flow cytometry using the nucleicacid stain SYBR Green I. Appl Environ Microbiol 63: 186−193

Mašín M, Jezbera J, Nedoma J, Straškrabová V, Hejzlar J,Šimek K (2003) Changes in bacterial community compo-sition and microbial activities along the longitudinal axisof two canyon-shaped reservoirs with different inflowloading. Hydrobiologia 504: 99−113

McDonald D, Price MN, Goodrich J, Nawrocki EP and oth-ers (2011) An improved Greengenes taxonomy with

203A

utho

r cop

y

http://dx.doi.org/10.1038/ismej.2011.139http://dx.doi.org/10.1023/B%3AHYDR.0000008512.04563.0bhttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=16535483&dopt=Abstracthttp://dx.doi.org/10.1007/s10021-003-0071-xhttp://dx.doi.org/10.1038/ismej.2012.26http://dx.doi.org/10.1007/s00248-007-9217-2http://dx.doi.org/10.1073/pnas.0903507106http://dx.doi.org/10.1371/journal.pone.0020956http://dx.doi.org/10.4319/lo.2004.49.2.0588http://dx.doi.org/10.3354/ame035243http://dx.doi.org/10.1016/j.mimet.2010.10.020http://dx.doi.org/10.1007/s00248-006-9013-4http://dx.doi.org/10.1007/s12237-009-9229-xhttp://dx.doi.org/10.1016/j.ecss.2013.05.001http://dx.doi.org/10.1080/10641260091129198http://dx.doi.org/10.1093/plankt/22.9.1801http://dx.doi.org/10.1093/plankt/fbm036http://dx.doi.org/10.1099/ijs.0.013292-0http://dx.doi.org/10.1002/rra.1321http://dx.doi.org/10.1101/gr.112730.110http://dx.doi.org/10.1007/s10126-009-9233-yhttp://dx.doi.org/10.1073/pnas.1217767110http://dx.doi.org/10.3989/scimar.2000.64n2197http://dx.doi.org/10.1371/journal.pone.0092579http://dx.doi.org/10.3354/ame01222http://dx.doi.org/10.3354/meps09070http://dx.doi.org/10.1007/s00248-008-9466-8http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=16347237&dopt=Abstracthttp://dx.doi.org/10.1093/bioinformatics/btr381

Aquat Microb Ecol 75: 187–205, 2015

explicit ranks for ecological and evolutionary analyses ofbacteria and archaea. ISME J 6: 610−618

Miletto M, Loy A, Antheunisse AM, Loeb R, Bodelier PL,Laanbroek HJ (2008) Biogeography of sulfate-reducingprokaryotes in river floodplains. FEMS Microbiol Ecol64: 395−406

Mitrovic SW, Westhorpe DP, Kobayashi T, Baldwin DS, RyanD, Hitchcock JN (2014) Short-term changes in zooplank-ton density and community structure in response to dif-ferent sources of dissolved organic carbon in an uncon-strained lowland river: evidence for food web support.J Plankton Res 36:1488–1500

Nair S, Simidu U (1987) Distribution and significance of het-erotrophic marine bacteria with antibacterial activity.Appl Environ Microbiol 53: 2957−2962

Newton RJ, Jones SE, Eiler A, McMahon KD, Bertilsson S(2011) A guide to the natural history of freshwater lakebacteria. Microbiol Mol Biol Rev 75: 14−49

NSWDPI (2003) The Hunter regulated river water sharingplan. New South Wales Department of Primary Indus-tries, Orange

Paerl HW, Pinckney JL, Fear JM, Peierls BL (1998) Ecosys-tem responses to internal and watershed organic matterloading: consequences for hypoxia in the eutrophyingNeuse River Estuary, North Carolina, USA. Mar EcolProg Ser 166: 17−25

Palijan G, Bogut I, Vidakovic J (2008) The impact of inunda-tion-isolation cycles on the culturable bacterioplanktonin the Danube River floodplain. Pol J Ecol 56: 391−403

Partensky F, Blanchot J, Vaulot D (1999) Differential distri-bution and ecology of Prochlorococcus and Synechococ-cus in oceanic waters: a review. Bull Inst Ocean MonacoSpec Issue 19: 457−476

Patel V, Munot H, Shouche YS, Madamwar D (2014)Response of bacterial community structure to seasonalfluctuation and anthropogenic pollution on coastal waterof Alang−Sosiya ship breaking yard, Bhavnagar, India.Bioresour Technol 161: 362−370

Petrone KC, Richards JS, Grierson PF (2009) Bioavailabilityand composition of dissolved organic carbon and nitro-gen in a near coastal catchment of south-western Aus-tralia. Biogeochemistry 92: 27−40

Pinhassi J, Gómez-Consarnau L, Alonso-Sáez L, Sala MM,Vidal M, Pedrós-Alió C, Gasol JM (2006) Seasonalchanges in bacterioplankton nutrient limitation and theireffects on bacterial community composition in the NWMediterranean Sea. Aquat Microb Ecol 44: 241–252

Poff NL, Zimmerman JK (2010) Ecological responses toaltered flow regimes: a literature review to inform thescience and management of environmental flows.Freshw Biol 55: 194−205

Pollard PC, Ducklow HW (2011) Ultrahigh bacterial produc-tion in a eutrophic subtropical Australian river: does virallysis short-circuit the microbial loop? Limnol Oceanogr56: 1115−1129

Pujalte MJ, Lucena T, Ruvira MA, Arahal DR, Macián MC(2014) The family Rhodobacteraceae, Vol 1. Springer,Berlin

Ramette A (2009) Quantitative community fingerprintingmethods for estimating the abundance of operationaltaxonomic units in natural microbial communities. ApplEnviron Microbiol 75: 2495−2505

Reshef D, Reshef Y, Finucane H, Grossman S and others(2011) Detecting novel associations in large datasets. Science 334: 1518−1524

Riemann L, Winding A (2001) Community dynamics of free-living and particle-associated bacterial assemblages dur-ing a freshwater phytoplankton bloom. Microb Ecol 42: 274−285

Rösel S, Grossart HP (2012) Contrasting dynamics in activityand community composition of free-living and particle-associated bacteria in spring. Aquat Microb Ecol 66: 169−181

Rossi P, Shani N, Kohler F, Imfeld G, Holliger C (2012) Ecol-ogy and biogeography of bacterial communities asso -ciated with chloroethene-contaminated aquifers. FrontMicrobiol 3: 260

Salles P, Bredeweg B, Araújo S (2006) Qualitative modelsabout stream ecosystem recovery: exploratory studies.Ecol Modell 194: 80−89

Santos L, Vaz L, Marcial Gomes NC, Vaz N, Dias JM, CunhaÂ, Almeida A (2014) Impact of freshwater inflow on bac-terial abundance and activity in the estuarine system Riade Aveiro. Estuar Coast Shelf Sci 138: 107−120

Schultz GE Jr, Kovatch JJ, Anneken EM (2013) Bacterialdiversity in a large, temperate, heavily modified river, asdetermined by pyrosequencing. Aquat Microb Ecol 70: 169−179

Seymour JR, Patten N, Bourne DG, Mitchell JG (2005) Spa-tial dynamics of virus-like particles and heterotrophicbacteria within a shallow coral reef system. Mar EcolProg Ser 288: 1−8

Sieczko A, Peduzzi P (2014) Origin, enzymatic response andfate of dissolved organic matter during flood and non-flood conditions in a river-floodplain system of theDanube (Austria). Aquat Sci 76: 115−129

Sinsabaugh RL, Findlay SEG (2003) Dissolved organic matter: out of the black box into the mainstream. In:Findlay SEG, Sinsabaugh RL (eds) Aquatic ecosystems:inter activity of dissolved organic matter. Elsevier, SanDiego, CA, p 479–498

Stahl DA, Hullar M, Davidson S (2006) The structureand function of microbial communities. In: DworkinM, Falkow S, Rosenberg E, Schleifer KH, StackebrandtE (eds) The prokaryotes. Springer, New York, NY,p 299–327

Steele JA, Countway PD, Xia L, Vigil PD and others (2011)Marine bacterial, archaeal and protistan association net-works reveal ecological linkages. ISME J 5: 1414−1425

Tang KW, Dziallas C, Grossart HP (2011) Zooplankton andaggregates as refuge for aquatic bacteria: protectionfrom UV, heat and ozone stresses used for water treat-ment. Environ Microbiol 13: 378−390

Tang X, Gao G, Qin B, Zhu L, Chao J, Wang J, Yang G(2009) Characterization of bacterial communities associ-ated with organic aggregates in a large, shallow,eutrophic freshwater lake (Lake Taihu, China). MicrobEcol 58: 307−322

Teeling H, Fuchs BM, Becher D, Klockow C and others(2012) Substrate-controlled succession of marine bac -terioplankton populations induced by a phytoplanktonbloom. Science 336: 608−611

Thiele S, Fuchs BM, Amann R, Iversen MH (2014) Coloniza-tion in the photic zone and subsequent changes duringsinking determine bacterial community composition inmarine snow. Appl Environ Microbiol 81:1463–1471

Tockner K, Malard F, Ward J (2000) An extension of theflood pulse concept. Hydrol Process 14: 2861−2883

Van der Gucht K, Vandekerckhove T, Vloemans N, Cousin Sand others (2005) Characterization of bacterial communi-

204A

utho

r cop

y

http://dx.doi.org/10.1016/j.femsec.2004.12.006http://dx.doi.org/10.1002/1099-1085(200011/12)14%3A16/17%3C2861%3A%3AAID-HYP124%3E3.0.CO%3B2-Fhttp://dx.doi.org/10.1128/AEM.02570-14http://dx.doi.org/10.1007/s00248-008-9482-8http://dx.doi.org/10.1111/j.1462-2920.2010.02335.xhttp://dx.doi.org/10.1038/ismej.2011.24http://dx.doi.org/10.1007/0-387-30741-9_13http://dx.doi.org/10.1007/s00027-013-0318-3http://dx.doi.org/10.3354/meps288001http://dx.doi.org/10.3354/ame01646http://dx.doi.org/10.1016/j.ecss.2013.12.021http://dx.doi.org/10.1016/j.ecolmodel.2005.10.018http://dx.doi.org/10.3389/fmicb.2012.00260http://dx.doi.org/10.3354/ame01568http://dx.doi.org/10.1007/s00248-001-0018-8http://dx.doi.org/10.1128/AEM.02409-08http://dx.doi.org/10.4319/lo.2011.56.3.1115http://dx.doi.org/10.1111/j.1365-2427.2009.02272.xhttp://dx.doi.org/10.3354/ame044241http://dx.doi.org/10.1007/s10533-008-9238-zhttp://dx.doi.org/10.1016/j.biortech.2014.03.033http://dx.doi.org/10.3354/meps166017http://dx.doi.org/10.1128/MMBR.00028-10http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=3435149&dopt=Abstracthttp://dx.doi.org/10.1093/plankt/fbu072http://dx.doi.org/10.1111/j.1574-6941.2008.00490.x

Carney et al.: Bacterioplankton response to a river inflow

ties in four freshwater lakes differing in nutrient load andfood web structure. FEMS Microbiol Ecol 53: 205−220

Vörösmarty CJ, McIntyre P, Gessner MO, Dudgeon D andothers (2010) Global threats to human water security andriver biodiversity. Nature 467: 555−561

Walsh S, Copeland C, Westlake M (2004) Major fish kills inthe northern rivers of NSW in 2001: causes, impacts andresponses. Fisheries final report series no. 68. New SouthWales Department of Primary Industries, Orange

Wear EK, Koepfler ET, Smith EM (2013) Spatiotemporalvariability in dissolved organic matter composition ismore strongly related to bacterioplankton communitycomposition than to metabolic capability in a blackwaterestuarine system. Estuaries Coasts 37: 119−133

Webster IT, Harris GP (2004) Anthropogenic impacts on theecosystems of coastal lagoons: modelling fundamentalbiogeochemical processes and management implica-tions. Mar Freshw Res 55: 67−78

Westhorpe DP, Mitrovic SM (2012) Dissolved organic carbon

mobilisation in relation to variable discharges and envi-ronmental flows in a highly regulated lowland river. MarFreshw Res 63: 1218−1230

Westhorpe DP, Mitrovic SM, Ryan D, Kobayashi T (2010)Limitation of lowland riverine bacterioplankton by dis-solved organic carbon and inorganic nutrients. Hydro -biologia 652: 101−117

Wilson JG, Devlin M (2013) Nutrients, food web and tropho-dynamic processes. Estuar Coast Shelf Sci 135: 1−4

Zhang J, Gilbert D, Gooday A, Levin L and others (2010)Natural and human-induced hypoxia and consequencesfor coastal areas: synthesis and future development. Biogeosciences 7: 1443−1467

Zhou J, Deng Y, Luo F, He Z, Tu Q, Zhi X (2010) Functionalmolecular ecological networks. MBio 1: e00169-10

Zwart G, Crump BC, Kamst-van Agterveld MP, Hagen F,Han SK (2002) Typical freshwater bacteria: an analysis ofavailable 16S rRNA gene sequences from plankton oflakes and rivers. Aquat Microb Ecol 28: 141−155

205

0

10

20

30

40

50

60

70

80

90

100

1 3 4 6 7 2 5

Chl

a (µ

g l–

1 )

Hunter River sites PatersonRiver

FebruaryMarchAprilMay June

WilliamsRiver

Appendix 1. Chl a concentrations sampled at the Hunter River and its tributaries (Paterson River andWilliams River) from February to June 2013

Editorial responsibility: Tom Battin, Vienna, Austria

Submitted: August 4, 2014; Accepted: April 22, 2015Proofs received from author(s): June 24, 2015

➤

➤

➤

➤

➤

➤

➤

➤

➤

Aut

hor c

opy

http://dx.doi.org/10.3354/ame028141http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20941329&dopt=Abstracthttp://dx.doi.org/10.5194/bg-7-1443-2010http://dx.doi.org/10.1016/j.ecss.2013.10.018http://dx.doi.org/10.1007/s10750-010-0322-8http://dx.doi.org/10.1071/MF12122http://dx.doi.org/10.1071/MF03068http://dx.doi.org/10.1007/s12237-013-9651-yhttp://dx.doi.org/10.1038/nature09440

cite123: cite124: cite125: cite10: cite16: cite30: cite18: cite32: cite27: cite41: cite8: cite38: cite52: cite45: cite54: cite47: cite61: cite56: cite49: cite63: cite65: cite58: cite67: cite81: cite74: cite69: cite90: cite78: cite85: cite87: cite1: cite89: cite96: cite5: cite98: cite9: cite11: cite13: cite22: cite15: cite24: cite26: cite40: cite19: cite28: cite42: cite6: cite37: cite51: cite35: cite33: cite100: cite101: cite55: cite102: cite104: cite71: cite64: cite107: cite80: cite59: cite73: cite82: cite75: cite109: cite84: cite91: cite111: cite93: cite112: cite113: cite88: cite95: cite115: cite117: cite118: cite119: cite99: cite120: cite121: cite122: