Impacts of Climate-Related Drivers on the Benthic Nutrient Filter in a Shallow Photic Estuary

Impacts of Climate-Related Drivers on the Benthic NutrientFilter in a Shallow Photic Estuary

Iris C. Anderson & Mark J. Brush & Michael F. Piehler & Carolyn A. Currin &

Jennifer W. Stanhope & Ashley R. Smyth & Johnathan D. Maxey &

Meaghan L. Whitehead

Received: 31 January 2013 /Revised: 13 June 2013 /Accepted: 17 June 2013# Coastal and Estuarine Research Federation 2013

Abstract In shallow photic systems, the benthic filter, in-cluding microphytobenthos and denitrifiers, is important inpreventing or reducing release of remineralized NH4

+ to thewater column. Its effectiveness can be impacted by climate-related drivers, including temperature and storminess, whichby increasing wind and freshwater delivery can resuspendsediment, reduce salinity and deliver nutrients, total suspendedsolids, and chromophoric dissolved organic matter (CDOM) tocoastal systems. Increases in temperature and freshwater de-livery may initiate a cascade of responses affecting benthicmetabolismwith impacts on sediment properties, which in turnregulate nitrogen cycling processes that either sequester (viamicrophytobenthos), remove (via denitrification), or increasesediment nitrogen (via remineralization, nitrogen fixation, anddissimilatory nitrate reduction to ammonium). We conducted aseasonal study at shallow stations to assess the effects offreshwater inflow, temperature, wind, light, and CDOM on

sediment properties, benthic metabolism, nitrogen cycling pro-cesses, and the effectiveness of the benthic filter. We alsoconducted a depth study to constrain seasonally varying pa-rameters such as temperature to better assess the effects of lightavailability and water depth on benthic processes. Based onrelationships observed between climatic drivers and responsevariables, we predict a reduction in the effectiveness of thebenthic filter over the long term with feedbacks that willincrease effluxes of N to the water column with the potentialto contribute to system eutrophication. This may push shallowsystems past a tipping point where trophic status moves fromnet autotrophy toward net heterotrophy, with new baselinescharacterized by degraded water quality.

Keywords Shallow photic estuary . Benthic filter . Climate .

Microphytobenthos .Benthicmetabolism .Remineralization .

Nitrogen fixation . Denitrification

Introduction

Shallow microtidal estuaries are common coastal features(Durr et al. 2011) comprising nearly half of estuarine surfacearea in the USA (NOAA (National Oceanic and AtmosphericAdministration) 2011). In shallow systems with photic sed-iments, phytoplankton production and biomass can be farless than predicted based on nutrient loading (Borum andSand-Jensen 1996; Nixon et al. 2001; McGlathery et al.2007). Supplies of external and remineralized nitrogen (N)that would otherwise support pelagic production in thesesystems are modulated by benthic autotrophs and hetero-trophs that sequester it through primary production or re-move it by denitrification (DNF) (Anderson et al. 2003,

Electronic supplementary material The online version of this article(doi:10.1007/s12237-013-9665-5) contains supplementary material,which is available to authorized users.

I. C. Anderson (*) :M. J. Brush : J. W. Stanhope : J. D. Maxey :M. L. WhiteheadVirginia Institute of Marine Science, College of William & Mary,P.O. Box 1346, Gloucester Point, VA 23062, USAe-mail: [email protected]

M. F. Piehler :A. R. SmythInstitute of Marine Sciences, The University of North Carolina,3431 Arendell St., Morehead City, NC 28557, USA

C. A. CurrinNational Oceanic and Atmospheric Administration, NationalOcean Service, National Centers for Coastal and Ocean Science,101 Pivers Island Rd, Beaufort, NC 28516, USA

Estuaries and CoastsDOI 10.1007/s12237-013-9665-5

2010; Sundbäck et al. 2004; Eyre et al. 2011). The ecologicalservice of N retention and removal by the benthic microbialcommunity is referred to here as the benthic filter.

The functions of the benthic filter depend on the commu-nity of microbial autotrophs and heterotrophs occurring over asteep redox gradient in the top few millimeters of sediment inshallow photic systems. This gradient is vulnerable to changesin temperature, available light, freshwater and nutrient deliv-ery, and resuspension events (MacIntyre et al. 1996; Sundbäcket al. 2004; de Jonge et al. 2012), all of which interact inunpredictable ways and often with contrasting effects on thebenthic filter.

Previous studies in a wide variety of shallow estuaries andlagoons have shown strong coupling between processes in thebenthic and pelagic zones. Ammonium (NH4

+) remineralizedfrom sediment organic matter along with that produced byeither nitrogen fixation (NFIX) or by dissimilatory nitratereduction to ammonium (DNRA), collectively called ammo-nification (AMN), can be oxidized to nitrite and nitrate(NO2

−+NO3−) by nitrification (NIT), flux to the water col-

umn, and support further pelagic primary production or beintercepted by the benthic filter. In photic sediments, benthicautotrophs drive the function of the benthic filter; otherwise,DNF will likely be the dominant process removing N(Sundbäck et al. 2000, 2004; Anderson et al. 2003, 2010;Eyre and Ferguson 2005; McGlathery et al. 2007).

The microphytobenthos (MPB) has the capacity to take upnutrients either from the water column or sediments providedthat light is sufficient to support photosynthesis (Sundbäcket al. 2004; Kromkamp et al. 2006; MacIntyre et al. 1996;Anderson et al. 2003; Hardison et al. 2011). In addition,MPB can indirectly regulate microbial processes, includingNIT, DNF, anammox, NFIX, and DNRA, by release of O2

and extracellular polymeric substances (Underwood andSmith 1998; Wolfstein et al. 2002) and by competition forsubstrates (MacIntyre et al. 1996; Kromkamp et al. 2006;Joye and Anderson 2008). Whereas an autotrophic benthosis likely to be a net sink, a heterotrophic benthos is likely tobe a net source of NH4

+ to the overlying water (Cook et al.2009; Engelsen et al. 2008). In addition, heterotrophic sed-iments accumulate sulfide from sulfate reduction and cancontribute to water column hypoxia and anoxia.

Photic sediments not only sustain MPB production andnutrient uptake but also support N2 uptake by autotrophicnitrogen fixers, which can offset N removal by DNF (Fulweileret al. 2007, 2013). Whereas MPB may outcompete denitrifiersfor substrate, theymay also support DNF by supplying organicmatter (OM) when NO3

− is not limiting (An and Joye 2001).The quality and quantity of OM deposited to the benthos mayfurther influence the relative importance of DNF and NFIX(Fulweiler et al. 2007, 2013). Temperature complicates com-petition between benthic processes by differential support, forexample, of respiration (R) over gross primary production

(GPP) (Harris et al. 2006; O’Connor et al. 2009) or of partic-ular N cycling processes (Veraart et al. 2011).

A series of research questions were posed to more fullyunderstand the impacts of multiple climatically controlled,physical and chemical drivers on benthic processes in a shal-low photic estuary. They included: (1) How does freshwaterdischarge impact the photic area of a shallow estuary? (2)How do sediment net metabolic and N cycling rates respondto the interacting effects of light, temperature, salinity, andMPB biomass? (3) How effective is the benthic filter acrossthe estuarine salinity gradient? We hypothesized that whensufficient light is available to support benthic autotrophy,MPB will both intercept N produced by benthic ammonifica-tion and take up water column N. In the absence of lightsufficient to support benthic photosynthesis, we predicted thatfixed N will be removed primarily by DNF; however, thisremoval of N will be partially offset by heterotrophic NFIX.

Methods

Site Description

This study was performed in the New River Estuary (NRE), asmall system of 7,810 ha located in Onslow County, NC,surrounded by Marine Corps Base Camp Lejeune (63,131 ha)with the city of Jacksonville at the head of the estuary and asurrounding community of approximately 180,000 people. Thismicrotidal estuary withmore than 50% of the benthic surface atdepths less than 2m (mean sea level,MSL) serves as amodel ofa shallow system with a substantial fraction of photic sedimentsin which the benthic filter is likely to play an important role inmodulating nutrient enrichment. The river-dominated NRE isvulnerable to both anthropogenic and natural disturbances withshifts in freshwater discharge, inorganic and organic nutrientloads, chromophoric dissolved organic matter (CDOM), salin-ity, and light availability occurring over short time scales inresponse to storm and other weather-related events. Flushingtime, measured using the fraction of fresh water method, withinthe NRE ranges from 8 to 187 days, averaging 70 days (Ensignet al. 2004). Freshwater discharge, nutrient, OM, and particulateloads to the NRE are primarily from the New River; dischargeand nutrient loads were measured by US Geological Survey atgauging station #02093000 at Gum Branch, NC. Land use inboth the New River and Southwest Creek watersheds is dom-inated by agriculture and includes numerous swine confinedanimal feeding operations (CAFOs). Prior to 1998 dischargesfrom CAFOs and wastewater treatment facilities (WWTF) inthe city of Jacksonville and on Marine Corps Base CampLejeune resulted inmassive phytoplankton blooms, widespreadhypoxia, and fish kills, such that the estuary was named one ofthe most eutrophic in the Southeastern United States (Brickeret al. 1999; Mallin et al. 2005). Although upgrades to the

Estuaries and Coasts

WWTFs in 1998 markedly improved water quality (Mallinet al. 2005), the NRE continues to receive high nutrient loadsfrom the New River watershed upstream of Jacksonville, epi-sodic WWTF spills and direct discharge from one WWTF, andadditional inputs via atmospheric deposition and tidal ex-changes with Onslow Bay.

Experimental Design

Two types of experiments were performed during this study.To identify the dominant drivers and interactions betweendrivers regulating benthic processes, we first performed aseries of seasonal shallow water measurements of metabolism,benthic fluxes, and N cycling process rates over a 3-yearperiod. All shallow water sites were located along the estuarinegradient at a depth of 0.5 m (mean low water), and metabolicand flux incubations were performed in situ during July andOctober 2008 and May and July 2009. In an additional studyperformed in March 2010, samples were incubated underambient conditions of light and temperature in an environmen-tal chamber. We identify the six sites used for the seasonalstudy as shallow seasonal sites [Fig. 1; Jacksonville (site 1),Southwest Creek (site 2), Wallace Creek (site 3), French Creek(site 4), Courthouse Bay (site 5), and Traps Bay (site 6)].

A second group of experiments were performed during twoseasons, April 2011 and July 2010, at multiple depths to allowus to constrain some of the drivers in order to better determinethe effects of light and water depth on benthic processes. Weidentify the sites used for the depth experiment as multidepthsites. They were randomly selected (three replicates per site)within three water depth contours (0.5, 1.5, and 3 mMSL), asdetermined from recent bathymetry, within each of three re-gions (upper, middle, and lower estuary) (Fig. 1).

Determinations of Sediment Characteristics

From 2007 to 2011, sampling at the six shallow seasonal sitesin the NRE was conducted for physical, chemical, and bio-logical characteristics. Triplicate sediment cores were collect-ed at each site to a depth of 5 cm for analyses of OM content,total N, total particulate organic carbon, and extractable NH4

+,as described by Anderson et al. (2003). Porewater, sampled at5 cm below the sediment surface using a stainless steel push-point sampler (2 cm screen; MHE Products, East Tawas, MI,USA) was analyzed for hydrogen sulfide (Cline 1969). Foranalysis of benthic chlorophyll a (bchla), triplicate sedimentsamples were collected using a 5-mL cut-off syringe (Normject,1.1 cm inner diameter (ID)), subsampled to a depth of 0.3 cm,stored frozen at −15 °C until extracted in 10 mL of a solution of90 % ethanol in deionized water. Samples were then vortexedand sonicated for 30 s and stored frozen for 24–28 h beforefiltration (Pall Acrodisc CR 0.45 μm PTFE membrane) andanalysis on a spectrophotometer (model DU800, Beckman

Coulter, Inc., Brea, CA, USA; Lorenzen 1967, as modified byPinckney et al. 1994). For the depth experiments, three samplesper water depth contour in each region of the estuary werecollected for determination of sediment characteristics.

Relationship Between Bchla and Wind Wave Energy

Additional sampling of surface sediments to determine the dis-tribution of bchla at various water column depths across theestuary was conducted during spring and fall 2008–2010 alongtransects at the six shallow seasonal sites plus an additional ninesites, chosen to represent a wide range of wind wave exposure.Each transect included samples obtained from 0.25, 0.5, 1.0, 1.5,and 2.0 m depths. Representative wave energy (RWE), withunits of Joules per meter, was calculated using a wave energymodel (WeMO), described byMalhotra and Fonseca (2007). Forthis exercise, we used the calculated RWE for the 1.0-m depthlocation to represent overall wave energy at each site. RWEcalculations utilized bathymetry, fetch, and wind data from theNew River Air Station (top 5 % of wind events between 2006and 2008) and represent a long-term characterization of the waveenergy experienced by a particular site.

Analyses of Water Quality Parameters

From 2007 to 2011, triplicate water samples were collectedseasonally at mid-depth from the shallow seasonal sites, filtered[0.45 μm Whatman polyethersulfone (PES)], and frozen untilanalyzed for dissolved inorganic nitrogen (DIN; NO3

−, NO2−,

and NH4+; Smith and Bogren 2001; Liao 2001), dissolved

Fig. 1 Bathymetry of the New River Estuary, NC (depths in metersbelow mean sea level). Locations of shallow seasonal (diamond) andmulti-depth (circle) sites

Estuaries and Coasts

organic nitrogen (Koroleff 1983), and dissolved inorganicphosphorus (DIP; PO4

3−, Knepel and Bogren 2001) with aLachat QuikChem 8000 automated ion analyzer (Lachat In-struments, Milwaukee, WI, USA; detection limits for NO3

−,NH4

+, and PO43− are 0.20, 0.36, and 0.16 μM, respectively).

For determinations of chlorophyll a (chl a) and phaeophytin,triplicate water samples were filtered (0.7 μm GF/F) andextracted as described in Anderson et al. (2003). Additionaltriplicate water samples for each site were filtered (0.22 μmpolycarbonate filter), frozen, and analyzed for CDOM absorp-tion at 440 nm (Kirk 1994) with a Beckman Coulter DU800spectrophotometer. Incident and underwater photosynthetical-ly active radiation (PAR) at multiple water depths were mea-sured at each site simultaneously with a Li-Cor quantum decksensor (LI-190SA) and underwater quantum sensor (LI-192SA; Li-Cor, Inc., Lincoln, NE, USA). The vertical lightattenuation coefficient (Kd(PAR)) was calculated using theBeer–Lambert law. Salinity, temperature, DO, in vivo chl a,and turbidity were measured in triplicate at mid-water columndepth at each site with a YSI 6600 datasonde (YSI, Inc.,Yellow Springs, OH, USA).

Surface Water Quality Mapping with Dataflow

Dataflow surface water quality mapping surveys were conductedseasonally from 2008 to 2011. The dataflow system, which wasfirst described by Madden and Day (1992), is currently instru-mented with a YSI 6600 datasonde, WET Labs CDOM sensor,Garmin global positioning system, and data acquisition system.Surveys were run along the shoreline of the entire estuary,collecting continuous water quality measurements (approximate-ly every 30 m; more than 4,500 datapoints). Calibration samplesfor CDOM, extracted chl a, and dissolved nutrients were collect-ed at ten stations in triplicate, including the six shallow seasonalsites. For each survey, regressions were conducted for (1) in situYSI chl a measurements versus laboratory analyzed chl a sam-ples and (2) in situ WET Labs CDOM measurements versusCDOM absorption coefficients. These regression relationshipswere utilized to predict extracted chl a and CDOM absorptioncoefficients for each datapoint in a specific dataflow survey. Datafrom the multiple seasonal field surveys during 2008 and 2009were combined to develop an empirical model of light attenua-tion, Kd(PAR). The composite multiple linear regression modelfor predicting Kd(PAR) (in per meter) as a function of chl a(extracted) (in micrograms per liter), turbidity (in nephelometricturbidity units), and CDOM (absorption coefficient at 440 nm; inper meter) was:

Kd PARð Þ ¼ 0:71þ ð0:005 � chlaÞ þ ð0:006 � NTUÞþð0:26 � CDOMÞ r2 ¼ 0:76; p < 0:001

� � ð1Þ

This model was applied to all dataflow surveys to predictKd(PAR) for each datapoint. Maps of interpolated surface

water quality measurements for the entire NRE were made inArcGIS 9.3 using Spatial Analyst and an inverse distanceweighting technique. Using the interpolated Kd(PAR) valuesand NRE bathymetry, the percent of bottom area in the NREreceiving at least 1 % of surface irradiance was determined.

Benthic Metabolism and Nutrient Fluxes—Seasonal Studies

Sediment cores (clear acrylic, 13.3 cm ID×40 cm tall),collected in triplicate to a depth of 20 cm at each of the sixshallow seasonal sites (sites 1–6), were used for concurrentdeterminations of GPP, R, net community production (NCP),and nutrient fluxes. Three additional cores, filled with watercollected from each site, served as water blanks to distinguishwater column from sediment processes. Capped sediment andwater blank cores from sites 1 and 2 were incubated at site 2,sites 3 and 4 at site 4, and sites 5 and 6 at site 5 under in situconditions for 24 h (1S of online electronic supplementarymaterial; ESM_1.pdf). Paired sites had similar light fields(Table 1). In March 2010, cores from only sites 2, 3, and 5were incubated in an environmental chamber at the VirginiaInstitute ofMarine Science (VIMS) at in situ light (underwaterPAR at sediment surface of 213, 360, and 501 μE m−2 s−1,respectively) and temperature conditions (13.8 °C). The watercolumns inside the cores were mixed with magnetic stirrersattached to clear lids. To determine net exchanges of nutrients,dissolved inorganic carbon (DIC), and DO between the sedi-ment and overlying water, water samples were collected at 4-to 6-h intervals during the day (11.5 to 14 h total depending onseason) and 10 to 12.5-h intervals at night, starting at dusk andending at the following dawn (Anderson et al. 2003). Coreswere connected to a reservoir so that water removed duringsampling was replacedwith water from the respective site. DOconcentrations in the sampled water were measured using aHach LDO101 Luminescent DO sensor (Hach Co., Loveland,CO, USA). Samples for DIC were collected in 8-mL hungatetubes (Bellco Glass Inc., Vineland, NJ, USA), pre-spiked with15 μL saturated mercuric chloride, and kept cold under wateruntil analyzed within 4 weeks using a Li-Cor 6252 infraredgas analyzer (Neubauer and Anderson 2003). Changes of DICin the light and dark were used for determination of rates ofbenthic and pelagic metabolism, including R, GPP, and NCP.Water samples taken concurrently with the DO and DICmeasurements were filtered and frozen until analyzed forDIN and DIP, as described above. Net uptake or release ofnutrients from sediment was determined from changes innutrient concentrations during the same incubation periods.Benthic DIC, DO, and nutrient fluxes were corrected for DIC,DO, and nutrient uptake or release measured in the waterblanks. Benthic metabolism (DIC based; in millimolesC per square meter per day) and daily nutrient fluxes(in millimoles N per square meter per day) were calcu-lated as follows:

Estuaries and Coasts

R ¼ Fd � 24h ð2Þ

GPP ¼ h1 � Fd−F1ð Þ ð3Þ

NCP ¼ − GPP−Rð Þ ð4Þ

Dailynutrient flux ¼ F1 � h1ð Þ þ Fd � hdð Þ ð5Þwhere Fd = hourly flux in the dark (in millimoles N per squaremeter per hour), Fl = hourly flux in the light (in millimoles Nper square meter per hour), hl = hours of light, and hd = hoursof dark. NCP (Eq. 4), shown as a negative number when GPPexceedsR, represents net autotrophy and net uptake of C; whenpositive, it represents net heterotrophy and net release of C.

MPBN demand (in millimoles N per square meter per day)was calculated (Eq. 6) based on 90% of GPP (in millimoles Cper square meter per day), which assumes that diatom respi-ration equals 10 % of gross primary production, as discussedin Cloern (1987), and a C/N of 9:1 (Sundbäck et al. 1991).Because MPB have been shown to exude as much as 70 % ofthe C taken up by photosynthesis as extracellular polymericsubstances and colloidal C (Wolfstein et al. 2002), rates ofMPB N demand were corrected based on the conservativeassumption that 50 % of the C fixed was exuded and did notcontribute to MPB biomass.

MPBN� Demand ¼ 0:5� 0:9� GPPð Þ � 1=9 ð6Þ

Benthic Metabolism and Nutrient Fluxes—DepthExperiments

For the depth experiments, benthic metabolism and nitrogenfluxes were measured in triplicate cores taken at random

locations within each of the nine multidepth sites (Fig. 1).Sediment cores (clear acrylic, 6.5 cm ID×30.6 cm tall) werecollected to a depth of 10 cm; overlying water was removedand replaced with filtered site-specific water to measurebenthic processes only. Site water was filtered through aseries of 142 mm filters: GF/D (2.7 μm), GF/F, and PES.Cores were incubated in site-specific water in clear fiberglasschambers under in situ conditions during July 2010 in pondsbehind the University of North Carolina Institute of MarineStudies and in a temperature- and light-controlled environ-mental chamber at VIMS in April 2011. Shade cloth wasused to simulate in situ irradiance at the sediment surfacemeasured during sampling. Sampling for DIC and nutrientsat dawn, dusk, and dawn was performed as described abovefor the seasonal studies.

Gross Ammonification Rates

To measure sources of NH4+ to the benthos in the NRE, gross

dark AMN rates were determined using the isotope pooldilution technique as described by Anderson et al. (1997)in sediment cores sampled at the same sites and concurrentlywith the metabolism and nutrient flux studies conductedduring both the shallow water seasonal and depth studies.Additional measurements, as described in Maxey (2012),were performed at the six shallow seasonal sites in October2009 and at three of these sites (sites 2, 3, 5) in June,September, and November 2010 and in March 2011. At eachsampling site, five sets of paired polycarbonate cores (5.7 cmID×20 cm tall) were collected to a depth of 10 cm. Follow-ing collection, cores were uncapped and immersed in site-specific water and held overnight in the dark with constantmixing and aeration. Immediately after injecting cores with15N-NH4

+ [3.6 mL of [NH4]2SO4 (30 at.%, 10 mM) through36 silicone holes in each core], sediments in one of the pairedcores of each set were extracted with equal volumes of 2 M

Table 1 Annual mean water quality and sediment characteristics of the six shallow seasonal sites of Jacksonville (site 1), Southwest Creek (site 2),Wallace Creek (site 3), French Creek (site 4), Courthouse Bay (site 5), and Traps Bay (site 6)

Site Salinity CDOMabsorption at440 nm (m−1)

Watercolumn chla (μg L−1)

Kd(m−1)

watercolumnDIN(μM)

watercolumnDIP (μM)

Sedimentorganicmatter (%)

SedimentextractableNH4

+

(mmol m−2)

PorewaterS2−a (mM)

Benthicchl a(mg m−2)

SedimentC/N(molar)

1 9.0 (1.5) 7.1 (0.8) 50.4 (14.1) 3.5 (0.2) 20.5 (5.9) 2.12 (0.33) 18.2 (2.4) 12.7 (3) 2.66 (1.22) 108.8 (14) 19 (0.2)

2 14.2 (1.7) 8.5 (1.5) 16.5 (1.8) 3.5 (0.4) 7.3 (1.8) 0.77 (0.20) 13.8 (1.0) 6.7 (1.2) 3.95 (0.54) 80.1 (10.6) 26.7 (1.2)

3 18.2 (1.3) 4.0 (0.6) 14.4 (1.8) 1.8 (0.2) 2.0 (0.8) 0.16 (0.02) 1.5 (0.4) 3.0 (0.6) 3.43 (1.89) 96.8 (8.0) 11.9 (1.7)

4 21.3 (1.2) 3.5 (0.6) 13.8 (1.2) 1.9 (0.2) 3.5 (0.9) BDb 0.8 (0.2) 3.2 (0.8) 0.22 (0.12) 92 (11.2) 10.8 (1.5)

5 31.9 (0.9) 1.4 (0.3) 8.7 (1.2) 1.5 (0.1) 1.5 (0.3) BD 1.1 (0.1) 2.4 (0.3) 0.07 (0.05) 48.1 (6.4) 11.9 (0.9)

6 32.0 (0.9) 1.5 (0.3) 14.5 (7.6) 1.7 (0.1) 1.3 (0.3) BD 0.8 (0.2) 2.1 (0.3) 0.39 (0.09) 57.3 (7.3) 10.1 (1.1)

Standard errors given in parentheses, n=8 to 21a Porewater sulfide (S−2 ) only measured in winter, spring, and summer 2011, n=3b Below detection limit of 0.16 μM

Estuaries and Coasts

potassium chloride with shaking for 1 h, representing T-zeromeasurements; the extractant was filtered (Whatman PES)and frozen until analyzed. The remaining cores of each of thepaired sets were capped, stirred, and incubated for 24 h in thedark in a temperature-controlled incubator at in situ temper-atures. At the end of the incubation, the sediments wereextracted as described above, representing T-final measure-ments. NH4

+ in extracts was trapped by diffusion onto acid-ified filters as described by Brooks et al. (1989). Previousstudies have shown that trapping efficiencies exceed 90 %.Samples were analyzed for 15N enrichment at the Universityof California Davis Stable Isotope Facility. Rates of grossdark AMN were calculated using a model described byWessel and Tietema (1992).

Nitrogen Fixation Rates

Five sediment cores, collected at each of the six shallowseasonal sites during May, July, and October 2009; March,June, and November 2010; and March and June 2011 and atthe multidepth sites during July 2010 and April 2011, wereanalyzed for NFIX using the acetylene reduction method asdescribed in Whitehead (2012) (Currin et al. 1996), assuminga ratio of 4 mol of acetylene reduced to 1 mol of N2 fixed(Mulholland et al. 2004). From each core, 1-cm-deep sub-samples (5-mL cut Normject syringe, 1.1 cm ID) were col-lected and incubated aerobically in the light and anaerobicallyin the dark for 6 h in 60-mL serum bottles amended with15 mL of acetylene at in situ water temperatures and lightadjusted to in situ irradiances with shade cloth. Ethylene wasmeasured by flame ionization gas chromatography (HewlettPackard model 5890 fitted with 6 ft Poropak N stainless steelcolumn; oven temperature 80 °C; detector 220 °C).

Denitrification Rates

Net dark N2 measurements were conducted for the seasonalshallow water study in July and October 2008 and May andJuly 2009 and for the depth study at 0.5 and 3 m in July 2010and April 2011. N2 was determined by analysis of N2/Ar usingmembrane inlet mass spectrometry (MIMS) as described byPiehler and Smyth (2011). The N2/Ar technique measures netN2 production since NFIX occurring concurrently with DNFmay offset some of the N2 produced by DNF. A positive net N2

flux indicates that DNF dominated the net N2 flux (DNFnet).Gross DNF (DNFg) was calculated by adding DNFnet to NFIX,measured concurrently. Triplicate sediment cores (6.8 cm ID),collected to a depth of 17 cm with addition of 400 mL ofoverlying water, were held in the dark in an environmentalchamber at ambient temperatures and connected to a 24-channel peristaltic pump (Lavrentyev et al. 2000; Piehler andSmyth 2011), which pulled site-specific water from a reservoircontinuously over the cores at a flow rate of 1 mL min−1 to

provide a well-mixed water column. Cores were pre-incubatedfor no less than 18 h prior to sampling to ensure that the systemwas at steady state (Eyre and Ferguson 2002). Five-mL watersamples were collected from both the outflow port of each coreand the site water inflow line at 18, 24, 36, and 48 h after thestart of incubation. Upon collection, samples were immediate-ly analyzed for N2, O2, and Ar using a Balzers Prisma QME200 quadrupole mass spectrometer (Kana et al. 1994) andconcentrations of N2 and O2 calculated based on their ratioswith Ar (Kana et al. 1998; Ensign et al. 2008). Measured ratesof DNFnet from each core were averaged to yield core specificvalues. To determine daily DNFnet rates, dark hourly rates weremultiplied by hours of dark, thus assuming no DNF in the lightand providing a conservative estimate of DNFnet.

Data Analysis

Preliminary analyses of all data (means, standard errors) werecompleted using Microsoft Excel. R (R Development CoreTeam 2011) was used to perform single and multiple linearand non-linear (e.g., natural log) regressions. Minitab 16(Minitab, Inc., State College, PA, USA) was used to conductanalysis of variance (ANOVA) on site characterization andprocess rate data to determine differences by site and seasonor region, season, and depth. Interactions between all variableswere tested. Levene’s test of homogeneity of variance wasconducted to determine if means had similar variances. If thetest was found to be significant (p<0.05), data were natural logtransformed. Tukey’s test was used to evaluate pair-wise com-parisons after a significant ANOVA; differences were consid-ered significant at p<0.05. When the assumption of homoge-neity of variance could not be met, non-parametric two-wayanalysis of similarity (ANOSIM) was conducted on the data todetermine differences by site and season using PRIMER 6(Primer-E, Inc., Plymouth, UK; Clarke 1993; Clarke and War-wick 2001). Euclidean distance matrices on normalized datawere used to determine a global R statistic for each main effect(site or season) as well as pairwise comparisons, similar to anF-statistic. Using ANOSIM, a permutation procedure (999permutations) was then performed to provide a p value forthe R statistic values. A Bonferroni correction (0.05/n; n =number of pairwise comparisons) was applied to adjust thesignificance level of the multiple pairwise comparisons.

Results

Site Characterization

Water quality data from the six shallow seasonal sites weretypical of a river-dominated estuary with a dominant sourceof fresh water and nutrients (New River) at the head. Salinityincreased whereas water column chl a, DIN, DIP, and Kd all

Estuaries and Coasts

decreased down-estuary (Table 1). Mixing curves (data notshown) demonstrated conservative transport of dissolvedorganic carbon and nitrogen and CDOM down estuary withrapid uptake of NOx and PO4

3− by phytoplankton primaryproduction at the head of the estuary. Average NOx concen-trations in 2008–2009 declined from 17 μM (SE=10) at thehead estuary (station #1) to 0.5 (SE=0.4) at station #3 andremained low throughout the remainder of the estuary.

Sediments similarly showed strong down-estuary gradientswith highest organic content, C/N, extractable NH4

+ (0–5 cmdepth), and porewater sulfide up-estuary declining down-estuary (Table 1). MPB biomass, as represented by bchla,averaged 108 mg m−2 (n=72, SE=7) in shallow subtidalsediments (0.25 m in depth) and 48 mg m−2 (n=88, SE=4)in deeper sediments (2.0 m in depth). The relationship be-tween bchla and depth varied along the estuarine gradient(Fig. 2a). In the upper NRE, bchla was the highest at 0.25-and 0.5-m depths and then sharply declined with depth,whereas at mid-estuary sites bchla was similar at all depthsfrom 0.25 to 1.5 m. At lower estuary sites bchla, averagedacross all stations also showed a decline with depth. Alongthe entire NRE, bchla values generally declined down-estuary across all depths, despite a decrease in Kd, possiblysuggesting nutrient limitation of MPB production (Fig. 2a).The distribution of bchla across the 0.25- to 2.0-m depthgradient varied considerably between stations as did the depthat which peak MPB biomass was found, which ranged from0.25 to 2.0 m (Fig. 2b). Overall, there was a positive rela-tionship between calculated representative wave energy ofeach station and the depth at which peak biomass was located(Fig. 2b), suggesting that at higher energy locations, waveenergy limited benthic microalgal biomass via shear stress atthe sediment surface.

Variations in Photic Area with Freshwater Discharge

Nearshore dataflow surveys along the perimeter of the NREwere used to determine the degree to which freshwaterdischarge with accompanying CDOM and chl a impactedphotic area, computed as the area of bottom receiving at least

1 % of incident irradiance. The benthic photic area of the NREranged from a low of 38 % in September 2011, a high dis-charge period following passage of Hurricane Irene, to a highof 81 % in July 2008, a period of drought. Mean freshwaterdischarge and benthic photic area were significantly relatedwith discharge explaining 69 % of the observed variation inphotic area (p≤0.001; Fig. 3). This linkage between dischargeand photic area extended to MPB biomass; winter FW dis-charge was negatively related to MPB biomass measured inspring, explaining 63 % of the observed variation (p=0.002;Fig. 4) and suggesting a months-long lag for a response inproduction of MPB biomass to photic area in winter.

Relationships Between Shallow Water Metabolismand Nutrient Fluxes: Seasonal Studies

To determine how environmental factors and their interac-tions affect the efficiency of the benthic filter, benthic me-tabolism (GPP, R, NCP), nutrient cycling processes (AMN,DNF, NFIX), and nutrient fluxes were measured at seasonalshallow sites under in situ conditions. Metabolic results arereported in units of C with negative bars for NCPrepresenting net autotrophy and positive bars net heterotro-phy (Fig. 5). Results of data analysis by two-way ANOVAand ANOSIM (by site and season) indicated that benthicGPP was significantly higher in summer and fall than spring,whereas R was the highest during summer (2S and 3S ofonline electronic supplementary material; ESM_1.pdf). Ben-thic NCP was net autotrophic at most sites and for mostseasons except occasionally for up- and mid-estuary sites.Regression of NH4

+ fluxes with benthic NCP showed ahighly significant relationship with NCP explaining 59 %of the observed NH4

+ flux (Fig. 6). Net autotrophic stationswere sinks and net heterotrophic stations sources of NH4

+ tothe water column (Fig. 6); thus, NCP serves as an importantindicator of the effectiveness of the benthic filter. DuringMay 2009 at the organic and sulfide-rich sites 1 and 2,benthic uptake of NO3

− explained 92 % of the flux ofNH4

+ out of sediments to the water column, suggesting theoccurrence of DNRA at those sites (data not shown).

0

20

40

60

80

100

120

140

160

180

0.0 0.5 1.0 1.5 2.0 2.5

Ben

thic

chl

a(m

g m

-2)

Depth (m)

UPMIDLOW

a

y = 0.29e0.0039x

p=0.03

0.0

0.5

1.0

1.5

2.0

2.5

0 100 200 300 400 500

wat

er d

epth

(m

) of

m

axim

um b

enth

ic c

hl a

Site RWE

UPMIDLOW

bFig. 2 Variations in meanbenthic chlorophyll (chl) a (±standard error; n=18 to 40) withwater depth at upper (square),middle (triangle), and lower(circle) regions of the NewRiver Estuary (a). Regression ofdepth of maximum benthic chl aconcentration with siterepresentative wave energycalculated at 1 m water depth(n=19) (b)

Estuaries and Coasts

Environmental Factors Regulating Metabolic and N-Processing Rates

Regression analyses demonstrated that GPP was most sensi-tive to light availability, temperature, and bchla. As shown inFig. 7, GPP was inversely proportional to Kd and the re-siduals directly proportional to both water temperature andbchla. Kd, bchla, and water temperature together explained52 % of the observed variability in GPP. As one wouldexpect, benthic R and AMN were also directly proportionalto water temperature, which explained 34 % of the variationin R and 22 % of the variation in AMN (Table 2).

Results of additional regression analysis of the seasonaldataset for significant relationships and interactions betweenvarious environmental drivers, their r2 values and slopes areshown in Table 2. Kd was negatively related to GPP, whichwas in turn negatively related to the flux of NH4

+ out of thesediment. Increased Kd negatively impacted NFIX but posi-tively affected dark DNFnet, suggesting the potential impor-tance of autotrophic N-fixing bacteria and a residual responseof dark DNFnet to GPP. GPP, R, AMN, and NFIX varieddirectly with temperature; however, since both GPP and Rresponded similarly to increased temperature, there was noclear effect on NCP. Higher temperatures, which increasedboth benthic AMN and NFIX, sources of N, also increasedMPB N-uptake, a major sink of N in shallow systems; thus,there was no significant relationship between temperature andNH4

+ fluxes out of sediments. The biomass of benthic auto-trophs, shown as bchla, was an indicator of net autotrophy(shown as negative NCP) and the effectiveness of the benthicfilter, represented by increased uptake of NH4

+ by the benthos(Table 2). NFIX responded quite differently than DNFnet toenvironmental variables. DNFnet rates demonstrated a signif-icant inverse relationship with salinity but were not related totemperature, whereas NFIX rates were positively related totemperature and were not significantly related to salinity(Fig. 8a; Table 2). NFIX, which can offset removal of N byDNFg and contribute to the DNFnet measured by the MIMSmethod, constituted an increasing percentage of DNFg, withdistance down-estuary (Fig. 8b).

Depth Experiments: Effects of Water Column Depthon Metabolic and N Cycling Rates

In order to better tease out the effects of light and depthwithout the confounding seasonal variability in temperature,freshwater discharge, and nutrient supply, measurements ofmetabolic parameters, N-cycling rates (AMN, DNF, NFIX),and DIN fluxes were made at the multidepth sites (Fig. 1)during two seasons. The percentage of surface irradiancereaching the benthos at these sites varied from approximately0.8 to 62 %. In July 2010 as light availability (% of surfaceirradiance) increased with decreasing water depth (0.5–3.0 m),bchla increased (Fig. 9a) and NCP became increasingly auto-trophic (Fig. 9b) with resultant declines in sediment–waterNH4

+ fluxes (Fig. 9c). Similar patterns were observed in April2011 (data not shown). Regressions of NCP versus NH4

+ fluxindicated a sharp transition. As NCP became net heterotro-phic, NH4

+ fluxed out of sediments; fluxes were linearly anddirectly proportional to the degree of benthic heterotrophy(Fig. 9d). Bchla again served as an indicator of the benthicfilter, explaining 33% of the variation in NCP and 30% of thesediment–water NH4

+ flux. Whereas GPP and NFIX weredirectly related to light availability (with r2 values of 0.63and 0.28, respectively), hourly DNFnet and NH4

+ flux were

Fig. 3 Regression of percentage of the NRE area with >1 % surfaceirradiance (I0) reaching the benthos versus daily mean freshwater dis-charge at the USGS Gum Branch Station (#02093000) (n=13)

Fig. 4 Variation of mean benthic chlorophyll (chl) a (± standard error;n=3), measured in April–June at 0.5 m water depth at three upper NREstations [Jacksonville (site 1) (square), Southwest Creek (site 2) (trian-gle), Wallace Creek (site 3) (circle)] with mean winter daily discharge(January–March) measured at the USGS Gum Branch Station(#02093000) from 2008 to 2011 (n=12)

Estuaries and Coasts

inversely related to light (with r2 values of 0.23 and 0.74,respectively) (Table 3).

Depth Experiments: Effectiveness of the Benthic Filter

The effectiveness of the benthic filter across the estuarinesalinity gradient as a function of water depth (light availability)

was determined during the July 2010 and April 2011 depthexperiments (April data not shown but followed the samepattern as July) by scaling square meter rates to the area ofeach estuarine segment within each depth interval (Table 4).DNFg, calculated by summing the absolute values of NFIXand DNFnet with their propagated errors, was used for thiscalculation rather than DNFnet since AMN includes contribu-tions from NFIX. Whereas MPB N-demand was primarilyresponsible for uptake of ammonified N in shallow waterthroughout the estuary, DNFg became increasingly importantin deeper waters of the upper and mid-estuary. The benthicfilter removed from 15 to 137 % of ammonified N with thehighest rates generally due to MPB uptake in shallow water.

Discussion

The studies described above, conducted during multiple yearswith different weather patterns, at shallow sites during multipleseasons, and across depth intervals within single seasons,allowed us to tease apart the interacting impacts of fresh waterdelivery, light, temperature, salinity, wind, and MPB biomass onbenthic processes. Based upon the relationships discerned duringthis study, we also make predictions of benthic responses to

100

150

200

250

300

P (

mm

ol C

m-2

d-1 )

July 08

Oct 08

May 09

July 09

Mar 10

a

100

150

200

250

atio

n (m

mol

C m

-2d-

1 )

b

3.0

3.5

4.0

1 )80

100

1200

50

100

150

GP

P (

mm

ol C

Mar 10

0

50

100

Res

pira

tion

(mm

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

NH

4+flu

x (m

mol

N m

-2d-

1 )

d

-40

-20

0

20

40

60

80

100

120

NC

P (

mm

ol C

m-2

d-1 )

c

1 2 3 4 5 6-1.5

-1.0

-0.5

0.0

0.5

Sites

NH

4+flu

-80

-60

-40

-20

1 2 3 4 5 6

NC

P (

mm

Sites

Fig. 5 Shallow water benthic gross primary production (a), respiration(b), net community production (c), and NH4

+ flux (d) by date andshallow seasonal site, Jacksonville (site 1), Southwest Creek (site 2),Wallace Creek (site 3), French Creek (site 4), Courthouse Bay (site 5),

and Traps Bay (site 6). Sites are arranged (left to right) from up- todown-estuary. March 2010 measurements were only made at sites 2, 3,and 5. Values are reported as means±1 SE (n=3 to 6)

Fig. 6 Variation of mean benthic NH4+ flux with mean benthic net

community production measured at the six shallow seasonal sites fromJuly 2008 to March 2010 (n=27)

Estuaries and Coasts

variables accompanying climate change including increasedfreshwater delivery and temperature.

Responses to Freshwater Inputs

The increased freshwater discharge accompanying storms, sea-sonal freshets, and high rainfall years with associated delivery

of total suspended solids (TSS), CDOM, and nutrients to shal-low estuaries have contrasting impacts on benthic primaryproduction. The benthos is exposed to increased nutrients butdecreased light availability. In this study, estuarine photic areawas inversely related to freshwater discharge, and MPB bio-mass measured in spring in the upper half of the estuary was thelowest during years with the highest winter freshwater dis-charge. Net heterotrophy of the system, as measured by NCP,was significantly and inversely related to MPB biomass, andGPP was significantly and negatively related to Kd (Table 2).Nutrients entering at the head of the estuary were rapidlyremoved by phytoplankton (Hall et al. 2013), andwater columnnutrients were generally low throughout the remainder of theestuary. Any response to nutrient enrichment in the NRE waslikely confounded by concurrent reduced light availability inthis blackwater system, which is characterized by high concen-trations of CDOM which play the largest role in light attenua-tion (Eq. 1). Although long-term exposure to increased nutrientloads might increase MPB biomass if sufficient light wereavailable (Cook et al. 2007; Hillebrand et al. 2000; Hillebrandand Sommer 1997), in the short term nutrient enrichment is lesslikely to affect MPB metabolism in the upper estuary becauseof high AMN rates and accumulation of NH4

+ in benthic porewater. In an experimental manipulation of sediment cores fromshallow sites throughout the NRE, Maxey (2012) did notobserve a metabolic response by the benthos after addingNH4

+ to concentrations as high as 100 μM to the overlyingwater.

Impacts of Wind

Wind wave energy, calculated in this study using WeMO toestimate representative wave energy (Malhotra and Fonseca2007), altered the distribution of bchla across a depth rangeof 0.25–2 m (MSL). Although we considered a muchnarrower depth range given the strong light attenuation inthe blackwater NRE, our results are similar to those ofForehead and Thompson (2010), who assessed MPB bio-mass, metabolism, nutrient cycling, and fluxes along a depthgradient of 1.5–14 m in Western Australia. In their study, thelowest MPB biomass, net heterotrophy, and efflux of NH4

+

were observed at 1.5 m, the depth exposed to the highestwave energy. At depths >4 m MPB biomass was higher thanat 1.5 m, and the sites were net autotrophic.

Effects of Salinity

Across the entire NRE, average annual salinity varied from 9to 32, while at an individual up-estuary site (site 2) between2007 and 2011, salinity varied from 1 to 26 depending onprecipitation for the period. Since salinity co-varies withmany of the other variables examined in this study, it wasdifficult to identify relationships with many of the individual

2

3

4

5

6ln

(GP

P+

10)

mm

ol C

m-2

d-1 )

a

y = -0.26x + 4.75R² = 0.28, p=0.005

0

1

2

3

0 2 4 6 8

ln(G

PP

(mm

ol C

Kd (PAR; m )

2b

-1

y = 0.04x - 0.82R² = 0.15, p=0.049

-1

0

1

2

Res

idua

ls o

f ln(

GP

P+

10)

(mm

ol C

m-2

d-1 )

b

-2

-1

0 10 20 30 40

Res

idua

l(m

m

water temperature (oC)2

)

y = 0.005x - 0.44R² = 0.20, p=0.019

-1

0

1

2

Res

idua

ls o

f ln(

GP

P+

10)

(mm

ol C

m-2

d-1 )

c

y = 0.005x - 0.44R² = 0.20, p=0.019

-2

-1

0 50 100 150 200 250

Res

Benthic chl a (mg m-2)

Fig. 7 Regressions of mean benthic gross primary production withvertical light attenuation (Kd) (a), residuals after Kd with water temper-ature (b), and residuals after Kd and water temperature with benthicchlorophyll a (bchla) (c), measured at the six shallow seasonal sitesfrom July 2008 to March 2010 (n=27)

Estuaries and Coasts

processes. For example, the influence of salinity on DNFnetis difficult to separate from that of NOx. In the NRE, NOx

was rapidly depleted at the head of the estuary due to uptakeby phytoplankton; regression analysis suggested that salinityplayed a more important role in influencing DNF than didNOx concentration. DNFnet was strongly and negativelyrelated to salinity (Fig. 8a) as has also been observed byothers (Rysgaard et al. 1999; Giblin et al. 2010), whereasNFIX was more weakly and positively related to salinity(Table 2). The efficiency of coupled NIT–DNF at lowersalinities may be explained by the increased availability of

NH4+ due to its greater adsorption to sediment particles

(Seitzinger et al. 1991; Rysgaard et al. 1999; Giblin et al.2010). In addition, porewater sulfide, which can inhibit NITand DNF, is expected to be lower in fresher water; however,in our study, sulfide concentrations were the highest at up-estuary sites (Table 1). The decreased DNFnet observeddown-estuary and during July 2009 in this study may bepartially explained by a larger percentage of DNFg offset byNFIX (Fig. 8b). Whitehead (2012) identified the organismsmost likely to be responsible for fixing N in the NRE assulfate reducing deltaproteobacteria. The role of NFIX in

Table 2 Single regressions for mean benthic metabolic and nitrogencycling rates measured at six shallow seasonal sites from 2008 to 2011with light attenuation (Kd per meter), temperature (in degree Celsius),

benthic chlorophyll a (in milligrams per square meter), salinity, andsediment organic matter content (in percent) as independent factors

GPP R NCPb NH4+ fluxc DNF AMN NFIXe

r2 Slope r2 Slope r2 Slope r2 Slope r2 Slope r2 Slope r2 Slope

Kd 0.28 −0.26a – – – – 0.17 0.28 0.38 0.76 – – 0.27 −0.56

Temperature 0.26 3.88 0.34 4.28 – – – – – – 0.22 0.53 0.39 0.10

Bchla – – – – 0.17 –0.26 0.19 –0.01 – – – – – –

Salinity – – – – – – – – 0.72 −1.17d – – 0.20 0.05

OM – – – – – – – – – – – – 0.18 −0.06f

The dependent factors are GPP, R, and NCP (in millimoles C per square meter per day); ammonium flux, DNF, AMN, NFIX (millimoles N per squaremeter per day). r2 and slope values are provided only for significant regressions (p<0.05). N=27 for GPP, R, NCP, and NH4

+ flux; n=23 for DNF;n=50 for AMN; and n=48 for NFIX

GPP gross primary production, R respiration, NCP net community production,DNF net denitrification, AMN ammonification, NFIXN-fixation,OMorganic matter, bchla benthic chlorophyll a, – non-significant regressiona GPP transformed as ln(x+10)b Positive NCP represents net heterotrophy; negative NCP represents net autotrophyc NH4

+ flux transformed as ln(x+0.8)d Salinity natural log transformede NFIX natural log-transformedf n=36

30

40

50

60

70

% o

f Gro

ss D

NF

May 09

July 09

b

y = 0.11x + 0.96R² = 0.45, p<0.001

3

4

5

6

7

8

et(m

mol

N m

-2d-

1 )

SalinityWC NOx

a

0

10

20

30

40

1 2 3 4 5 6

Nfix

% o

f Gro

s

Site

No data

y = 0.11x + 0.96R² = 0.45, p<0.001

y = -1.17ln(x) + 4.71R² = 0.72, p<0.001

0

1

2

3

4

0 10 20 30 40 50

DN

Fne

t(m

mol

N

Salinity or NOx (µM)

Fig. 8 Regressions of mean net denitrification (DNFnet) with salinityand water column nitrate + nitrite (NOx) measured at the six shallowseasonal sites from July 2008 to July 2009 (n=23) (a). Benthic NFIX(0–1 cm depth horizon) as a percentage of estimated gross DNF

measured at six shallow seasonal sites, Jacksonville (site 1), SouthwestCreek (site 2), Wallace Creek (site 3), French Creek (site 4), CourthouseBay (site 5), and Traps Bay (site 6), in May and July 2009 (b). Sites arearranged (left to right) from up- to down-estuary

Estuaries and Coasts

partially offsetting removal of N2 by DNFg should, therefore,be considered when determining the efficiency of N removalby DNF. Whereas increased storm activity may result inincreased DNF due to freshwater discharge, the coincidentdecrease in light availability will reduce N uptake by MPBand, thus, likely decrease the efficiency of the benthic filter.

Impacts of Light: Benthic Metabolism and N Cycling AcrossDepth Gradients

By constraining environmental variables, including temper-ature, discharge, and nutrient inputs, the depth experimentsperformed during this study provided further information on

y = -1.00x + 26.33R² = 0.66, p<0.001

0 20 40 60 80-80

-60

-40

-20

0

20

40

60

% incident irradiance

NC

P (

mm

ol C

m-2

d-1)

b

y = -0.46ln(x) + 1.42R² = 0.74, p<0.001

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0 20 40 60 80

NH

4+flu

x (m

mol

N m

-2d-1

)

% incident irradiance

c

y = 2.60x + 10.06R² = 0.63, p<0.001

0

50

100

150

200

250

300

0 20 40 60 80

Ben

thic

chl

a(m

g m

-2)

% incident irradiance

UPMIDLOW

a

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

-80 -60 -40 -20 0 20 40 60

NH

4+flu

x (m

mol

N m

-2d-1

)

NCP (mmol C m-2 d-1)

d

Fig. 9 Regressions of benthicchlorophyll a (a), benthic netcommunity production (NCP)(b), and benthic NH4

+ flux vs.percent of surface irradiance (c)and plot of benthic NH4

+ fluxversus benthic NCP (d) formultidepth stations in the upper(square), middle (triangle), andlower (circle) regions of theNew River estuary, measured inJuly 2010 (n=26)

Table 3 Single regressions for mean benthic metabolic and nitrogencycling rates measured during the depth experiment in July 2010 withpercent surface irradiance reaching the sediment surface, benthic

chlorophyll a (in milligrams per square meter), salinity, and sedimentorganic matter content (in percent) as independent factors

GPP R NCPb NH4+ flux DNF NFIXe

r2 Slope r2 Slope r2 Slope r2 Slope r2 Slope r2 Slope

% Surface irradiance 0.63 0.03a – – 0.66 −1.00 0.74 −0.46c 0.28 −1.05d 0.26 0.03

Bchl 0.57 0.35 0.17 0.13 0.33 −0.21 0.30 −0.48c – – 0.42 0.01

Salinity – – 0.23 −1.35 0.19 −1.43 – – – – – –

% OM – – – – 0.34 2.13 0.35 0.07 – – 0.13 −0.25c

The dependent factors are GPP, R, and NCP (in millimoles C per square meter per day); ammonium flux (in millimoles N per square meter per day);DNF (micromoles N per square meter per hour); NFIX (in millimoles N per square meter per day). r2 and slope values are provided only forsignificant regressions (p<0.05). N=26 for GPP, R, NCP, and NH4

+ flux; n=18 for DNF; and n=43 for NFIX

GPP gross primary production, R respiration, NCP net community production, DNF dark net denitrification, NFIX N-fixation, OM organic matter,bchla benthic chlorophyll a, – non-significant regressiona GPP transformed as ln(x+10)b Positive NCP represents net heterotrophy; negative NCP represents net autotrophyc Independent factor natural log-transformedd% surface irradiance data used in this regression were measured in the field when cores were collected because DNF cores were incubated in thedarke NH4

+ flux transformed as ln(x+1)

Estuaries and Coasts

the effects of light on benthic metabolism and N uptake.Results showed strong positive and significant relationshipsbetween % surface irradiance, related to water depth, andGPP, MPB biomass, and NFIX, and negative relationshipsbetween % surface irradiance and NCP (net heterotrophy),hourly dark DNFnet, and NH4

+ flux. The negative relation-ship between DNFnet, measured in the dark and % surfaceirradiance, suggests a residual response to prior light on thesurface benthos or the possibility of continued competitionfor nutrients due to dark uptake by MPB as observed by Joyeet al. (2003) and Dalsgaard (2003). Similar to our observa-tions, Sundbäck et al. (2004) and Ferguson et al. (2007)reported that greater light availability increased benthic au-totrophy with greater assimilation of N, decreased DNF, and,thus, decreased fluxes of DIN from sediments to the watercolumn. Although net benthic autotrophy might be expectedto increase DNF by producing labile organic carbon (Cooket al. 2007) and DO, which can support coupled NIT–DNF,the MPB also compete with NIT for NH4

+. Results of thisstudy and others (Rysgaard et al. 1995; Sundbäck and Miles2000; Eyre and Ferguson 2005) suggest that in most cases,benthic autotrophs outcompete denitrifiers for resources,provided that sufficient light is available. As observed in thisstudy, benthic autotrophy also provides an OM subsidysupporting R through either turnover of biomass or decom-position of dissolved OM exuded during grazing or as extra-cellular polymeric substances (Eyre and Ferguson 2005;Cook et al. 2007); however, although R was significantlyand positively related to bchla, it was not significantly relat-ed to light availability (Table 3).

Temperature Effects

In this present study, temperature explained 26, 34, 22, and39 % of the variability observed in GPP, R, AMN, and NFIX,respectively (Table 2). R increased more rapidly than GPP

with increasing temperature, suggesting that the benthos islikely to become more heterotrophic as water temperatureswarm supporting allometric predictions from the metabolictheory of ecology (Harris et al. 2006). Similar to observa-tions in this study, Alsterberg et al. (2011, 2012) demonstrat-ed that a 4 °C increase in temperature had a greater effect onheterotrophic than autotrophic variables in a shallow sedi-ment system; however, their system remained net autotro-phic in spring and only weakly heterotrophic in fall, whenwarming increased N mineralization, benthic fluxes ofNH4

+, and potential denitrification. Hopkinson and Smith(2005) in their review of benthic respiration across a widevariety of systems noted that OM inputs explained much ofthe variance in R across systems on an annual basis whiletemperature was a more important predictor on seasonal timescales. In shallow systems with inputs of OM to the benthosfrom deposition of phytoplankton blooms (Hall et al. 2013)and high rates of benthic GPP, as observed in this study, wepredict increasing net benthic heterotrophy as temperatureswarm. In shallow systems, NCP is an important predictor ofthe efflux of NH4

+ from sediments (Figs. 6 and 9d). Weexpect that increased heterotrophy will be accompanied byhypoxia/anoxia and sulfide accumulation in the sedimentprofile with increased fluxes of NH4

+ to the water column,the magnitude of which will depend on inhibition of coupledNIT–DNF by the combination of sulfide and anoxia andincreased NFIX by sulfate reducers. With regard to theinfluence of temperature on DNF, our results in the NREare in contrast to those reported by Veraart et al. (2011), whopredicted a significant increase of DNF with temperature duemainly to anoxia resulting from decreased oxygen solubilityand increased R. While their results would apply to a systemwith high water column NO3

−, removed primarily by directDNF, in systems like the NRE with low nitrate increasedtemperature with accompanying low DO would more likelyreduce DNF due to NIT inhibition and substrate limitation.

Table 4 Benthic N sources (AMN), sinks (MPB N-demand, DNF) (in kilomoles N per day) and percentage of AMN removed by the benthic sinksby region and depth contour of the New River Estuary in July 2010 depth experiment

Region Depth Gross AMN MPB N-Demand Gross DNF % Removed

Up 0.5 25 (13) −26 (5) −8 (1) 137

1.5 50 (18) −8 (3) −6 (1) 29

3 209 (102) −3 (3) −28 (3) 15

Mid 0.5 73 (19) −17 (5) −15 (1) 43

1.5 90 (25) −9 (4) −12 (1) 24

3 160 (81) −3 (2) −75 (4) 49

Low 0.5 17 (3) −17 (2) −3 (0) 112

1.5 35 (3) −9 (1) −7 (2) 43

3 7 (3) −1 (1) −2 (0) 48

Standard errors given in parentheses, n=3 for MPB N-Demand and Gross DNF and n=5 for AMN

N nitrogen, AMN ammonification, MPB microphytobenthic, DNF gross denitrification

Estuaries and Coasts

Effectiveness of the Benthic Filter

Over the depth range 0.5 to >3m, the benthic filter removed orsequestered from 15 to 137 %, 24 to 49%, and 43 to 112 % ofammonified N in the upper, middle, and lower estuary, re-spectively (Table 4). As noted below, these conservativevalues may underestimate both sequestration of N by MPBand removal by denitrification. The % of N sequestered byMPB ranged from 5 to 75%with the largest uptake in shallowwater (0.5-mMSL). Whereas we calculated MPB uptake of Nbased on the assumption that 50 % of the carbon fixed wassubsequently exuded, Cook et al. (2009) using inverse model-ing analysis determined that 20 and 30 % was lost underheterotrophic and autotrophic conditions, respectively. GrossDNF, which was measured in this study only in the dark andlikely underestimated daily rates, was responsible for remov-ing from 15 to 30 % of ammonified N along the same depthgradient in the NRE over a temperature range of 25–31 °C;however, removal was very similar to the 22 % observed in astudy of Australian estuaries by Eyre and Ferguson (2005)across a temperature range of 15–29 °C and the 20 % ofremineralized N observed by Sundbäck et al. (2004) on thewest coast of Sweden over a temperature range of 7–14 °C.The similarity in rates of removal of remineralized N by DNFover a wide range of temperatures and water depths suggeststhat, as observed in this study, variables such as quality andquantity of OM, light availability, porewater sulfide, andsalinity play a more important role than temperature in regu-lating DNF across shallow coastal systems. In NarragansettBay sediments amended with OM, simulating deposition ofphytoplankton, and incubated over a 200-d period, Fulweileret al. (2013) observed that rates of net DNF were not directlyrelated to temperature but were related to both the quality andquantity of OM deposited to sediments. Time since depositionof OM corresponded to both a decline in DNF rates and anincrease in expression of nifH, representing the activity ofnitrogen fixing bacteria. In the Narragansett Bay study, as inthe study by Whitehead (2012) in the NRE, bacteria express-ing nifH were primarily sulfate reducers. Fulweiler et al.(2013) hypothesized that nitrogen fixing sulfate reducers canoutcompete denitrifiers for recalcitrant OM and in the processproduce sulfide, which inhibits both NIT and DNF (Joye andHollibaugh 1995; Joye and Anderson 2008). In the NRE,DNF rates were the highest and NFIX the lowest in the upperestuary where we noted both the highest concentrations ofsediment OM but also the highest accumulations of sulfide inpore water. The percent of DNFg offset by NFIX was thehighest down-estuary during July. Whitehead (2012) furthershowed that NFIX rates across the NREwere significantly andinversely related to rates of AMN and in the presence ofmolybdate, an inhibitor of sulfate reduction, were reducedby 44–83 %, further supporting the role of sulfate reducersas nitrogen fixers in estuaries. However, in the NRE, the role

of degraded OM and sulfide inhibition in regulating DNFremains uncertain.

Predicted Responses of Benthic Processes to ClimateChange Variables

Predictions of the impacts of climate change on mid-Atlanticcoastal areas of the USA suggest that temperature increasesranging from 3 to 6 °C by 2070–2099 and increased precip-itation and intensity of storm events, accompanied by greaterwind and river flow especially during winter and spring(Hayhoe et al. 2007; IPCC (International Panel on ClimateChange) 2007; Allan and Soden 2008; Najjar et al. 2009,2010). A trend of long-term warming in the Chesapeake Bay

Fig. 10 Cascade of benthic responses to weather- and climate-relatedphysical–chemical drivers: proximate variables (downward triangles),metabolic processes (circles), sediment properties (upward triangles),nitrogen cycling processes (hexagons). (→) increases; (⇢) decreases.Both solid and dashed lines connect net community production (NCP)to dissolved inorganic nitrogen (DIN) since negative NCP (net autotro-phy) decreases DIN and positive NCP (net heterotrophy) increasesDIN. DIN and NO3

−/NH4+ located separately to reduce diagram

complexity

Estuaries and Coasts

has been observed with temperatures in the 1990s approxi-mately 1 °C warmer than in the 1960s (Najjar et al. 2010).

We expect that the shallow photic systems that are com-mon along the US Atlantic and Gulf coasts will be particu-larly vulnerable to the impacts of climate change because ofthe importance of benthic processes and the benthic filter tomediation of nutrient enrichment in these systems. In Fig. 10,we conceptualize a predicted cascade of responses to theinteracting drivers expected to accompany climate change.We predict that in response to climate change:

& Increased freshwater discharge with accompanying loadsof TSS, CDOM, and nutrients will attenuate light andsupport phytoplankton blooms reducing the availabilityof light to the benthos. DNF may increase due to de-creased competition for nutrients fromMPB and reducedinhibition by sulfide.

& Bchla and benthic GPP will decrease in response toreduced light.

& Shallow water sediments exposed to high shear stressdue to elevated wind energy will demonstrate reducedbchla resulting in decreased benthic primary production.

& Benthic R will increase faster than GPP in response toclimate warming.

& Benthic NCP will become increasingly net heterotrophicresulting in increased fluxes of NH4

+ to the water col-umn, further supporting pelagic primary production andincreasing shading of the benthos.

& Sediments will become increasingly anoxic reducingNIT and accumulate sulfide due to anaerobic respiration,further inhibiting NIT.

& Although increasing temperature may increase DNF insystems where nitrate is not limiting, it will decreaseDNF in systems where nitrate is supplied primarily byNIT.

& Increased anoxia and sulfide will result in increasedDNRA relative to DNF with more NH4

+ accumulationand efflux to the water column; the contribution of NFIXto gross DNF will increase offsetting removal of N as N2.

The cascading set of responses summarized above islikely to reduce the effectiveness of the benthic filter withfeedbacks that will increase effluxes of N to the water col-umn. This increased release of bioavailable N has the poten-tial to contribute to the eutrophication of these systems withconcomitant decreases in overlying water quality. Thesechanges also have the potential to push coastal systems pasta tipping point where trophic status moves away from netautotrophy and towards net heterotrophy. Taken together,these non-linear feedbacks are expected to accelerate degra-dation of shallow systems as climate change proceeds, withthe potential for regime shifts to more degraded stable statesand/or shifting baselines, limiting the potential for nutrient

abatement to reverse the effects of cultural eutrophication(Scheffer et al. 2001; Duarte et al. 2009).

Acknowledgments This work would not have been possible withoutthe assistance of H. Walker, S. Lake, J. Giordano, H. Wiseman, A.Hardison, K. Stark, B. Lawson, L. Ott, and G. Secrist of VIMS; A.Hilting, M. Greene, J. Wernly, P. Delano, and L. Cowart of NOAA; andR. Swartz of UNC-IMS. We are also thankful to S. Cohen of NAVFACand the Camp Lejeune Environmental Management Division staff.Funding for this research was provided by the Strategic EnvironmentalResearch and Developmental Program—Defense Coastal/Estuarine Re-search Program Project SI-1413 and National Science Foundation Pro-jects DEB-0542645 and DGE-0840804. The views expressed are thoseof the authors and do not represent the policies or opinions of the USDepartment of Defense, Department of Commerce, NOAA, or associ-ated services. This paper is Contribution No. 3289 of the VirginiaInstitute of Marine Science, The College of William and Mary.

References

Allan, R.P., and B.J. Soden. 2008. Atmospheric warming and the amplifi-cation of precipitation extremes. Science 321: 1481–1484.

Alsterberg, C., S. Hulth, and K. Sundbäck. 2011. Response of a shallow-water sediment system to warming. Limnology and Oceanography56: 2147–2160.

Alsterberg, C., K. Sundbäck, and S. Hulth. 2012. Functioning of a shallow-water sediment system during experimental warming and nutrientenrichment.PLoSOne 7: e51503. doi:10.1371/journal.pone.0051503.

An, S., and S.B. Joye. 2001. Enhancement of coupled denitrification bybenthic photosynthesis in shallow subtidal estuarine sediments.Limnology and Oceanography 46: 62–74.

Anderson, I.C., C.R. Tobias, B.B. Neikirk, and R.L. Wetzel. 1997.Development of a process-based mass balance model for a Virgin-ia Spartina alterniflora salt marsh: implications for net DIN flux.Marine Ecology Progress Series 159: 13–27.

Anderson, I.C., K.J. McGlathery, and A.C. Tyler. 2003. Microbial medi-ation of “reactive” nitrogen in a temperate lagoon. Marine EcologyProgress Series 246: 73–84.

Anderson, I.C., J.W. Stanhope, A.K. Hardison, and K.J. McGlathery.2010. Sources and fates of nitrogen in Virginia coastal bays. InCoastal lagoons: critical habitats of environmental change, ed. M.Kennish and H. Paerl, 43–72. Boca Raton: CRC.

NOAA (National Oceanic and Atmospheric Administration), 2011. NOAACoastal Geospatial Data project. http://coastalgeospatial.noaa.gov/data_gis.html.

Borum, J., and K. Sand-Jensen. 1996. Is total primary production inshallow coastal marine waters stimulated by nitrogen loading? Oikos76: 406–410.

Bricker, S.B., C.G. Clement, D.E. Pihalla, S.P. Orlano, and D.R.G. Farrow.1999. National estuarine eutrophication assessment: effects of nutrientenrichment in the nation’s estuaries. Silver Spring: National Oceanicand Atmospheric Administration, National Ocean Service, Special Pro-jects Office and the National Centers for Coastal Ocean Science.

Brooks, P.D., J.M. Stark, B.B. McInteer, and T. Preston. 1989. Diffusionmethod to prepare soil extracts for automated nitrogen-15 analysis.Proceedings Soil Science Society of America 53: 1707–1711.

Clarke, K.R. 1993. Non-parametric multivariate analyses of changes incommunity structure. Australian Journal of Ecology 18: 117–143.

Clarke, K.R., and R.M. Warwick. 2001. Change in marine communi-ties: an approach to statistical analysis and interpretation (2ndedition). Plymouth: PRIMER-E.

Cline, J.D. 1969. Spectrofluorometric determination of hydrogen sul-fide in natural waters. Limnology and Oceanography 14: 454–458.

Estuaries and Coasts

Cloern, J.E. 1987. Turbidity as a control on phytoplankton biomass andproductivity in estuaries. Continental Shelf Research 7: 1367–1381.

Cook, P.L.M., B. Veuger, S. Böer, and J.J. Middelburg. 2007. Effect ofnutrient availability on carbon and nitrogen incorporation andflows through benthic algae and bacteria in near-shore sandysediment. Aquatic Microbial Ecology 49: 165–180.

Cook, P.L.M., D. Van Oevelen, K. Soetaert, and J. Middelburg. 2009.Carbon and nitrogen cycling on intertidal mudflats of a temperateAustralian estuary. IV. Inverse model analysis and synthesis. Ma-rine Ecology Progress Series 394: 35–48.

Currin, C.A., S.B. Joye, and H.W. Paerl. 1996. Diel rates of N-2-fixation and denitrification in a transplanted Spartina alternifloramarsh: implications for N-flux dynamics. Estuarine, Coastal andShelf Science 42: 597–616.

Dalsgaard, T. 2003. Benthic primary production and nutrient cycling insediments with benthic microalgae and transient accumulation ofmacroalgae. Limnology and Oceanography 48: 2138–2150.

de Jonge, V.N., W.F. de Boer, D.J. de Jong, and V.S. Brauer. 2012. Long-term mean annual microphytobenthos chlorophyll a variation cor-relates with air temperature. Marine Ecology Progress Series 468:43–56.

Duarte, C.M., D.J. Conley, J. Carstensen, and M. Sánchez-Camacho.2009. Return to Neverland: shifting baselines affect eutrophicationrestoration targets. Estuaries and Coasts 32: 29–36.

Durr, H.H., G.G. Laruelle, C.M. van Kempen, C.P. Slomp, M.Meybeck, and H. Middelkoop. 2011. Worldwide typology ofnearshore coastal systems: defining estuarine filter of river inputsto the oceans. Estuaries and Coasts 34: 441–458.

Engelsen, A., S. Hulth, L. Pihl, and K. Sundbäck. 2008. Benthic trophicstatus and nutrient fluxes in shallow-water sediments. Estuarine,Coastal and Shelf Science 78: 783–795.

Ensign, S.H., J.N. Halls, and M.A. Mallin. 2004. Application of digitalbathymetry data in an analysis of flushing times of two largeestuaries. Computers and Geosciences 30: 501–511.

Ensign, S., M. Piehler, and M. Doyle. 2008. Riparian zone denitrifica-tion affects nitrogen flux through a tidal freshwater river. Biogeo-chemistry 91: 133–150.

Eyre, B.D., and A.J.P. Ferguson. 2002. Comparison of carbon produc-tion and decomposition, benthic nutrient fluxes and denitrificationin seagrass, phytoplankton, benthic microalgae- and macroalgae-dominated warm-temperate Australian lagoons. Marine EcologyProgress Series 229: 43–59.

Eyre, B.D., and A.J.P. Ferguson. 2005. Benthic metabolism and nitro-gen cycling in a subtropical east Australian estuary Brunswick:temporal variability and controlling factors. Limnology andOceanography 50: 81–96.

Eyre, B.D., A.J.P. Ferguson, A. Webb, D. Maher, and J.M. Oakes. 2011.Denitrification, N-fixation and nitrogen and phosphorus fluxes indifferent benthic habitats and their contribution to the nitrogen andphosphorus budgets of a shallow oligotrophic sub-tropical coastalsystem (southern Moreton Bay, Australia). Biogeochemistry 102:111–133.

Ferguson, A., B. Eyre, J. Gay, N. Emtage, and L. Brooks. 2007. Benthicmetabolism and nitrogen cycling in a sub-tropical coastal embay-ment: spatial and seasonal variation and controlling factors. Aquat-ic Microbial Ecology 48: 175–195.

Forehead, H.I., and P.A. Thompson. 2010. Microbial communities ofsubtidal shallow sandy sediments change with depth and wavedisturbance, but nutrient exchanges remain similar. Marine Ecol-ogy Progress Series 414: 11–26.

Fulweiler, R.W., S.W. Nixon, B.A. Buckley, and S.L. Granger. 2007.Reversal of the net dinitrogen gas flux in coastal marine sediments.Nature 448: 180–182.

Fulweiler, R.W., S.M. Brown, S.W. Nixon, and B.D. Jenkins. 2013.Evidence and a conceptual model for the co-occurrence of nitrogen

fixation and denitrification in heterotrophic marine sediments. Ma-rine Ecological Progress Series. doi:10.3354/meps10240.

Giblin, A.E., N.B. Weston, G.T. Banta, J. Tucker, and C.S. Hopkinson.2010. The effects of salinity on nitrogen losses from an oligohalineestuarine sediment. Estuaries and Coasts 33: 1054–1068.

Hall, N.S.,H.W. Paerl, B.L. Peierls, A.C.Whipple, andK.L.Rossignol. 2013.Effects of climatic variability on phytoplankton biomass and communitystructure in the eutrophic, microtidal, New River Estuary, North Caro-lina, USA. Estuarine, Coastal and Shelf Science 117: 70–82.

Hardison, A., I.C. Anderson, E.A. Canuel, C. Tobias, and B. Veuger.2011. Carbon and nitrogen dynamics in shallow photic systems:interactions between macroalgae, microalgae, and bacteria. Lim-nology and Oceanography 56: 1489–1503.

Harris, L.A., C.M. Duarte, and S.W. Nixon. 2006. Allometric laws andprediction in estuarine and coastal ecology. Estuaries and Coasts29: 340–344.

Hayhoe, K., C.P. Wake, T.G. Huntington, L.F. Luo, M.D. Schwartz, J.Sheffield, E.Wood, B. Anderson, J. Bradbury, A. DeGaetano, T.J. Troy,andD.Wolfe. 2007. Past and future changes in climate and hydrologicalindicators in the US Northeast. Climate Dynamics 28: 381–407.

Hillebrand, H., and U. Sommer. 1997. Response of epilithicmicrophytobenthos of the Western Baltic Sea to in situ experimentswith nutrient enrichment.Marine Ecology Progress Series 160: 35–46.

Hillebrand, H., B. Worm, and H.K. Lotze. 2000. Marine microbenthiccommunity structure regulated by nitrogen loading and herbivorepressure. Marine Ecology Progress Series 204: 27–38.

Hopkinson, C.S., and E.M. Smith. 2005. Estuarine respiration: anoverview of benthic, pelagic and whole system respiration. InRespiration in aquatic ecosystems, ed. P.A. del Giorgio and PJleBWilliams, 122–146. USA: Oxford University Press.

IPCC (International Panel on Climate Change). 2007. Summary forpolicymakers. In Climate Change 2007: the physical science basis.Contribution ofWorkingGroup I to the Fourth Assessment Report ofthe Intergovernmental Panel on Climate Change, ed. S. Solomon,D. Qin, M. Manning, Z. Chen,M.Marquis, K.B. Averyt, M. Tignor,and H.L. Miller. Cambridge: Cambridge University Press.

Joye, S.B., and I.C.Anderson. 2008.Nitrogen cycling in coastal sediments. InNitrogen in the marine environment, ed. D.G. Capone, D.A. Bronk,M.R. Mulholland, and E.J. Carpenter, 867–915. New York: Academic.

Joye, S.B., and J.T. Hollibaugh. 1995. Influence of sulfide inhibition ofnitrification on nitrogen regeneration in sediments. Science 270:623–625.

Joye, S, W. Porubsky, N. Weston, and R. Lee. 2003. Benthic microalgalproduction and nutrient dynamics in intertidal sediments. In: Bio-geochemistry of tidal flats. Edited by J. Rullkötter, 67–70.Forschungszentrum Terramare Berichte Nr. 12, Proceedings of aWorkshop held at the Hanse Institute of Advanced Study, Delmen-horst, Germany.

Kana, T.M., C. Darkangelo, M. Hunt, J. Oldham, G. Bennett, and J.Cornwell. 1994. Membrane inlet mass-spectrometer for rapidhigh-precision determination of N2, O2, and Ar in environmentalwater samples. Analytical Chemistry 66: 4166–4170.

Kana, T.M., M. Sullivan, J. Cornwell, and K. Groszkowski. 1998. Deni-trification in estuarine sediments determined by membrane inletmass spectrometry. Limnology and Oceanography 43: 334–339.

Kirk, J.T.O. 1994. Light and photosynthesis in aquatic ecosystems, 2nded. Cambridge: Cambridge University Press.

Knepel, K., and K. Bogren. 2001. Determination of orthophosphate byflow injection analysis. QuikChem Method 31-115-01-1-H.Lachat Instruments, Milwaukee, WI, USA.

Koroleff, F. 1983. Total and organic nitrogen. In Methods of seawateranalysis, ed. K. Grasshoff, M. Ehrhardt, and K. Kremling, 162–169. Weinheim: Verlag-Chemie.

Kromkamp, J.C., J.F.C. de Brouwer, G.F. Blanchard, R.M. Forster, and V.Créach (Eds). 2006. Functioning of microphytobenthos in estuaries.In Proceedings of the Colloquium, Amsterdam, the Netherlands.

Estuaries and Coasts

August 21–23, 2003. Amsterdam: Royal Netherlands Academy ofArts and Sciences.

Lavrentyev, P., W. Gardner, and L. Yang. 2000. Effects of the zebramussel on nitrogen dynamics and the microbial community at thesediment–water interface. Aquatic Microbial Ecology 21: 187–194.

Liao, N. 2001. Determination of ammonia in brackish or seawater byflow injection analysis. QuikChem Method 31-107-06-1-B. Mil-waukee: Lachat Instruments.

Lorenzen, C. 1967. Determination of chlorophyll and phaeopigments: spec-trophotometric equations. Limnology and Oceanography 12: 343–346.

MacIntyre, H.L., R.J. Geider, and D.C.Miller. 1996. Microphytobenthos:the ecological role of the “secret garden” of unvegetated, shallow-water marine habitats. I. Distribution, abundance and primary pro-duction. Estuaries 19: 186–201.

Madden, C., and J. Day. 1992. An instrument system for high-speedmapping of chlorophyll a and physico-chemical variables in sur-face waters. Estuaries 15: 421–427.

Malhotra, A., and M. S. Fonseca. 2007. Wave exposure model (WEMo):formulation, procedures and validation. National Oceanic and At-mospheric Administration Technical Memorandum. National OceanService, National Centers for Coastal Ocean Science, number 65.

Mallin, M.A., M.R. McIver, H.A. Wells, D.C. Parsons, and V.L. John-son. 2005. Reversal of eutrophication following sewage treatmentupgrades in the New River Estuary, NC. Estuaries 28: 750–760.

Maxey, J.D. 2012. Shedding light on the estuarine coastal filter: Therelative importance of benthic microalgae in shallow photic systems.M.S. thesis. College of William and Mary, Williamsburg, VA.

McGlathery, K.J., K. Sundbäck, and I.C. Anderson. 2007. Eutrophica-tion in shallow coastal bays and lagoons: the role of plants in thecoastal filter. Marine Ecology Progress Series 348: 1–18.

Mulholland, M.R., D.A. Bronk, and D.G. Capone DG. 2004. Dinitrogenfixation and release of ammonium and dissolved organic nitrogen byTrichodesmium IMS101. Aquatic Microbial Ecology 37: 85–94.

Najjar, R.G., L. Patterson, and S. Graham. 2009. Climate simulations ofmajor estuarine watersheds in the Mid-Atlantic region of theUnited States. Climatic Change 95: 139–168.

Najjar, R.G., C.R. Pyke, M.B. Adams, D. Breitburg, C. Hershner, M.Kemp, R. Howarth, M.R. Mulholland, M. Paolisso, D. Secor, K.Sellner, D. Wardrop, and R. Woodm. 2010. Potential climate-change impacts on the Chesapeake Bay. Estuarine, Coastal andShelf Science 86: 1–20. doi:10.1016/j.ecss.2009.09.026.

Neubauer, S.C., and I.C. Anderson. 2003. Transport of dissolved inorganiccarbon from a tidal freshwater marsh to the York and Pamunkey riverestuary. Limnology and Oceanography 48: 299–307.

Nixon, S.W., B.A. Buckley, S.L. Granger, and J. Bintz. 2001. Re-sponses of very shallow marine ecosystems to nutrient enrichment.Human and Ecological Risk Assessment 7: 1457–1481.

O’Connor, M.I., M.F. Piehler, D.M. Leech, A. Anton, and J.F. Bruno.2009. Warming and resource availability shift food web structureand metabolism. PLoS Biology 7(8): e1000178. doi:10.1371/journal.pbio.1000178.

Piehler, M.F., and A. Smyth. 2011. Habitat-specific distinctions inestuarine denitrification affect both ecosystem function and ser-vices. Ecosphere 2: 1–16.

Pinckney, J., R. Papa, and R. Zingmark. 1994. Comparison of high-performance liquid chromatographic, spectrophotometric, and fluo-rometric methods for determining chlorophyll a concentrations inestuarine sediments. Journal of Microbiological Methods 19: 59–66.

R Development Core Team. 2011. R: a language and environment forstatistical computing. Vienna: R Foundation for Statistical Com-puting. ISBN 3-900051-07-0, http://www.R-project.org/.

Rysgaard, S., P.B. Christensen, and L.P. Nielsen. 1995. Seasonal vari-ation in nitrification and denitrification in estuarine sediment col-onized by benthic microalgae and bioturbating infauna. MarineEcology Progress Series 126: 111–121.

Rysgaard, S., P. Thastum, T. Dalsgaard, P.B. Christensen, and N.P. Sloth.1999. Effects of salinity on NH4+ adsorption capacity, nitrification, anddenitrification in Danish estuarine sediments. Estuaries 22: 21–30.

Scheffer, M., S. Carpenter, J.A. Foley, C. Folke, and B. Walker. 2001.Catastrophic shifts in ecosystems. Nature 413: 591–596.

Seitzinger, S.P., W.S. Gardner, and A.K. Spratt. 1991. The effect ofsalinity on ammonium sorption in aquatic sediments: implicationsfor benthic nutrient cycling. Estuaries 14: 167–174.

Smith, P., and K. Bogren. 2001. Determination of nitrate and/or nitrite inbrackish or seawater by flow injection analysis colorimetry.QuikChemMethod 31-107-04-1-E. Milwaukee: Lachat Instruments.

Sundbäck, K., and A. Miles. 2000. Balance between denitrification andmicroalgal incorporation of nitrogen in microtidal sediments, NEKattegat. Aquatic Microbial Ecology 22: 291–300.

Sundbäck, K., V. Enoksson, W. Graneli, and K. Pettersson. 1991. Influ-ence of sublittoral microphytobenthos on the oxygen and nutrientflux between sediment and water: a laboratory continuous-flowstudy. Marine Ecology Progress Series 74: 263–279.

Sundbäck, K., A. Miles, and E. Göransson. 2000. Nitrogen fluxes,denitrification and the role of microphytobenthos in microtidalshallow-water sediments: an annual study. Marine Ecology Prog-ress Series 200: 59–76.

Sundbäck, K., F. Linares, F. Larson, A. Wulff, and A. Engelsen. 2004.Benthic nitrogen fluxes along a depth gradient in a microtidalfjord: the role of denitrification and microphytobenthos. Limnolo-gy and Oceanography 49: 1095–1107.

Underwood, G.J.C., and D.J. Smith. 1998. Predicting epipelic diatomexopolymer concentrations in intertidal sediments from sedimentchlorophyll a. Microbial Ecology 35: 116–125.

Veraart, A.J., J.J.M. de Klein, and M. Scheffer. 2011. Warming can boostdenitrification disproportionately due to altered oxygen dynamics.PLoS One 6(3): e18508. doi:10.1371/journal.pone.0018508.

Wessel, W.W., and A. Tietema. 1992. Calculating gross N transforma-tion rates of 15N pool dilution experiments with acid forest litter:analytical and numerical approaches. Soil Biology and Biochem-istry 24: 931–942.

Whitehead, M. 2012. The role of benthic nitrogen fixation as a source ofnew nitrogen to the New River Estuary, North Carolina. Masters ofMarine Science, College of William and Mary, Williamsburg, VA.

Wolfstein, K., J.F.C. de Brouwer, and L.J. Stal. 2002. Biochemicalpartitioning of photosynthetically fixed carbon by benthic diatomsduring short-term incubations at different irradiances. MarineEcology Progress Series 245: 21–31.

Estuaries and Coasts