Vertical activity distribution of dissimilatory nitrate reduction in coastal marine sediment

Introduction Conclusions References


Introduction
The balance between retention and loss of fixed nitrogen, especially NO − 3 , in coastal marine ecosystems is crucial for their predisposition to eutrophication (Burgin and Hamilton, 2007;Herbert, 1999;King and Nedwell, 1985).Sediments play a key role in the biological turnover of fixed nitrogen in shallow aquatic environments by hosting microbially mediated processes such as nitrification, anaerobic ammonia oxidation Figures (anammox), and denitrification (Thamdrup and Dalsgaard, 2008).Among these processes, anammox and denitrification have the potential to contribute to fixed nitrogen removal because they convert fixed nitrogen into dinitrogen that can leave the ecosystem.A third anaerobic process involved in fixed nitrogen conversion is dissimilatory nitrate reduction to ammonium (DNRA).Ammonium produced via DNRA is recycled either within the sediment or in the water column into which it diffuses and hence DNRA may sustain coastal eutrophication.In the anoxic layer of marine sediments, denitrification (DEN) and DNRA directly compete for NO − 3 as an electron acceptor and for organic carbon, sulfide, and others as electron donors.The outcome of this competition determines whether marine sediments act as source or sink of fixed nitrogen, which has impacts for the trophic status of the whole ecosystem.
Studies on the identification of these possible controlling factors have mostly used slurry incubations of sediment (Bonin et al., 1998;Fernandes et al., 2012;Lansdown , 2012), whole sediment core incubations with a final destructive sampling of the upper sediment layers (Christensen et al., 2000;Dong et al., 2009;Dunn et al., 2012), or whole sediment core incubations in which only the in-and outflow of the water column were analyzed (Gardner and McCarthy, 2009;Gardner et al., 2006;Smyth et al., 2013).The major limitation of these approaches is that the controlling factors are not studied directly in the intact nitrate-reducing sediment layer.In slurry incubations, all in situ gradients are destroyed and the conditions formerly established in the nitratereducing and the neighboring sediment layers are blended.Furthermore, rates determined in slurries often overestimate the in situ rates (Christensen et al., 2000;Revsbech et al., 2006).Whole core incubations have the advantage that the biological and chemical stratification of the sediment stays intact during the incubation (but not necessarily during experimental sampling).The distinct investigation of the nitrate-reducing sediment layer, however, is not targeted by this method, neither in terms of nitrogen conversions, nor in terms of the controlling factors.
We therefore investigated sediment cores with intact biological and chemical stratification during experimental incubation and sampling with respect to the vertical distribution of DEN and DNRA activities and their hypothesized controlling factors in the sediment.Coastal marine sediments were sampled at five field sites that differed in several environmental and sediment parameters and were analyzed in the laboratory.DEN activity was measured with the acetylene blocking technique combined with N 2 O microsensor measurements, whereas DNRA activity was measured with a newly developed gel probe-stable isotope technique.In parallel sediment cores, the vertical distribution of possible controlling factors was analyzed.For a methodical comparison, sediment from two contrasting field sites was investigated in both whole core and slurry incubations.Introduction

Conclusions References
Tables Figures

Back Close
Full These sites were chosen to cover a range of sediment characteristics that might influence the rates of dissimilatory nitrate reduction pathways (e.g., organic carbon and sulfide contents).Site characteristics and sampling details are given in Table 1.
At each site, 6-10 sediment cores were taken with acrylic core liners with an inner diameter of 9 cm and a length of 20 cm.The final height of the sediment and the water column in the core liners were 15 and 5 cm, respectively.Care was taken to avoid macrofauna burrows and shell debris during coring.The sediment cores were transported to the laboratory within 1-6 h and then immediately connected to the experimental setup as described below.If the transport could not be arranged on the same day, cores were stored over night at 15 • C.
For additional sediment slurry experiments, surface sediment (0-2 cm depth) was sampled from Dorum and from two sites of Janssand (i.e., from the upper sand flat and from the low-water line near a sulfidic seep) in October 2012.

Experimental setup and sampling design
Six intact sediment cores were connected to an incubation set-up, in which the overlying seawater was aerated and continuously exchanged from a reservoir (10 L) to maintain stable conditions at the sediment surface.The water level was kept constant by drawing off excess overlying water with a peristaltic pump.Seawater was prepared Introduction

Conclusions References
Tables Figures

Back Close
Full from Red Sea Salt (Red Sea Fish Farm, Israel) at the salinity of the respective sampling site and a pH of 8.0-8.4.The seawater was amended with NaNO 3 to a final concentration of 50 µmol L −1 NO − 3 .The sediment cores were incubated at a constant temperature that was close to the in situ temperature at the time of sediment collection (Table 1).After starting the pumps, the overlying water of the cores reached a stable concentration of 50 µmol L −1 NO − 3 within 1 day, but a further incubation for 3-5 days was scheduled to allow steady-state conditions to develop inside the sediments.The NO − 3 concentration of the overlying water was monitored each day and corrected if necessary.Additional cores for sediment analyses were kept submersed in an aquarium under the same conditions as in the incubation set-up.After the pre-incubation period, the vertical distribution of DNRA and DEN activities and of physical-chemical parameters assumed to influence these two pathways were measured in whole sediment cores.

Microsensor measurements
Microsensors for O 2 (Revsbech, 1989), NO − 3 (Larsen et al., 1997), H 2 S (Jeroschewski et al., 1996), N 2 O (Andersen et al., 2001), and pH (Schulthess et al., 1981) were constructed at the Max Planck Institute for Marine Microbiology in Bremen (Germany).The sensors were calibrated each day and used for profiling the 6 sediment cores that were connected to the incubation set up.In each core 3 to 12 profiles were measured at randomly selected spots.The custom-made programs µ-Profiler, DAQserver, and LINPOS-server were used for measurement automation and data acquisition (http://www.microsen-wiki.net).Vertical profiles were recorded in steps of 250 or 500 µm, starting at 3 mm above the sediment surface and ending 10-20 mm below the sediment surface.
To determine the vertical distribution of DEN activity in the sediment, the acetylene inhibition technique was used (Sørensen, 1978).Acetylene inhibits the last step of denitrification so that N 2 O becomes the end product which accumulates in the sediment and can be measured with an N 2 O microsensor.Three cores were incubated over Introduction

Conclusions References
Tables Figures

Back Close
Full night at 10 % acetylene saturation in the overlying water.As a measure of DEN activity, the N 2 O flux (J) between the layer of N 2 O production (which coincides with the layer of NO − 3 consumption) and the sediment surface was calculated from the steady-state N 2 O concentration profiles using Fick's law of diffusion.Since the sediment cores were closed at the bottom, the continuous N 2 O production led to an accumulation of N 2 O in the deeper sediment layers and thus to a progressively flattening concentration gradient.Instead of using this non-steady-state gradient to calculate the downward N 2 O flux, the upward N 2 O flux was multiplied by 2 for the quantitative comparison with the NO − 3 flux from the sediment surface to the layer of NO − 3 consumption.Several profiles were offset-corrected, setting the lowest measured value in the overlying water to zero.
Total dissolved sulfide (i.e., the sum of H 2 S, HS − , and S 2− ) was calculated from the H 2 S and pH microprofiles according to Jeroschewski et al. (1996).The pK 1 value (i.e., the dissociation coefficient for the equilibrium between H 2 S and HS − ) was corrected for temperature and salinity of the respective sampling site according to Millero et al. (1988).Introduction

Conclusions References
Tables Figures

Back Close
Full

Combined gel probe and isotopic labeling technique
The depth distribution of DNRA activity was measured using the gel probe stable isotope technique of Stief et al. (2010) with minor modifications.Briefly, the pre-hydrated polyacrylamide gel in the probe, deoxygenated with He, were inserted into the sediment.Fourty-eight hours later, the overlying water was amended with 15 N-labeled NO − 3 (99 % 15 N atom %, Cambridge Isotope Laboratories, Andover, MA, USA) to a final concentration of 50 µmol L −1 .The probes were left in the sediment for another 24-48 h for complete equilibration with the pore water.After retrieveing the probes, the gel was immediately cut into a series of 20 1 mm pieces with a home-made cutter.Each slice was placed in a pre-weighed 3-mL vial (Exetainer; Labco, High Wycombe, UK), weighed again, and flushed twice with He for 60 s (with 1 min equilibration time in between).Samples from the Mississippi Delta and Janssand experiments were spiked with 50 µL of 10 µmol L −1 14 NH + 4 .12 M NaOH and hypobromite were injected to convert NH + 4 to N 2 (Warembourg, 1993).Samples were left for 3 days at 21 • C in the dark to allow the reaction to N 2 to proceed.In headspace samples of 100-250 µL, the isotope ratio of 28 N 2 , 29 N 2 , and 30 N 2 was determined by gas chromatography-isotopic ratio mass spectrometry (VG Optima, ISOTECH, Middlewich, UK) against air standards.Calibration standards were prepared with MilliQ water adjusted to different 15 NH + 4 concentrations (0, 5, 10, and 25 µmol L −1 ).Gel probes were immersed in the standard solutions and allowed to equilibrate for 24 h.After incubation, the gels were treated in the same way as described above.For each 15 NH + 4 concentration, 3-5 replicate gel slices were analyzed.
The 15 NH + 4 concentration was calculated from the isotope ratios of 28 N 2 , 29 N 2 , and 30 N 2 in the sample and in air standards using equations given by Risgaard-Petersen et al. (1995).
As a measure of DNRA activity, the + 4 upward flux was multiplied with 2, assuming that the downward flux was the same as the upward flux (see explanation above).In addition, all profiles were corrected for the abundance of 15 NH + 4 in the pore water of coastal marine sediment as found by Prokopenko et al. (2011) (i.e., 0.374 15 N-Atom %), assuming a similar natural enrichment in 15 NH + 4 at all sampling sites investigated here.The diffusion coefficient of NH + 4 (D w ) was taken as 1.5 × 10 −5 , 1.8 × 10 −5 and 2.0 × 10 −5 cm 2 s −1 at 15, 21 and 25 • C, respectively (Li and Gregory, 1974).

Porewater analyses
For NO − 3 and NH + 4 analyses, 3 sediment sub cores (inner diameter of 2.6 mm) were taken from each sampling site and cut into 2 mm slices down to a depth of 20 mm.To each sediment slice 2 mL of artificial seawater (Red Sea Fish Farm, Israel) adjusted to the respective in situ salinity were added.After thorough mixing, the sediment suspension was centrifuged at 1000g for 10 min and the supernatant was analyzed for NO − 3 and NH + 4 .The dilution with artificial seawater and the weight of the sediment slices were taken into account for the calculation of pore water concentrations.NO − 3 was analyzed in 25-µL samples after chemical conversion to NO that was quantified by the chemoluminescence detector of a NO x -analyser (CLD 66, EcoPhysics, Germany) (Braman and Hendrix, 1989).In the following, the NO x data are reported as NO − 3 concentrations, since NO − 2 concentrations were generally very low.NH + 4 was analyzed with the salicylate-hypochlorite method scaled down to 1-mL samples (Bower and Holm-Hansen, 1980).
For the determination of total dissolved iron after Viollier et al. (2000), 3 sub cores were cut into 2 mm slices down to 20 mm.All plastic ware was cleaned with acid (suprapure HNO 3 Merck, Darmstadt) and solutions were made with deoxygenated water.Cutting of sediment sub cores and handling of the sediment slices were done in a dinitrogen-flushed glove box.The sediment slices were weighed and 1.5 mL of anoxic Introduction

Conclusions References
Tables Figures

Back Close
Full MilliQ H 2 O was added.The mixed samples were centrifuged (5 min at 3000 g) and the supernatant was removed completely and filtered (Millipore, Millex-GN 13 mm).The pelleted sediment was used for solid-phase iron analysis described below.For pore water analyses 1 mL of the filtered sample was taken (Viollier et al., 2000).Samples were measured undiluted, whereas the standards (prepared from a 1 mM stock solution diluted with a NaCl solution at the salinity of the respective sampling site) were diluted 1 : 2 with 0.5 M HCl.

Sedimentary adsorption of ammonium
The adsorption of DNRA-derived NH + 4 to sediment was quantified in sediment sub cores cut into the layers 0-2, 2-7, and 7-12 mm.Sediment from the Mississippi Delta was cut into the layers 0-5 and 5-10 mm only.Each slice was split into two pieces of approximately equal size.The sediment pieces were weighed and amended with 1.5 mL of one of the following anoxic solutions: (A.) NaCl adjusted to the respective in situ salinity as a blank, (B.) NaCl enriched with 50 µmol L −1 14 NH + 4 to mimic newly produced NH + 4 .The sediment was incubated for 30 min and vigorously shaken every 10 min.Following this incubation, the samples were centrifuged for 5 min at 3000 g.In the supernatant, NH + 4 was analyzed as described above.

CNS content
CNS was analyzed in freeze-dried sediment aliquots by combustion gas chromatography (Carlo Erba NA-1500 CNS analyzer).

Total organic carbon
To estimate the total organic carbon content of the sediments, the loss on ignition was determined after heating the samples to 550 • C for 4 h.Introduction

Conclusions References
Tables Figures

Back Close
Full Easily extractable sulfide (mainly FeS) was measured after Simpson (2001) in sediment sub cores cut for all sampling sites as described above for the iron determination.Samples were handled in a dinitrogen-flushed glove box and all the plastic ware was cleaned as described before.Solutions were made with deoxygenated water.The sediment slices were weighed and a subsample of 0.150-0.650g wet weight was used for the subsequent analysis.Calibration standards were made in deoxygenated water out of a 100 mM Na 2 S stock solution and diluted 1 : 10 with 1 M suprapure H 2 SO 4 (Merck, Darmstadt).

Solid-phase iron content
Extraction of solid-phase iron from the sediment was made with 0.5 M HCl for 1 h (Kostka and Luther, 1994).The extracts were filtered (Millipore, Millex-GN 13 mm) and handled as described above for pore water iron analysis.The extracted sediment samples were diluted like the standard (see paragraph for pore water iron analyses) and measured after Viollier et al. (2000).

Sediment slurry incubations
In addition to the whole core incubations, slurry experiments were conducted with sediment from Dorum and Janssand (upper flat vs. low-water line near a sulfidic seep).
Approximately 1 g of homogenized sediment was transferred into 14 6-mL exetainers (7 for DEN and 7 for DNRA rate measurements) and mixed with 3 mL of anoxic 35 ‰ artificial seawater (Read Sea Salt, Red Sea Fish Farm, Israel).The water was amended with 50 µmol L −1 15 NO − 3 .The exetainers were flushed for 3 min with He to create anoxic conditions and then continuously rotated in a dark incubator at room temperature.At each of 7 sampling time points, 2 exetainers were sacrificed, 1 for DEN and 1 for DNRA rate measurements.Biological activity in the DEN exetainers was terminated by adding Introduction

Conclusions References
Tables Figures

Back Close
Full

Characteristics of sampling sites and sediments
The five sampling sites and sediments covered a wide range of environmental parameters and sediment characteristics (Tables 1 and 2).The five coastal marine sediments were muddy, sandy or muddy-to-sandy with porosities ranging from 45 to 85 %.At the time of sampling the sediments, NO − 3 concentrations in the water column were generally low, with one notable exception at the freshwater-impacted Limfjord (124 µmol L −1 NO − 3 , 2 ‰ salinity).In situ temperatures ranged from 2.9 to 30.5 • C.
Total carbon contents differed less than expected between the sediments and ranged from 0.6 to 3.0 %, while nitrogen contents were particularly low at Dorum (0.02 %) and highest at Aarhus Bight (0.30 %).The sediments differed largely in their capacity to adsorb NH + 4 produced by DNRA, with virtually no adsorption at Aarhus Bight and very high adsorption at Janssand (46 %).Total organic carbon content was highest at the Mississippi Delta (32.9 ± 4.6 %).

Vertical gradients of pore water concentrations
Steady-state microprofiles of pore water solutes directly or indirectly involved in the activity of both DEN and DNRA were measured in laboratory-incubated sediment cores Introduction

Conclusions References
Tables Figures

Back Close
Full from five coastal marine sampling sites (Fig. 1a-o).The penetration depth of O 2 into the sediment was mostly around 2.5 mm, except for the Janssand and Aarhus Bight sediments (1.5 and 3.5 mm, respectively).NO − 3 penetration always exceeded O 2 penetration and was particularly deep in the Mississippi Delta sediment (9 mm).In the Dorum sediment, the NO − 3 profiles revealed substantial nitrification activity at the surface, which increased NO − 3 availability in the sediment.From the linear concentration gradient below the sediment surface, the NO − 3 flux into the layer of dissimilatory nitrate reduction was calculated and was highest in the Mississippi Delta and lowest at Janssand (Fig. 2).
Ammonium concentrations generally increased with sediment depth.The concentra- Total sulfide profiles derived from H 2 S and pH microprofiles revealed substantial differences between the five sediments (Fig. 1k-o), with low concentrations in Dorum and the Aarhus Bight, intermediate concentrations in the Mississippi Delta and high concentrations in the Limfjord and Janssand sediments.In the Janssand sediment, heterogeneity was particularly high and at several spots total sulfide concentrations reached up to 3.9 ± 1.7 mmol L −1 .After incubation with 10 % acetylene, a distinct N 2 O concentration peak indicating DEN activity developed in the anoxic layer of the sediments from all sampling sites.The corresponding N 2 fluxes were between 3.3 ± 0.7 and 11.1 ± 1.0 nmol N cm −2 h −1 (Fig. 2).In none of the sediments, N 2 O was detectable without acetylene inhibition (Fig. 1f-j).

Vertical activity distribution of dissimilatory nitrate reduction
A distinct concentration peak of 15 NH + 4 in the layer of NO − 3 reduction indicating DNRA activity was only detected in Janssand (7.8 ± 1.8 µmol L −1 ) (Fig. 1j).High 15 NH + 4 concentrations were also found in the Mississippi Delta (9.6 ± 1.2 µmol L −1 ), but in the scattered depth profile a distinct concentration peak in the layer of NO − 3 reduction was not discernible.The 15 NH + 4 flux in Janssand was more than three times lower (0.5 ± 0.2 nmol N cm −2 h −1 ) than the corresponding N 2 fluxes.The calculated flux is possibly significantly underestimated since adsorption of NH + 4 to this sediment was particularly high (Table 2).
The relative partitioning between DEN and DNRA was assessed for each sampling site by a mass balance based on the NO − 3 , N 2 , and 15 NH + 4 fluxes (Fig. 2, Table 3).
At most sites, NO − 3 (the flux of which was set to 100 %) was quantitatively reduced to nitrogen gas.In sediment from Dorum (130.6 %), the Mississippi Delta (116.9 %) and the Limfjord (103.2 %), even more N 2 was produced than pore water NO − 3 was consumed.In sediment from Janssand, however, only 59.0 % of the NO − 3 consumed ended up as N 2 , while 8.9 % was converted to NH

Easily extractable solid-phase sulfide and iron
Acid volatile sulfide (mainly FeS) showed for the Mississippi Delta, Limfjord and Janssand sediments a linear increase in concentration starting at the sediment surface (Fig. 3).In combination with the sulfide freely dissolved in the pore water, these Introduction

Conclusions References
Tables Figures

Back Close
Full three sampling sites had the highest amount of readily available sulfide species in the sediment.
Total dissolved iron had highest values in the Limfjord sediment, with a continuous increase from the sediment surface down to 20 mm (Fig. 3).A distinct peak was measured in Janssand starting at 3 mm.Solid phase iron showed no distinct distribution pattern at any of the sampling sites (Fig. 3), with concentrations ranging from 4.1 ± 1.8 to 12.6 ± 3.2 µmol g −1 wet weight.

Slurry experiment
In addition to the whole core experiment, slurry incubations were conducted to test the differences in DEN and DNRA activities in a diffusive (whole core) vs. an advective setting (slurry).
In sediment from all three sampling sites (Dorum, Janssand upper flat and low-water line near a sulfidic seep), the reduction of NO − 3 was accompanied by the simultaneous production of 15 N 2 (DEN activity) and 15 NH + 4 (DNRA activity).In contrast to the whole core incubations, the relative share of DNRA in dissimilatory nitrate reduction was substantial in the slurry incubations of all three sediments (Table 4).77, 56, and 37 % of the observed NO − 3 reduction in sediment from Janssand (low-water line), Janssand (upper flat), and Dorum, respectively, was explained by DNRA activity, while the remainder was explained by DEN activity (Table 4).The production of N 2 O was negligible in all slurried sediments (Table 4).Introduction

Conclusions References
Tables Figures

Back Close
Full In all coastal marine sediments studied as intact cores, denitrification (DEN) rather than DNRA was the dominant NO − 3 reduction pathway.DEN dominated irrespective of a large range of variation in sediment characteristics that are often discussed to favor either DEN or DNRA (e.g., sulfide concentration, (An and Gardner, 2002;Brunet and GarciaGil, 1996), carbon-to-nitrate ratio (Tiedje et al., 1982;Yin et al., 2002)).In all but one sediment, NO − 3 was mainly reduced to N 2 within the bounds of accuracy of the methodical approach.In some cases, the N 2 flux even exceeded the NO A notable exception to these observations was found in the sediment from Janssand (sampling site near sulfidic seeps).In this case, the gel probe technique revealed a distinct concentration peak of 15  4 adsorption in this sediment (46 %), the relative partitioning of dissimilatory nitrate reduction between DEN and DNRA might have been close to 82 and 18 %, respectively.These values are within the range found in other marine sediments, with DNRA activity accounting for 11-75 % and DEN accounting for 5-98 % of the total reduced NO − 3 (An and Gardner, 2002;Bonin et al., 1998;Dong et al., 2011;Koop-Jakobsen and Giblin, 2010;Porubsky et al., 2009).
The nitrogen budget in the Janssand sediment (and in Aarhus Bight sediment) was not closed; in these the NO  et al., 2011;McHatton et al., 1996;Risgaard-Petersen et al., 2006;Sayama, 2001).The latter scenario, however, would only apply to non-steady-state conditions when nitrate-storing microorganisms with exhausted stores fill up their vacuoles.Additionally, the high sulfide concentrations in the sediment pore water may have alleviated the inhibition of N 2 O reduction by acetylene, which is known to underestimate DEN rates (Sørensen et al., 1987).The opposite phenomenon (i.e., the combined flux of N 2 and 15 NH + 4 exceeding the NO − 3 flux) was observed in sediment from Dorum and the Mississippi Delta.Also in this case, the intracellular storage of NO − 3 by vertically migrating sulfide-oxidizing bacteria, foraminifera, and diatoms may serve as a possible explanation (again only under non-steady-state conditions).NO − 3 taken up at the sediment surface and transported to deep layers will not be reflected in the steady-state pore water profile of NO The downward transport of 15 NO − 3 by migrating cells may also be responsible for the shape of the 15 NH + 4 concentration peak observed in Janssand sediment.A closer look at this peak reveals that it extends to well below the layer of NO − 3 penetration.Non-steady-state modeling confirmed that the shape of this peak cannot be explained by downward diffusion and accumulation of 15 NH + 4 at depth during the exposure time of the gel probe of 2 days (data not shown).Hence, a faster spreading of the 15 NH + 4 concentration peak by moving cells seems more likely.Nitrate-storing and migrating microorganisms performing DNRA might be responsible for the deep occurrence of 15 NH + 4 and for the scatter in the 15 NH + 4 profiles measured in the other four sediments.

Slurry incubations
In all three sediments incubated as slurries, both DEN and DNRA contributed substantially to dissimilatory nitrate reduction.A direct comparison with whole core incubations is possible for the sediments from Dorum and Janssand (sampling site near sulfidic seeps) which have been used in both types of incubation.For Dorum sediment, the whole core incubation revealed DEN activity exclusively, while the slurry incubation revealed a relative partitioning between DEN and DNRA of 61 and 39 %, respectively of the total reduced NO − 3 .For Janssand sediment, the slurry incubation shifted the partitioning from a dominance of DEN (i.e., 87 vs. 13 %) in the whole core incubation to a dominance of DNRA (i.e., 18 vs.82 %).In sediment from the upper flat in Janssand and Dorum tested in slurries, the nitrogen budget was closed, leaving no room for a substantial involvement of intracellular NO − 3 storage, microbial NO − 3 assimilation, or NH + 4 adsorption.At the low water line from Janssand 6 % of the total nitrogen budget are missing that could account for the above mentioned additional NO − 3 sinks.Additionally, N 2 O production did not exceed 1 % of the NO − 3 consumption, but interestingly the highest N 2 O production rate was found in the most sulfidic sediment, probably due to partial inhibition of dissimilatory N 2 O reduction (Brunet and GarciaGil, 1996;Sørensen et al., 1980).Introduction

Conclusions References
Tables Figures

Back Close
Full It could be assumed that the different results obtained by the two types of sediment incubation are explained by much higher NO − 3 consumption rates in the slurry incubations due to the advective vs. diffusive substrate supply.However, the NO − 3 consumption rates measured in slurry and whole core incubations were either not significantly different (Janssand: −42.4 ± 4.1 nmol cm −3 h −1 and −47.0 nmol cm −3 h −1 , respectively), or even lower in the slurry incubations (Dorum: and −69.7 nmol cm −3 h −1 , respectively).Alternatively, it may be speculated that in sediment slurries many more DNRA bacteria are supplied with NO − 3 than in intact sediment cores, especially in the absence of advective porewater transport.When NO − 3 is exclusively supplied by diffusion, many microorganisms capable of DNRA might be cut off the NO − 3 supply from above because they reside deeper in the sediment than microorganisms capable of DEN.DNRA microorganisms are active only in sediment layers that are completely anoxic and in which strongly reducing conditions prevail, often characterized by near absence of NO − 3 and presence of sulfide (Tiedje et al., 1982).
In contrast, DEN microorganisms can cope well with oxic-anoxic shifts (e.g., due to porewater irrigation by the tides or burrowing animals) and therefore are active closer to the oxic-anoxic interface in the sediment (Brettar and Rheinheimer, 1991;Thamdrup and Dalsgaard, 2008;Tiedje et al., 1982).In fact, the gel probe technique previously revealed that the activity maximum of DNRA was located slightly deeper in stream sediment than the activity maximum of DEN (Stief et al., 2010).In stratified sediments, DEN microorganisms have thus the potential to outcompete DNRA microorganisms for NO − 3 .The slurry incubation of sediment disrupts these stratifications and exposes all microorganisms to homogeneous conditions with respect to substrates and products.DNRA microorganisms and rates, even if irrelevant in situ (Christensen et al., 2000;Laverman et al., 2006;Revsbech et al., 2006) might thus get more important in the slurry incubations because here they are not nitrate-limited any longer.Taken together, slurry incubations may overestimate DNRA rates due to enhanced NO − 3 supply, Introduction

Conclusions References
Tables Figures

Back Close
Full

Environmental factors controlling the partitioning between DEN and DNRA
The five sampling sites were chosen to cover a range of environmental factors that are proposed to promote or repress either DEN or DNRA.These factors (e.g., availability of NO − 3 and electron donors such as organic carbon, sulfide, or reduced iron) are thought to influence the partitioning of dissimilatory nitrate reduction between DEN and DNRA.In the following, the contrasting results observed for sediment from Janssand and the remaining sampling sites will be viewed in the light of these partitioning factors.Introduction

Conclusions References
Tables Figures

Back Close
Full content.Also with respect to organic carbon, no correlation with DNRA activity could be shown.

Sulfide
High sulfide concentrations in the sediment pore water have a strong influence on the activity of both DEN and DNRA (An and Gardner, 2002;Brunet and GarciaGil, 1996;Burgin and Hamilton, 2007;Christensen et al., 2003;Nizzoli et al., 2006).Sulfide can serve as an electron donor for DEN and DNRA, but at very high concentration it inhibits the last step of DEN, but not DNRA.Sulfidic sediments therefore tend to have a high capacity to reduce NO − 3 and to produce N 2 O and NH + 4 (An and Gardner, 2002;Brunet and GarciaGil, 1996).
An almost gradual increase in free total sulfide concentrations was observed in the sediments reaching from Aarhus Bight via Dorum, Mississippi, and Janssand, to Limfjord, but this increase was not reflected in increases of NO − 3 consumption and N 2 O or 15 NH + 4 production.The only striking findings were the substantial DNRA activity and the low DEN activity measured in Janssand sediment in which the second highest sulfide concentrations occurred.However, even in this sediment, DEN activity dominated dissimilatory nitrate reduction.Possibly, heterotrophic denitrifiers were outcompeted by autotrophic denitrifiers who can oxidize sulfide to sulfate in the presence of NO − 3 and therefore exist even in sediments high in sulfide (Brinkhoff et al., 1998;Shao et al., 2009).The role of sulfide in stimulating DNRA can be further questioned based on the results of the slurry incubations.During the experimental procedure, the in situ concentration of freely dissolved sulfide was diluted approximately 3-fold by the addition of sulfide-free seawater.Despite the diluted sulfide concentrations, clear shifts from DEN to DNRA were observed for the Dorum and Janssand sediments, contradicting the sulfide hypothesis.Finally, there was also no correlation between the sedimentary Introduction

Conclusions References
Tables Figures

Back Close
Full Even though the possible role of a high background concentration of NH + 4 for DNRA activity remains unclear, it might still be used as an indicator of highly reduced sediment where DNRA is more likely to occur than in less reduced sediment.

Iron
Besides carbon and sulfide, reduced iron can serve as another electron donor for dissimilatory nitrate reduction.For Geobacter and Dechloromonas spp., it has been shown that iron is used to reduce NO − 3 quantitatively to NH + 4 (Weber et al., 2006a,c).In the present study, no distinct correlation between the appearance of iron (pore water and solid-phase) and DNRA or DEN activity in the sediments was observed.The solid-phase iron contents were comparable at all five sampling sites and porewater iron concentrations were clearly not the highest in Janssand, the only sediment with measurable DNRA activity in whole core incubations.

Temperature
Seasonally or habitat-specific high temperatures were shown to favor DNRA over DEN activity (Dong et al., 2011;Jørgensen, 1989;Ogilvie et al., 1997).In our study, in situ temperatures ranged from ∼ 3 to 30 • C, but no relationship between the incubation temperature and the rates of DEN and DNRA was observed.At the particularly warm sampling site in the Mississippi Delta, high DNRA activity was expected (Dong et al., 2011).Indeed, the 15 NH + 4 pore water concentrations were the highest ones encountered in this study with values of up to 9.6 µmol L −1 (Fig. 1).However, the variability of data from replicate gels was very pronounced and only the average profile suggested set temperature was not a key factor that explained the presence or absence of DNRA activity.

Conclusions
The special conditions that prevail at Janssand apparently create a microenvironment in which DEN and DNRA can co-occur.At the low-water line, there is a high advective input of reduced compounds (e.g., sulfide and NH + 4 ) from the body of the tidal flat towards the sediment surface and a diffusive or advective input of O 2 and NO − 3 from the water column into the sediment (Billerbeck et al., 2006;Jansen et al., 2009;Røy et al., 2008).It can be assumed that microorganisms capable of DNRA cope better with the millimolar-range sulfide concentrations compared to DEN microorganisms.So unlike in non-sulfidic sediments, the in situ conditions at Janssand may have allowed DNRA microorganisms to thrive particularly well due to their sulfide tolerance and the lack of competition for NO − 3 with DEN microorganisms.However, the whole core incubation turned the Janssand sediment into a non-seep sediment without advective inputs of sulfide from below and O 2 and NO − 3 from above.Thus, the whole core incubation presumably underestimates the relative share of DNRA in dissimilatory nitrate reduction.On the contrary, the slurry incubation of Janssand sediment may overestimate the relative share of DNRA because of unlimited NO − 3 supply to DNRA bacteria that are outcompeted for NO − 3 by DEN bacteria in stratified sediments.The true partitioning of dissimilatory nitrate reduction between DNRA and DEN may consequently lie in between the values found in whole core and slurry incubations.Thus, while the gel probe method is a good alternative for DNRA measurements in diffusive settings, further improvements should aim at DNRA activity measurements in whole core incubations with realistic advective dynamics.Introduction

Conclusions References
Tables Figures

Back Close
Full  Full  Full Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | et al.
Intact sediment cores were collected at five coastal marine sites between September 2009 and July 2011.The sampling sites were Dorum, an intertidal flat north of Bremerhaven (Germany), Station M5 in Aarhus Bight (Denmark), the Mississippi Delta near Chauvin (USA), the Hjarbaek Fjord within the Limfjord (Denmark) and the low-water line of Janssand (near sulfidic seeps), a back barrier tidal flat of Spiekeroog Island (Germany).
Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 2.6.4Acid-volatile sulfide (AVS) content Discussion Paper | Discussion Paper | Discussion Paper | 500 µL of 50 % ZnCl 2 and stored for later analysis of 15 N 2 and N 2 O by GC-IRMS (VG Optima, ISOTECH, Middlewich, UK) and GC 7890 (Agilent Technologies), respectively.The DNRA exetainers were amended with 1 mL of 3 M KCl to aid desorption of NH + 4 from the sediment particles.Liquid subsamples were quickly taken from the DNRA exetainers and transferred into fresh vials for subsequent analysis of NO − above.Linear concentration changes over the time were used for rate calculations.
Discussion Paper | Discussion Paper | Discussion Paper | of NH + 4 in the layer of NO − 3 reduction was high in Janssand, Limfjord and Mississippi Delta sediment (100 to 167 µmol L −1 ) and low in Dorum and Aarhus Bright (2 to 28 µmol L −1 ).

Fluxes
from the concentration profiles shown in Fig. 1 (f-j) and used as measures of DEN and DNRA activity, respectively.In all sediments analyzed here, DEN rather than DNRA was the dominant NO − 3 respiration pathway and only in the Janssand sediment, a significant 15 NH + 4 flux directed from the layer of NO − 3 consumption to the sediment surface could be measured.Discussion Paper | Discussion Paper | Discussion Paper |

+ 4 .
Taking the NH + 4 adsorption of 46 % into account, the proportion of DNRA in NO − 3 consumption might have been as high as 19.3 % in this sediment.
Discussion Paper | Discussion Paper | Discussion Paper | resulted from DEN activity by nitrate-storing microorganisms or from Anammox activity.In four out of five sediments, the vertical 15 NH + 4 profiles measured with gel probes did not feature a distinct concentration peak in the layer of NO − 3 reduction, which would be the strongest argument for DNRA activity.Since the NH + 4 adsorption capacity of these four sediments did not exceed 15 %, the majority of newly produced NH + 4 would not have gone undetected by the gel probe technique.In fact, in all four sediments, the measured 15 NH + 4 concentrations clearly exceeded the natural abundance of 15 NH + 4 usually found in the pore water of coastal marine sediment (Prokopenko et al., 2011).However, the scattered vertical distribution of this 15 NH + 4 suggests production mechanisms other than dissimilatory reduction of porewater NO − 3 such as intracellular nitrate storage and DNRA activity by migrating microorganisms (see below).
NH + 4 that partially overlapped with the layer of NO − Discussion Paper | Discussion Paper | Discussion Paper | nitrate-reducing layer made up 8.9 % of the NO − 3 flux into the nitrate-reducing layer.Taking into account the high percentage of NH + microorganisms or intracellular storage of NO − 3 by sulfide-oxidizing bacteria, foraminifera, and diatoms (Kamp

− 3 ,
whereas its dissimilatory reduction in deep layers will be reflected in the porewater profiles of N 2 (measured as N 2 O) and 15 NH + 4 .In fact, intracellularly stored NO − 3 was detected in Dorum sediment (up to 22.3 µmol NO − 3 dm −3 , Stief et al., 2013) where the discrepancy between NO − 3 and N 2 fluxes was particularly pronounced, but in Mississippi Delta sediment no stored NO − 3 could be detected.Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | whereas whole core incubations may underestimate DNRA rates due to diminished NO − 3 supply, especially in the absence of advection.
Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | be questioned, however, because it also revealed high 15 NH + 4 concentrations well below the NO − 3 penetration depth that are maybe explained by DNRA activity of nitrate-storing and migrating microorganisms.We conclude that in our limited data Introduction Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Figure 1.

Fig. 1 .
Fig. 1.Vertical profiles of O 2 , NO − 3 and NH + 4 (a-e), 15 NH + 4 and N 2 O (f-j) and pH and Sulfide tot (k-o) measured in intact sediment cores from different coastal marine sampling sites.The NH + 4 profiles were measured in extracted pore water, 15 NH + 4 profiles (indicating DNRA activity) were measured with gel probes, while the other profiles were measured with microsensors.The DEN activity profiles (represented by the N 2 O profiles with acetylene) were measured after inhibition of the last step of denitrification with acetylene.Means ± standard deviation of 3-9 profiles are shown.

Fig. 3 .
Fig. 3. Pore water sulfide (Sulfide tot ), AVS (acid-volatile sulfide), Pore water iron (Fe tot PW) and Solid-phase iron (Fe tot Sed) of the different sampling sites.Solid-phase pools (AVS and solidphase iron) are shown per gram wet weight (WW).Means ± standard deviation of 3 replicate sub cores are shown.

Table 1 .
Location and characteristics of sampling sites.

Table 3 .
Mass balance for dissimilatory nitrate reduction in intact sediment cores sampled at five coastal marine investigation sites.Fluxes of N-NO − 3 were set to 100 % and fluxes of N-N 2 (indicating DEN activity) and 15 N-NH + 4 (indicating DNRA activity) were calculated as relative shares of NO − 3 fluxes.n.d.: not detected.