Nitrate consumption in sediments of the German Bight (North Sea)

Journal of Sea Research xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Journal of Sea Research journal homepage: www.elsevier.com/locate/seares

Nitrate consumption in sediments of the German Bight (North Sea) Andreas Neumanna,⁎, Justus E.E. van Beusekoma,b, Moritz Holtappelsc, Kay-Christian Emeisa a

Helmholtz Zentrum Geesthacht Center for Materials and Coastal Research, Max-Planck Str. 1, D-21502 Geesthacht, Germany University of Hamburg, Center for Earth System Research and Sustainability, Institute for Hydrobiology and Fisheries Science, Große Elbstraße 133, D-22767 Hamburg, Germany c Alfred-Wegener-Institute, Helmholtz Centre for Polar and Marine Research, Am Alten Hafen 26, D-27568 Bremerhaven, Germany b

A R T I C L E I N F O

A B S T R A C T

Keywords: North Sea Anthropogenic N-load Denitrification Diffusive porewater exchange Advective pore water exchange Permeability

Denitrification on continental margins and in coastal sediments is a major sink of reactive N in the present nitrogen cycle and a major ecosystem service of eutrophied coastal waters. We analyzed the nitrate removal in surface sediments of the Elbe estuary, Wadden Sea, and adjacent German Bight (SE North Sea) during two seasons (spring and summer) along a eutrophication gradient ranging from a high riverine nitrate concentrations at the Elbe Estuary to offshore areas with low nitrate concentrations. The gradient encompassed the full range of sediment types and organic carbon concentrations of the southern North Sea. Based on nitrate penetration depth and concentration gradient in the porewater we estimated benthic nitrate consumption rates assuming either diffusive transport in cohesive sediments or advective transport in permeable sediments. For the latter we derived a mechanistic model of porewater flow. During the peak nitrate discharge of the river Elbe in March, the highest rates of diffusive nitrate uptake were observed in muddy sediments (up to 2.8 mmol m− 2 d− 1). The highest advective uptake rate in that period was observed in permeable sediment and was tenfold higher (up to 32 mmol m− 2 d− 1). The intensity of both diffusive and advective nitrate consumption dropped with the nitrate availability and thus decreased from the Elbe estuary towards offshore stations, and were further decreased during late summer (minimum nitrate discharge) compared to late winter (maximum nitrate discharge). In summary, our rate measurements indicate that the permeable sediment accounts for up to 90% of the total benthic reactive nitrogen consumption in the study area due to the high efficiency of advective nitrate transport into permeable sediment. Extrapolating the averaged nitrate consumption of different sediment classes to the areas of Elbe Estuary, Wadden Sea and eastern German Bight amounts to an N-loss of 3.1 ∗ 106 mol N d− 1 from impermeable, diffusion-controlled sediment, and 5.2 ∗ 107 mol N d− 1 from permeable sediment with porewater advection.

1. Introduction The German Bight is part of the southern North Sea and is semienclosed by a densely populated and industrialized hinterland from which major continental European rivers (Rhine, Maas, Elbe, Weser, and Ems) transport significant amounts of nutrients to the coastal waters (Los et al., 2014). Riverine nitrogen loads into the German Bight reached a maximum in the 1980s (e.g. Radach and Patsch, 2007). Stratification from high freshwater discharge in combination with high riverine nutrient loads led to large phytoplankton blooms and oxygen deficiencies during the 1980′s (Westernhagen von et al., 1986; Hickel et al., 1993). Eutrophication promoted blooms of phytoplankton, spread of green macroalgae and a decrease in seegrass, especially in the Wadden Sea, one of the largest intertidal ecosystems on earth (Cadée

and Hegeman, 2002; Reise and Siebert, 1994; van Dolch et al., 2013; van Katwijk et al., 1997). Although various mitigation efforts constantly reduced the nutrient load since then (Amann et al., 2012), the entire SE North Sea is still flagged as an eutrophication problem area (OSPAR, 2010). Natural attenuation mechanisms such as denitrification and anammox in sediments counteract eutrophication by converting reactive nitrogen in suboxic sediment layers to inert N2. These processes are a significant global sink for nitrate (Middelburg et al., 1996; Seitzinger et al., 2006), and are particularly effective in sediments of continental margins, shelf seas, and coastal permeable sediments with elevated concentrations of organic matter (e.g., Cornwell et al., 1999). In the face of increasing developments and modifications of sea floors in offshore North Sea areas such as bottom trawling, dredging, removal of

⁎

Corresponding author. E-mail addresses: [email protected] (A. Neumann), [email protected] (J.E.E. van Beusekom), [email protected] (M. Holtappels), [email protected] (K.-C. Emeis). http://dx.doi.org/10.1016/j.seares.2017.06.012 Received 1 February 2016; Received in revised form 20 June 2017; Accepted 23 June 2017 1385-1101/ © 2017 Published by Elsevier B.V.

Please cite this article as: Neumann, A., Journal of Sea Research (2017), http://dx.doi.org/10.1016/j.seares.2017.06.012

Journal of Sea Research xxx (xxxx) xxx–xxx

A. Neumann et al.

sand and gravel, and offshore construction (Emeis et al., 2015), there is a need to assess the role of sediments and different sediment types in the turn-over of organic matter and nutrients. Of particular interest is their capacity to sequester or eliminate potentially problematic substances such as reactive nitrogen - a natural environmental service of considerable value to society (e.g., Costanza et al., 1997). In spite of its acknowledged importance as an ecosystem function, benthic denitrification rate measurements on continental shelves and in coastal seas are scarce and have rarely been investigated in terms of sediment texture, organic matter loading and seasonality. The available data base for North Sea sediments is highly local and sporadic, and model attempts to estimate the elimination of reactive nitrogen in German Bight sediments have no recent data basis (e.g., Pätsch et al., 2010). Exploitation of the available data sets for a systematic assessment is also hampered by the heterogeneity of methods used. Schroeder et al. (1996) measured nitrate consumption rates in the Elbe estuary in the late 1980s with benthic chambers. Lohse et al. (1993, 1996) investigated several stations encompassing different sediment types in the outer German Bight in the early 1990s applying the acetylene blocking method, core flux incubations and isotope pairing to determine denitrification rates. More recently, Deek et al. (2012) measured N-turnover in Wadden Sea sediment using core flux incubations and isotope pairing. Gao et al. (2012) and Marchant et al. (2014, 2016) reported denitrification rates in intertidal and subtidal permeable sediments obtained from slurry incubations and percolated sediment cores. However, these measurements cover a limited area, especially in subtidal waters, and do not trace the N-loss along major river runoff characterized by constantly changing environmental conditions with respect to nutrient loads (Amann et al., 2012) and primary production (e.g. Cadée and Hegeman, 2002; Philippart et al., 2007; van Beusekom et al., 2009), both of which are among the most important determinants of denitrification rates. In this paper we report pore water concentration profiles of nitrate for a wide range of North Sea sediment types and for various bottom water concentrations of nitrate. Assuming steady state conditions, the nitrate profiles reflect the balance of transport intensity (either by diffusive or advective processes) and sediment reactivity. If the nitrate transport is high in comparison with reactivity, then nitrate penetrates deeply into the sediment until it is eventually consumed by denitrifying bacteria and archaea. Conversely, all nitrate is consumed in the uppermost sediment layer in cases where the nitrate import from the bottom water is low compared to the sediment reactivity. It is thus possible to estimate nitrate turnover rates on the basis of transport intensity and the depth of nitrate penetration into the sediment. Assuming steady-state conditions such interpretation of porewater profiles is a standard procedure to calculate diffusive solute fluxes (e.g. Berg et al., 1998). Based on the concentration gradient, the calculation of the diffusive flux requires only the diffusivity of the solute in water, corrected for effects of temperature and porosity. As implemented in the one-dimensional transport-reaction model of Berg et al. (1998), this method interprets the first and second derivative of the nitrate concentration profile to calculate the diffusive nitrate flux and associated reaction rates. If influx and efflux are not balanced then the model assumes either production or consumption of nitrate to balance fluxes and turnover. However, mass transport by molecular diffusion is slow compared to porewater advection along pressure gradients in permeable sediments. Coarse grained permeable sediments are found predominantly in shallow coastal waters with strong tidal currents that usually form a rippled seabed topography. Across the ripples, the water current induces pressure gradients that pump bottom water into the sediment where reactive compounds such as nitrate are consumed. Although this form of porewater advection has been well described (e.g. Elliott and Brooks, 1997; Precht and Huettel, 2003; Precht et al., 2004, and Huettel et al., 2014) it was often neglected in field measurements and biogeochemical modelling. Here, we interpret measured nitrate profiles in two

Table T1 Overview of sampling campaigns in the Elbe Estuary, German Bight and North-Frisian (NF) Wadden Sea. Cruise

Vessel

Date (month/year)

area

Pr-0309 HE-304 Pr-0909 HE-318

RV RV RV RV

03/2009 05/2009 09/2009 02/2010

Elbe Estuary, NF Wadden Sea German Bight Elbe Estuary, NF Wadden Sea German Bight, Dogger Bank

L. Prandtl Heincke L. Prandtl Heincke

ways. First, we employ the method of Berg et al. (1998) for a conservative estimate of the diffusive nitrate flux by assuming steady-state diffusion as the dominant benthic transport mode. Second, we expand the analytical model for porewater flow by Elliott and Brooks (1997) and Ahmerkamp et al. (2015) to derive an estimate of sediment nitrate consumption for advective pore water regimes. Finally, we use the rates determined by both methods and apply them to areas with specific permeability properties to estimate total nitrate consumption in the SE German Bight. 2. Material and methods 2.1. Study site and sampling The sampling campaign in Elbe estuary, German Bight and NorthFrisian Wadden Sea was carried out between March 2009 and February 2010 during 4 cruises with RV Ludwig Prandtl and RV Heincke (Table T1). The sampled stations are depicted in Fig. F1. Temperature, salinity and oxygen saturation of the bottom water were measured with an OTS 1500 multiprobe (Meerestechnik-Elektronik) during the cruises Pr-0309 and Pr-0909, and with a SBE911plus (Seabird) during the cruises HE304 and HE-318. Sediment cores were retrieved with a multicorer equipped with acrylic glass tubes (PMMA, 60 cm long and 10 cm wide). A subset of these tubes was prepared for pore water sampling by drilling holes in 1 cm intervals and sealing them with a septum prior to deployment. Directly after retrieval, the supernatant of the cores was carefully removed and the pore water was extracted with rhizon core solution samplers (Rhizosphere Research) connected to disposable syringes (Meijboom and van Noordwijk, 1992; Seeberg-Elverfeldt et al., 2005). The first few hundred microliters of pore water were discarded to remove air bubbles and oxygenated pore water. The porewater samples were then transferred to evacuated Exetainers (Labco) and stored frozen until nutrient analysis. One core of each station was sliced in 1 cm intervals and stored frozen for further analysis of sediment characteristics. 2.2. Sediment characteristics The frozen sediment slices were freeze-dried, and the resulting weight loss was used to calculate the water content and porosity based on the assumed mean grain density of 2.65 g cm− 3. The dried sediment was then sieved through mesh sizes of 1000 μm, 500 μm, 250 μm, 125 μm, and 63 μm to establish the grain size distributions using Gradistat (Blott and Pye, 2001). Additional subsamples of the dry sediments were used to determine the concentrations of total nitrogen and organic carbon with an Elemental Analyzer (Thermo Flash EA) calibrated against acetanilide. 2.3. Pore water nutrient analysis The pore water samples were kept frozen in septum capped Exetainers (Labco). After fast thawing in a water bath at room temperature, the samples were acidified with 6 M hydrochloric acid (1% v/ v final concentration) to stabilize any gaseous ammonia as ammonium 2

Journal of Sea Research xxx (xxxx) xxx–xxx

A. Neumann et al.

Fig. F1. Overview of the sampled stations in Elbe Estuary, German Bight and North Frisian Wadden Sea. German Bight refers to the area indicated by the dashed rectangle (east of 5° E and south of 55.5° N), which is also the area used for extrapolation.

of the induced porewater flow start and end at respective low and high pressure fields at the sediment surface, which results in a complex 2dimensional porewater flow well described by Elliott and Brooks (1997) and Huettel et al. (2014, 1996, 1998). The 2-D flow field can be integrated in the horizontal plane as shown by Elliott and Brooks (1997) to established the spatially averaged penetration depth of a solute (entering the sediment at t = 0) as a non-linear function of time (t):

by lowering pH < 7. Aliquots were then taken with syringes through the septum and analyzed with a nutrient autoanalyzer (AA3, Seal Analytical) for NH4+, NO3−, NO2− PO43 − according to Grasshoff et al. (1983). 2.4. Estimation of penetration depth The nitrate penetration depth was determined as the interval from the sediment surface down to the sediment depth where the local nitrate concentration falls below 5% of the nitrate concentration in the bottom water or in the internal nitrate peak, if present.

L=

1 t ln ⎛0.4 k 2 K hm + 1⎞ k θ ⎠ ⎝

(1)

where θ is the porosity. Here, Eq. (1) is the dimensionalized form of E25 in (Elliott and Brooks, 1997). The velocity (u) at which the penetration depth increases is given by the first derivative of Eq. (1) with respect to t. When corrected by the porosity, the velocity u represents the volume flux at the penetration depth for a specific time t

2.5. Calculation of diffusive nitrate fluxes Reaction rates and diffusive fluxes of nitrate across the sedimentwater interface were calculated on the basis of the concentration profiles and sediment porosities using the algorithm of Berg et al. (1998), assuming steady state profiles. The local effective diffusion coefficients were corrected for effects of porosity and temperature. The bottom water concentrations have been excluded from the calculations, since the modeled flux over the water-sediment interface and thus the total reaction rate is very sensitive to errors in the bottom water concentration and the porosity of the top layer. The depth-integrated nitrate reaction rate is used as areal consumption rate.

u=

k K hm θ k 2 K hm t + 2.5 θ

(2)

Solving Eq. (1) for t and inserting the result in Eq. (2) gives the volume flux under steady state conditions at a specific penetration depth L:

u=

k K hm 2.5 exp (L k )

(3)

In case of denitrification, the lower bound of the denitrification layer is given by the nitrate penetration depth, while the upper bound is defined by the oxygen penetration depth. We assume that denitrification is completely inhibited in the oxic layer and we further neglect nitrification, assuming that merely the bottom water nitrate enters the denitrification layer at the oxycline. From Eq. (3) it becomes apparent that the volume flux depends on the depth L, so that uox at the oxygen penetration depth Lox must be larger than unit at the deeper nitrate penetration depth Lnit. The divergence represents the fraction of

2.6. Estimation of advective nitrate fluxes To establish the advective nitrate flux, we applied the following mechanistic model: In subtidal permeable sands, the bulk porewater flow is a function of the hydraulic conductivity of the sediment (K) and the hydraulic head at the sediment surface (hm). The latter is induced by the bottom water flow over a rippled seabed topography that can be characterized by the ripple wave number k = 2π / ripple length. The pathways 3

Journal of Sea Research xxx (xxxx) xxx–xxx

A. Neumann et al.

for a variety of sediment types and different nitrate concentrations in the overlying bottom water (Fig. F2). These profiles indicate that the sediment depth to which nitrate penetrates into the sediment reflects the reactivity of the sediment. For example, the sediment at stations ML and NW (cruises Pr-0309 and Pr-0909) is similar with respect to permeability and bottom water nitrate, but differs in TOC as a proxy for reactivity. Sediment at ML had 2.8–3.5% TOC and nitrate entered 1–2 cm into the sediment, which suggests a high reactivity. By contrast, sediment at station NW had less TOC (0.3–1.2% dry weight) and deeper nitrate penetration (5–6 cm), which suggests a lower reactivity as compared to ML sediment (Table T2). It is possible to identify a few characteristic profile types (Fig. F2). In a first case, the nitrate concentration decreases uninflected with the sediment depth, and the curvature is typical for transport by diffusion. Examples are stations ML, EI, TS, and 196, which all have impermeable sediment. In a second case, a distinct nitrate maximum is present within the sediment, which indicates benthic nitrification (e.g. Pr-0309-STS, Pr-0909-FP, HE304-327, HE318-210). Shallow nitrification zones in 1 cm sediment depth might receive oxygen from the bottom water by diffusion. Deeper zones (3–5 cm) may suggest advective transport. Similar penetration depths under advection-dominated conditions are reported by Ahmerkamp et al. (2017) from in-situ measurements. A remarkable example is station Pr-0909-FP where bottom water nitrate was zero, while the peak concentration reached 5 μmol L− 1 due to internal nitrification thereby emphasizing the significance of coupled nitrification-denitrification in permeable sediment (Marchant et al., 2016). The depth of the internal nitrate peak exceeded 5 cm in Pr-0909SOS and 7 cm in HE318–199, respectively illustrating a constraint of the employed sediment sampling by means of a Multicorer. This sampling device tended to recover only short sediment cores from permeable, sandy sediment which restricts the length of the obtained pore water profiles. In the two samples above, the nitrate penetration exceeded the length of the recovered cores and the profiles are thus incomplete. Apart from these two examples, the pore water profiles shown in Fig. F2 generally indicate that the Rhizon sampling method is reliable to examine the penetration of nitrate since profiles of duplicate cores show good agreement (e.g. Pr0309-STS, Pr0909-KL, HE304-350). Profiles from offshore stations under summer conditions (HE-304) had very low concentrations of around 1 μmol L− 1, which was close to the detection limit.

porewater that is released from the denitrification layer directly into the bottom water. For the smaller volume flux unit at the nitrate penetration depth it appears that all bottom water nitrate has been consumed. By multiplication of unit with the bottom water nitrate concentration (C0), we construct a conservative estimate of the nitrate flux due to denitrification:

J = unit C0 =

κ K hm C0 2.5 exp (Lnit κ )

(4)

Estimates for wave number (k) and hydraulic conductivity (K) were derived from empirical functions of measured median grain size D50 of the sediment. The sediment permeability (κ) and K was calculated employing the method of Neumann et al. (2016). Bedform wave number was estimated according to Yalin (1985). The hydraulic head (hm) at the sediment surface was calculated according to Ahmerkamp et al. (2015) and Elliott and Brooks (1997) applying a mean bottom water velocity of 16 cm s− 1 based on recent measurements throughout the German Bight (Jana Friedrich, 18 deployments, unpublished data). 2.7. Average fluxes of sediment classes Based on permeability, the sediment within the study area was classified as impermeable (κ < 3 ∗ 10− 13 m2), moderately permeable (3 ∗ 10− 13 m2 < κ < 3 ∗ 10− 12 m2) and permeable (κ > 3 ∗10− 12 m2). The annually average of diffusive and advective nitrate consumption was calculated for each sediment class by grouping the flux estimates according to sediment class (impermeable, moderately permeable, and permeable) and transport mode (diffusion and advection). The impact of outliers was reduced by discarding the lowest and highest sixtiles (1/6) of each group, which keeps the central 2/3 of the data. 2.8. Maps The maps showing the spatial distribution of sampled stations and the values of flux estimates were composed using the software Ocean Data View (Schlitzer, 2011). 3. Results and discussion 3.1. Site characteristics

3.3. Nitrate consumption rates The majority of sampled sediment was sand with median grain sizes in the range 70–390 μm. The TOC content was in the range of 0.0–2.8% dry weight. Particles with low sinking velocity such as clay, silt and organic particles are concentrated in areas with reduced flow velocity, which permits the sedimentation of fine and light-weight particles. Such areas with elevated concentrations of fines and TOC are the shallow margins of the back-barrier tidal flats along the coast, the ‘Helgoland mud area’ (Hebbeln et al., 2003), and the outer German Bight along the submersed valley of the Palaeo-Elbe. By contrast, the sediment in areas with coarser sand contained very low TOC concentrations (Table T2). A dominant source of nitrate in the eastern part of the German Bight is the Elbe estuary that discharges brackish waters with nitrate concentrations as high as 300 μmol L− 1 during winter. The Elbe plume veers northwards and flows along the coast, so that Elbe nitrate is detectable in the eastern part of the German Bight (east of 7° E) during winter conditions (cruises Pr-0309, HE-318). During summer (cruises Pr-0909, HE-304), nitrate concentrations were generally low and elevated nitrate concentrations were restricted to the immediate vicinity of the Elbe estuary (Table T2).

Nitrate penetration depths ranged from 1 cm in muddy, impermeable sediments up to 8 cm in sandy, permeable sediments (Fig. F2, Table T2). In a first analysis we assume that molecular diffusion is the dominant transport mechanism and thus applied the method of Berg et al. (1998) to establish nitrate consumption rates based on the concentration gradients. We regarded the depth-integrated reaction rate as the nitrate consumption which may differ from the nitrate flux across the sediment water interface since we considered nitrate production by nitrification. Highest nitrate consumption rates were estimated for muddy sediments overlain by bottom water with high nitrate concentration (e.g. station Pr-0309-ML: 2800 μmol m− 2 d− 1, Table T2). The high rates correspond with the increased organic carbon content of 2.8% dry weight. Consequently the high nitrate consumption results in low nitrate penetration. Sediments low in TOC and otherwise similar characteristics with respect to grain size, temperature, and bottom water concentrations of nitrate and oxygen had significantly deeper nitrate penetration and decreased nitrate consumption rates (e.g. Pr0309-ML vs. Pr-0309-NW). Nitrate consumption was also low if the bottom water had a low nitrate concentration. As such nitrate consumption was virtually undetectable in post-bloom conditions when nitrate was depleted in bottom waters (e.g. HE304, stations 334–338). In a second analysis, we assume porewater advection to be the dominant transport mechanism. Especially in coarse grained sediments

3.2. Nitrate concentrations in pore water We measured pore water profiles of nitrate in the surface sediment 4

Journal of Sea Research xxx (xxxx) xxx–xxx

A. Neumann et al.

Table T2 Characteristics of sampled sites with median grain size D50, permeability κ, organic carbon content TOC, as well as nitrate penetration depth L, depth of internal nitrate peak, diffusioncontrolled nitrate consumption Jdiff, and advection-controlled nitrate consumption Jadvec. Values which were not determined are indicated with ‘n.d.’. Cruise

Station

D50 mm

κ 10− 12 m2

TOC % d.w.

NO3− BW μmol L− 1

NO3− L cm

Pr-0309

ML STS BR NW EI WW TS TR KL 327 334 335 336 337 338 339 341 343 350 351 353 354 ML STS BR NW KL TR TS FP SOS 177 180 181 182 191 195 196 197 199 209 210 212 214 216

0.10 0.39 0.09 0.08 0.11 0.16 0.11 0.15 0.15 0.15 0.15 0.16 0.15 0.14 0.15 0.16 0.12 0.29 0.10 0.14 0.21 0.25 0.07 0.39 0.09 0.07 0.12 0.15 0.11 0.12 0.27 0.18 0.16 0.18 0.18 0.14 0.22 0.11 0.11 0.26 0.17 0.17 0.21 0.19 0.16

0.06 98.83 0.03 0.02 0.09 0.72 0.11 0.51 0.52 0.54 0.44 0.79 0.57 0.41 0.53 0.67 0.15 19.21 0.04 0.37 2.97 8.29 0.01 98.83 0.03 0.01 0.13 0.61 0.08 0.17 12.55 1.58 0.80 1.32 1.59 0.30 4.66 0.08 0.08 9.76 1.03 1.07 3.51 2.08 0.83

2.80 0.42 1.55 0.34 0.57 0.08 0.81 0.21 0.14 0.06 0.02 0.31 0.25 0.20 0.12 0.05 0.39 0.29 0.35 0.23 0.15 0.02 3.51 0.04 0.10 1.22 0.25 0.36 0.14 0.02 0.02 0.18 0.02 0.01 0.00 0.04 0.02 0.28 0.19 n.d. 0.01 0.01 0.02 0.02 0.01

147.56 305.95 199.65 225.33 267.74 162.98 107.65 80.48 219.36 1.23 1.13 0.36 0.39 0.69 0.98 1.90 7.68 6.27 5.077 3.426 4.25 5.072 54.74 102.35 93.01 111.86 30.73 8.74 1.72 0.10 0.72 8.80 0.08 3.21 3.14 6.26 6.97 14.44 14.85 35.65 6.11 7.92 28.05 22.54 6.69

2 8 3 6 5 3 3 5 6 3 8 8 3 8 7 4 1 5 1 3 3 5 1 8 5 5 5 2 4 6 >5 4 4 7 4 3 3 2 4 >7 3 4 4 7 2

HE-304

Pr-0909

HE-318

NO3− peak cm

3

2

3 1 3

1

2 5

3 >5 1 3

1

>7 2 5 1

Jdiff μmol/m2 d

Jadvec μmol/m2 d

2844 349 857 337 2021 1137 93 106 131 2 0 3 0 1 0 1 16 1 25 2 4 6 837 18 158 1159 109 26 10 3 n.d. 16 4 5 5 5 10 52 6 n.d. 22 8 9 20 8

8 32,186 5 1 7 173 18 35 43 1 0 0 0 0 0 1 2 85 1 1 16 34 1 10,768 1 0 2 12 0 0 n.d. 12 0 2 3 3 64 2 1 n.d. 11 11 134 25 15

T2) with maximum rates of 32,200 μmol N m2 d− 1. Direct measurements of benthic nitrate consumption under advective conditions are scarce for the German Bight. Recently, Marchant et al. (2016) estimated denitrification in permeable sediment (κ = 2–7 ∗ 10− 11 m2) within our study area and report denitrification rates in the range 520–2280 μmol m2 d− 1, which agrees well with our results (Fig. F3B). In addition, they estimated nitrate penetration depths in the range of 1–5 cm, which also compares well with our measurements. Gao et al. (2012) analyzed nitrate consumption in permeable sediment (7–9 ∗ 10− 12 m2) of an intertidal sand flat and reported nitrate consumption rates in the range 230–26,400 μmol m2 d− 1, which further supports our high estimates for permeable sediment (e.g. Pr-0309-STS, Table T2). By comparison of sites with different permeability and otherwise similar conditions with respect to temperature and bottom water concentrations of oxygen and nitrate, the highest advection- fueled nitrate consumption is one order of magnitude higher than the highest rate enabled by diffusion. Examples are the muddy site Pr-0309-ML (2800 μmol m− 2 d− 1) vs. sandy site Pr-0309-STS (32,200 μmol m− 2 d− 1) or muddy site Pr-0909-ML (840 μmol m− 2 d− 1) vs. sandy site Pr-0909-STS (10,800 μmol m− 2 d− 1). This indicates that advective transport in permeable sediments is not just replacing molecular diffusion but that advection sustains substantially higher consumption rates than diffusive transport.

advective solute transport has the potential to exceed diffusive transport by several orders of magnitude. We applied Eq. (4) to estimate the advective net nitrate flux into the sediment. As outlined in Section 2.6, we conservatively estimated the nitrate consumed in the denitrification layer from the modeled volume flux at the nitrate penetration depth, assuming that only bottom water nitrate enters the denitrification layer. This approach is conservative for several reasons. Firstly, it neglects nitrification within the sediment, which can significantly increase the total N-loss in permeable sediment as shown by Marchant et al. (2016). Secondly, our approach neglects the nitrate depletion of the porewater fraction that does not cross the nitrate penetration depth but is flushed from denitrification zone directly into the water column. Finally, we consider only the average tidal and residual currents as sources of the pressure field. The pressure gradients and thereby the porewater volume flux may be enhanced by the oscillatory flow of surface waves (e.g. Precht and Huettel, 2003). Diffusive and advective nitrate uptake was calculated for all stations irrespective of the estimated permeability. The advective uptake rates at stations with quasi impermeable sediments (κ < < 10− 12 m2) were reduced to 1–18 μmol N m− 2 d− 1. The advective uptake increased substantially in permeable sediments and exceeded 103 μmol N m2 d− 1 (Table 5

Journal of Sea Research xxx (xxxx) xxx–xxx

A. Neumann et al.

Fig. F2. Pore water concentrations of sediment from German Bight, North-Frisian Wadden Sea and Elbe Estuary. If pore water was sampled from duplicate cores then the second profile is represented by open circles.

permeability, and advective pore water exchange is virtually absent. Concomitantly, this sediment typically has a high reactivity and low nitrate penetration depth, which results in steep gradients and high

The widely differing estimates assuming either sole diffusion or advection appear conflicting, but both are actually complementary. Muddy sediment generally has a low grain size and very low 6

Journal of Sea Research xxx (xxxx) xxx–xxx

A. Neumann et al.

diffusive fluxes (Fig. F3A). Coarser, sandy sediments on the other hand exhibit lower reactivity and deeper nitrate penetration with less steep concentration gradients, which reduces any diffusive nitrate uptake. However, this coarser sediment is substantially more permeable and the lower reactivity is compensated by the fast advective nitrate influx (Fig. F3B). The sediment types analyzed in our study represent a continuum with respect to permeability, and both processes, diffusion and advection, occur simultaneously. We thus propose that the true nitrate consumption rate is the sum of both rates. The ratio of the two processes is expressed as the Péclet number, which was calculated as the ratio of the advection-controlled nitrate consumption (Jadvec) over the diffusioncontrolled consumption (Jdiff). The calculated Péclet numbers of our flux estimates are < 1 for sediment with permeability < 10− 12 m2 and > 1 for sediment with permeability > 10− 12 m2 (Fig. F3 C), which perfectly matches the accepted classification of impermeable and permeable sediment with a permeability of 10− 12 m2 as the threshold (e.g., Huettel et al., 2014). 3.4. Estimation of reactive nitrogen removal in German Bight sediment Benthic nitrate consumption clearly depends on the nitrate availability from the bottom water. During periods of high riverine nitrate loads (winter), high nitrate concentrations in the range of 10–60 μmol L− 1 are found in the Elbe plume northwards along the North-Frisian Coast (van Beusekom et al., 2008) and enable high rates of benthic nitrate consumption. The highest rates were observed in the Elbe Estuary at nitrate concentration of around 200–300 μmol L− 1. West of the Elbe Plume (west of 7° E) in the open German Bight, nitrate consumption rates decrease with the nitrate availability (Fig. F4 A, C). During summer when riverine discharge is low, substantial nitrate concentrations are restricted to the Elbe Estuary and are low in the western German Bight (< 1 μmol L− 1). Consequently, high nitrate consumption rates are only found in the Elbe Estuary (Fig. F4 B, C). Advection-fueled nitrate consumption dominates only in clearly permeable sediment (Fig. F4, dark grey shaded areas). We have demonstrated that the relative importance of nitrate consumption driven by diffusion and advection varies with sediment permeability (Fig. F3). For assessing the relative contribution of different sediment types to the total nitrate consumption in our study area, we defined three sediment classes. Here we expand the dual classification in either permeable or impermeable (e.g. Huettel et al., 2014) by introducing a transition interval between permeable and impermeable as a third sediment class. This new class is termed ‘moderately permeable’ and spans the permeability interval 3 ∗ 10− 13 − 3 ∗ 10− 12 m2 where both transport modes contribute equally to nitrate consumption. Hence, sediment classified as ‘impermeable’ in the following has a permeability < 3 10− 13 m2 and is characterized by a dominance of diffusion over advection. Likewise, sediment in the class ‘permeable’ has a permeability > 3 ∗ 10− 12 m2 where advection clearly dominates over diffusion. From recent observations of sediment permeability (Neumann et al., 2016) we conclude that diffusion controlled, impermeable sediment is present in 24% of our study area (Fig. F1, dashed outline, 57,000 km2), moderately permeable sediment accounts for 39%, and clearly permeable sediment is present at 37% of the study area (Table T3). Using the average of the nitrate consumption rates of each sediment class (Table T3), we estimate an annually averaged nitrate consumption of 3 ∗ 106 mol N d− 1 in impermeable sediment, mainly fuelled by diffusion. Moderately permeable sediment consumed 3 ∗ 105 mol N d− 1, while permeable sediment removed 5 ∗ 107 mol N d− 1, mainly fuelled by advection. This implies that among the different sediment types especially the permeable sediment is the most efficient nitrate sink accounting for up to 90% (5.2 ∗ 107 mol N d− 1) of the total benthic nitrate consumption (5.5 ∗ 107 mol N d− 1). The presence of internal nitrification may further enable substantial rates of nitrate consumption even when the bottom water concentration of nitrate is low (Marchant et al., 2016).

Fig. F3. Calculated benthic nitrate consumption assuming either diffusive (A) or advective (B) transport in a sediment permeability continuum. Impermeable sediment has a permeability < 10− 12 m2, permeable sediment has a permeability > 10− 12 m2. Results from Marchant et al. (2016) are represented by open diamonds. The Péclet number (C) represents the ratio of advective and diffusive transport.

7

Journal of Sea Research xxx (xxxx) xxx–xxx

A. Neumann et al.

Fig. F4. Map of benthic nitrate consumption fuelled by diffusion (blue) and advection (orange) during winter conditions (A, C) with high riverine nitrate load, and during summer conditions (B, C) with low riverine nitrate load. Isolines indicate sediment permeability: impermeable (white), moderately permeable (light grey), permeable (dark grey). Note that panels A & B have different scale than panels C & D. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

We compared the estimated daily N-loss of the study area (57.000 km2) in the German Bight with the daily nitrate input of the Elbe river, calculated from the Elbe discharge volume (Anonymous, 2009) and corresponding measured nitrate concentrations. For winter conditions during February and March we calculated an average nitrate input of 3 ∗ 107 mol N d− 1, while the nitrate input in summer is decreased to 1 ∗ 106 mol N d− 1. The calculated nitrate input suggests that the permeable sediment in our study area has the potential to substantially reduce the nitrate load during winter, while during summer, the permeable sediment has the potential to completely consume the riverine nitrate input. Denitrification measurements in the German Bight are rare and a

major challenge has been to reconcile measurements and budgets (e.g. Beddig et al., 1997; van Beusekom et al., 1999). Beddig et al. (1997) made an N budget for the German Bight based on cruise data for 1989–1991 and noted that the budget was not balanced. Their budget did not account for an N-deficit of about 6 Gmol y− 1, while available measurements of benthic denitrification at that time (Wadden Sea plus German Bight) supported a total denitrification of about 1.5 Gmol N y− 1. The various incubation methods used to quantify denitrification (e.g. the Acetylen Blocking Method) may underestimate the true N-fluxes to a differing extent as discussed by Lohse et al. (1996) and van Beusekom et al. (1999). On top of that, the incubation approach itself may play a role: Most incubations that have been used in

Table T3 Area of sediment permeability classes in the German Bight and truncated average nitrate consumption fueled by molecular diffusion (Jdiff) and pore water advection (Jadvec) for each sediment class. Sediment type

Area km2

Mean Jdiff μmol m− 2 d− 1

Mean Jadvec μmol m− 2 d− 1

Total diffusion mol d− 1

Total advection mol d− 1

Impermeable κ < 3 ∗ 10− 13 m2 Mod. permeable Permeable κ > 3 ∗ 10− 12 m2 Weighted average total

12,200 19,600 18,800 50,600

249 8 11 68

2 7 2762 1027

3.1 ∗ 106 1.6 ∗ 105 2.0 ∗ 105 3.4 ∗ 106

2.1 ∗ 104 1.5 ∗ 105 5.2 ∗ 107 5.2 ∗ 107

8

Journal of Sea Research xxx (xxxx) xxx–xxx

A. Neumann et al.

MH: Data analysis, preparation of the manuscript. KE: Preparation of the manuscript.

the German Bight so far exclude hydrodynamic forcing of advective pore water exchange (Lohse et al., 1993, 1996; Deek et al., 2012). Only recently, methods used by Gao et al. (2012), and Marchant et al. (2014, 2016) considered pore water advection in permeable sediments which resulted in higher denitrification rates compared to previous estimates based on diffusion-dominated methods. Given the methodological issues, budget approaches may help to constrain N budgets (e.g. Luijn van et al., 1996). Hydes et al. (1999) estimated an average denitrification rate of about 260 mmol m− 2 y− 1 for the North Sea, based on budget considerations (~ 6.5 Gmol y− 1 for German Bight and adjacent Wadden Sea), which is in line with the above mentioned nitrogen budget of the German Bight. The average denitrification rates in the German Bight derived in the present study of about 375 mmol m− 2 y− 1 compare favourably with both, the BeddigBudget and the Hydes-Budget. Given the good agreement between previous measurements and our estimates under diffusive conditions, the discrepancy between the above N budgets and available denitrification measurements is probably due to a previous underestimation of denitrification/nitrate consumption under advective conditions. The Péclet number points out to which extent diffusion-dominated methods of flux measurements underestimate the true flux in permeable sediment, which might be as much as 100-fold (Fig. F3 C). For example, Deek et al. (2012) employed a stagnant wholecore incubation method to examine North-Frisian Wadden Sea sediment, and found consistently higher denitrification in the impermeable sediment (0.7–6.2 mmol N m− 2 d− 1) than in the permeable sand (0.3–1.4 mmol N m− 2 d− 1). Marchant et al. (2014) examined permeable sand from the East-Frisian Wadden Sea with a percolation method to enable pore water advection, and found significantly higher denitrification rates in the range of 10–27 mmol N m− 2 d− 1 (top 5 cm of the sediment). These high rates under advective conditions are in line with the present denitrification rates of up to 32 mmol N m− 2 d− 1 in permeable sediments and compare well with estimates from other sandy shelf regions such as the Mauritanian shelf (up to 9 mmol N m− 2 d− 1, Sokoll et al., 2016) and the Great Barrier Reef (up to 16 mmol N m− 2 d− 1, Santos et al., 2012). Approximately one third of the German Bight sediment is classified here as permeable and accounts for up to 90% of the total benthic nitrate consumption (Table T3).

Notation C0 D50 J hm k K Κ L Lnit Lox m m′ m* m*′ t TOC t* u unit uox

initial solute concentration, in mol/m3 median grain size, in m solute flux, in mol/m2s hydraulic head amplitude at the sediment surface, in m bed form wave number, in 1/m hydraulic conductivity of the sediment, in m/s intrinsic permeability, in m2 averaged sediment penetration depth, in m averaged nitrate penetration depth, in m averaged penetration depth, in m mass transfer per unit bed area, in m instantaneous net mass transfer per unit bed area, in m/s normalized value of m, unitless normalized value of m′, unitless time scale, in s total organic carbon normalized value of t, unitless averaged velocity of a solute front in the sediment, in m/s averaged velocity of the nitrate front, in m/s averaged velocity of the oxygen front, in m/s

Acknowledgements We wish to thank the crews and scientific parties of the research vessels “Ludwig Prandtl” and “Heincke” for their excellent support (Grant No AWI-HE304_00 and AWI-HE318_00). Soeren Ahmerkamp is acknowledged for valuable comments on an earlier draft. This study received financial support by the German National Science Foundation DFG (Em 37/29). Appendix A. Supplementary data The porewater data are available online in the Pangaea database via http://dx.doi.org/http://dx.doi.org/10.1594/PANGAEA.876462.

4. Conclusions References We present nitrate pore water profiles in various North Sea sediments exhibiting a broad range of permeabilities, carbon contents and for various bottom water nitrate concentrations. We used these profiles to estimate the advection-fuelled nitrate consumption by employing an extended model of pore water advection based on Elliott and Brooks (1997) and compared these fluxes with results from the established diffusion-based method of Berg et al. (1998). Our results confirm the previous classification stating that molecular diffusion dominates at sediment permeability < 10− 12 m2, while pore water advection is the dominant transport mode at permeability > 10− 12 m2. In similar conditions with respect to temperature and bottom water concentrations of oxygen and nitrate, the nitrate consumption rates fuelled by advection in permeable sediment were one order of magnitude higher than the rates sustained by diffusion in impermeable sediment. Although sediment with permeability > 3 ∗ 10− 12 m2 accounts only for one third of the German Bight sediment, it accounts for up to 90% of the total benthic nitrate consumption of this area. In summary, permeable sediments appear as the most significant sink for reactive nitrogen in the German Bight.

Ahmerkamp, S., Winter, C., Janssen, F., Kuypers, M.M.M., Holtappels, M., 2015. The impact of bedform migration on benthic oxygen fluxes. J. Geophys. Res. Biogeosci. 120, 2229–2242. Ahmerkamp, S., Winter, C., Krämer, K., de Beer, D., Janssen, F., Friedrich, J., Kuypers, M.M.M., Holtappels, M., 2017. Regulation of benthic oxygen fluxes in permeable sediments of the coastal ocean. Limnol. Oceanogr. http://dx.doi.org/10.1002/lno. 10544. Amann, T., Weiss, A., Hartmann, J., 2012. Carbon dynamics in the freshwater part of the Elbe estuary, Germany: implications from improving water quality. Estuar. Coast. Shelf Sci. 107, 112–121. Anonymous, 2009. Deutsches Gewässerkundliches Jahrbuch, Elbegebiet, Teil 3, 2009. Hamburg Port Authority, Freie und Hansestadt Hamburg (ISSN 0949-3654). Beddig, S., Brockmann, U., Dannecker, W., Körner, D., Pohlmann, T., Puls, W., 1997. Nitrogen fluxes in the German Bight. Mar. Pollut. Bull. 34, 382–394. Berg, P., Risgaard-Petersen, N., Rysgaard, S., 1998. Interpretation of measured concentration profiles in sediment pore water. Limnol. Oceanogr. 43 (7), 1500–1510. Beusekom van, J.E.E., Brockmann, U.H., Hesse, K.-J., Hickel, W., Poremba, K., Tillmann, U., 1999. The importance of sediments in the transformation and turnover of nutrients and organic matter in the Wadden Sea and German Bight. Ger. J. Hydrogr. 51 (2/3), 245–266. Beusekom van, J.E.E., Loebl, M., Martens, P., 2008. Long-term variability of winter nitrate concentrations in the northern Wadden Sea driven by freshwater discharge, decreasing riverine loads and denitrification. Helgol. Mar. Res. 62 (1), 49–57. Beusekom van, J.E.E., Loebl, M., Martens, P., 2009. Distant riverine nutrient supply and local temperature drive the long-term phytoplankton development in a temperate coastal basin. J. Sea Res. 61, 26–33. Blott, S.J., Pye, K., 2001. GRADISTAT: a grain size distribution and statistics package for the analysis of unconsolidated sediments. Earth Surf. Process. Landf. 26 (11), 1237–1248. Cadée, G.C., Hegeman, J., 2002. Phytoplankton in the Marsdiep at the end of the 20th century; 30 years monitoring biomass, primary production, and Phaeocystis blooms.

Contributions AN: Sampling, nutrient analysis, data analysis, preparation of the manuscript. JvB: Nutrient analysis, additional field work, preparation of the manuscript. 9

Journal of Sea Research xxx (xxxx) xxx–xxx

A. Neumann et al.

mass balance approach. Water Res. 30, 893–900. Marchant, H.K., Lavik, G., Holtappels, M., Kuypers, M.M.M., 2014. The fate of nitrate in intertidal permeable sediments. PLoS One 9 (8), e104517. http://dx.doi.org/10. 1371/journal.pone.0104517. Marchant, H.K., Holtappels, M., Lavik, G., Ahmerkamp, S., Winter, C., Kuypers, M.M.M., 2016. Coupled nitrification – denitrification leads to extensive N loss in subtidal permeable sediments. Limnol. Oceanogr. http://dx.doi.org/10.1002/lno.10271. Meijboom, F.W., van Noordwijk, M., 1992. Rhizon soil solution samplers as artificial roots. In: Kutschera, L. (Ed.), Root Ecology and its Practical Application.- Proc 3rd ISSR Symp., 2.-6. Sep 1991, Wien Austria, pp. 793–795. Middelburg, J.J., Soetaert, K., Herman, P.M.J., Heip, C., 1996. Denitrification in marine sediments: a model study. Glob. Biogeochem. Cycles 10 (4), 661–673. Neumann, A., Möbius, J., Hass, H.C., Puls, W., Friedrich, J., 2016. Permeability of surface sediments in the German Bight (North Sea). J. Sea Res. http://dx.doi.org/10.1016/j. seares.2016.12.002. OSPAR, 2010. Quality Status Report 2010. OSPAR Commission, London (176 pp). Pätsch, J., Serna, A., Dähnke, K., Schlarbaum, T., Johannsen, A., Emeis, K.-C., 2010. Nitrogen cycling in the German Bight (SE North Sea) - Clues from modelling stable nitrogen isotopes. Cont. Shelf Res. 30, 203–213. Philippart, C.J.M., Beukema, J.J., Cadée, G.C., Dekker, R., Goedhart, P.W., Iperen, J.M.V., Leopold, M.F., Herman, P.M.J., 2007. Impact of nutrients on coastal communities. Ecosystems 10, 95–118. Precht, E., Huettel, M., 2003. Advective pore-water exchange driven by surface gravity waves and its ecological implications. Limnol. Oceanogr. 48 (4), 1674–1684. Precht, E., Franke, U., Polerecky, L., Huettel, Markus, 2004. Oxygen dynamics in permeable sediments with wave-driven pore water exchange. Limnol. Oceanogr. 49 (3), 693–705. Radach, G., Patsch, J., 2007. Variability of continental riverine freshwater and nutrient inputs into the North Sea for the years 1977–2000 and its consequences for the assessment of eutrophication. Estuar. Coasts 30, 66–81. Reise, K., Siebert, I., 1994. Mass occurrence of green algae in the German Wadden Sea. Deutsche Hydrographische Zeitschrift 1, 171–188. Santos, I.R., Eyre, B.D., Glud, Ronnie N., 2012. Influence of porewater advection on denitrification in carbonate sands: evidence from repacked sediment column experiments. Geochim. Cosmochim. Acta 96 (1), 247–258. Schlitzer, R., 2011. Ocean Data View. http://odv.awi.de. Schroeder, F., Wiltshire, K.H., Klages, D., Mathieu, B., Blöcker, G., Knauth, H.D., 1996. Nitrogen and oxygen processes in sediments of the Elbe estuary. Arch. Hydrobiol. Suppl. 110, 311–328. Seeberg-Elverfeldt, J., Schlüter, M., Feseker, T., Kölling, M., 2005. Rhizon sampling of pore waters near the sediment/water interface of aquatic systems. Limnol. Oceanogr. Methods 3, 361–371. Seitzinger, S., Harrison, J.A., Bohlke, J.K., Bouwman, A.F., Lowrance, R., Peterson, B., Tobias, C., Van Drecht, G., 2006. Denitrification across landscapes and waterscapes: a synthesis. Ecol. Appl. 16 (6), 2064–2090. Sokoll, S., Lavik, G., Sommer, S., Goldhammer, T., Kuypers, M.M.M., Holtappels, M., 2016. Extensive nitrogen loss from permeable sediments off North-West Africa. J. Geophys. Res. Biogeosci. 121. http://dx.doi.org/10.1002/2015JG003298. Westernhagen von, H., Hickel, W., Bauerfeind, E., Niermann, U., Kröncke, I., 1986. Sources and effects of oxygen deficiencies in the south-eastern North Sea. Ophelia 26, 457–473. Yalin, M., 1985. On the Determination of Ripple Geometry. J. Hydraul. Eng. 111, 1148–1155.

J. Sea Res. 48, 97–110. Cornwell, J.C., Kemp, W.M., Kana, T.M., 1999. Denitrification in coastal ecosystems: Methods, environmental controls, and ecosystem level controls, a review. Aquat. Ecol. 33, 41–54. Costanza, R., d'Arge, R., de Groot, R., Farber, S., Grasso, M., Hannon, B., Limburg, K., Naeem, S., O'Neill, R.V., Paruelo, J., Raskin, R.G., Sutton, P., van den Belt, M., 1997. The value of the world`s ecosystem services and natural capital. Nature 387, 253–260. Deek, A., Emeis, K., van Beusekom, J., 2012. Nitrogen removal in coastal sediments of the German Wadden Sea. Biogeochemistry 108, 467–483. Dolch, T., Buschbaum, C., Reise, K., 2013. Persisting intertidal seagrass beds in the northern Wadden Sea since the 1930s. J. Sea Res. 82, 134–141. Elliott, A.H., Brooks, N.H., 1997. Transfer of nonsorbing solutes to a streambed with bed forms: theory. Water Resour. Res. 33 (1), 123–136. Emeis, K.-C., van Beusekom, J., Callies, U., Ebinghaus, R., Kannen, A., Kraus, G., Kröncke, I., Lenhart, H., Lorkowski, I., Matthias, V., Möllmann, C., Pätsch, J., Scharfe, M., Thomas, H., Weisse, R., Zorita, E., 2015. The North Sea - a shelf sea in the Anthropocene. J. Mar. Syst. 141, 18–33. Gao, H., Matyka, M., Liu, B., Khalili, A., Kostka, J.E., Collins, G., Jansen, S., Holtappels, M., Jensen, M.M., Badewien, T.H., Beck, M., Grunwald, M., de Beer, D., Lavik, G., Kuypers, M.M.M., 2012. Intensive and extensive nitrogen loss from intertidal permeable sediments of the Wadden Sea. Limnol. Oceanogr. 57 (1), 185–198. Grasshoff, K., Erhardt, M., Kremling, K., 1983. Methods of Seawater Analysis. Verlag Chemie, Weinheim. Hebbeln, D., Scheurle, C., Lamy, F., 2003. Depositional history of the Helgoland mud area, German Bight, North Sea. Geo-Mar. Lett. 23, 81–90. Hickel, W., Mangelsdorf, P., Berg, J., 1993. The human impact in the German Bight: eutrophication during three decades (1962–1991). Helgoländer Meeresun. 47, 243–263. Huettel, M., Ziebis, W., Forster, S., 1996. Flow-induced uptake of particulate matter in permeable sediments. Limnol. Oceanogr. 41 (2), 309–322. Huettel, M., Ziebis, W., Forster, S., Luther III, G.W., 1998. Advective transport affecting metal and nutrient distributions and interfacial fluxes in permeable sediments. Geochim. Cosmochim. Acta 62 (4), 613–631. Huettel, M., Berg, P., Kostka, J.K., 2014. Benthic exchange and biogeochemical cycling in permeable sediments. Annu. Rev. Mar. Sci. 2014 (6), 23–51. Hydes, D.J., Kelly-Gerreyn, B.A., Le Gall, A.C., Proctor, R., 1999. The balance of supply of nutrients and demands of biological production and denitrification in a temperate latitude shelf sea - a treatment of the southern North Sea as an extended estuary. Mar. Chem. 68 (1–2), 117–131. van Katwijk, M.M., Vergeer, L.T.H., Schmitz, G.H.W., Roelofs, J.G.M., 1997. Ammonium toxicity in eelgrass Zostera marina. Mar. Ecol. Prog. Ser. 157, 159–173. Lohse, L., Malschaert, J.F.P., Slomp, C.P., Helder, W., van Raaphorst, W., 1993. Nitrogen cycling in North Sea sediments: interaction of denitrification and nitrification in offshore and coastal areas. Mar. Ecol. Prog. Ser. 101, 283–296. Lohse, L., Kloosterhuis, R., van Raaphorst, W., Helder, W., 1996. Denitrification rates in continental shelf sediments of the North Sea: acetylene block technique versus isotope pairing. Mar. Ecol. Prog. Ser. 132, 169–179. Los, F.J., Troost, T.A., van Beek, J.K.L., 2014. Finding the optimal reduction to meet all targets - Applying Linear Programming with a nutrient tracer model of the North Sea. J. Marine Syst. 131, 91–101. Luijn van, F., Boers, P.C.M., Lijklema, L., 1996. Comparison of denitrifcation rates in lake sediments obtained by N2 flux method, the 15N isotope paring technique and the

10