Seasonality of Organic Matter Degradation Regulates Nutrient and Metal Net Fluxes in a High Energy Sandy Beach

During seawater circulation in permeable intertidal sands, organic matter degradation alters the composition of percolating fluids and remineralization products discharge into surficial waters. Concurrently, coastal seawater nutrient and organic matter composition change seasonally due to variations in pelagic productivity. To assess seasonal changes in organic matter degradation in the intertidal zone of a high energy beach (Spiekeroog Island, southern North Sea, Germany), we analyzed shallow pore waters for major redox constituents (oxygen [O2], manganese [Mn], and iron [Fe]) and inorganic nitrogen species (nitrite [NO2−], nitrate [NO3−], and ammonium [NH4+]) in March, August, and October. Surface water samples from a local time series station were used to monitor seasonal changes in pelagic productivity. O2 and NO3− were the dominating pore water constituents in March and October. Dissolved Mn, Fe, and NH4+ were more widely distributed in August. Seasonal changes in seawater temperature as well as organic matter and nitrate supply by seawater were assumed to affect microbial rates and degradation pathways. Pore water and seawater variability led to seasonally changing constituent effluxes to surface waters. Mn, Fe, and NH4+ effluxes are minimal in March and reached their maximum in August. Furthermore, the intertidal sands switched from a net dissolved inorganic nitrogen sink in March to a net source in August. In conclusion, seasonal effects on intertidal pore water biogeochemistry affect constituent fluxes across the sediment‐water interface. The seasonality of the beach bioreactor must be considered when fluxes are extrapolated to annual timescales.


Introduction
Worldwide, about one third of ice-free shorelines are covered by sandy beaches (Luijendijk et al., 2018) and many are exposed to mesotidal to macrotidal conditions (Flemming, 2005). The subterranean estuaries of these high energy beaches are complex and dynamic environments due to pore water advection. Wave pumping and bottom water currents cause advective pore water flow within surficial sediments, while tidal pumping induces advective flow into deeper layers (Santos et al., 2012). Combined with fresh groundwater discharge, these factors often shape a seawater circulation cell (Robinson et al., 2018;Santos et al., 2012).
AHRENS ET AL.
1 of 21 Advective flow of seawater through sediments supplies benthic microbial communities with oxygen (O 2 ) and nitrate (NO 3 − ), as well as particulate and dissolved organic matter, thus leading to high rates of organic matter degradation and nutrient turnover (Anschutz et al., 2009;Huettel & Rusch, 2000b;Kim et al., 2019). As such, coastal permeable sediments can contribute substantially to marine organic carbon remineralization, thereby affecting coastal nutrient and metal budgets (Huettel et al., 1998;Huettel et al., 2014;Van Raaphorst et al., 1990).
Organic matter degradation pathways in intertidal beach sediments may be restricted to aerobic respiration and NO 3 − reduction (Charbonnier et al., 2013) but can also include Mn and Fe reduction as well as sulfate reduction (McAllister et al., 2015). Rates of organic matter degradation coupled to aerobic respiration are usually high compared to other metabolisms (Froelich et al., 1979;Jørgensen & Sørensen, 1985) but are restricted to the layer where O 2 is available. This region is shaped by O 2 consumption and resupply and represents an important redox boundary with respect to nitrogen transformation and iron (Fe) and manganese (Mn) cycling (Froelich et al., 1979;Shaw et al., 1990).
Nitrate reduction in subterranean estuaries can be driven by inputs of NO 3 − from fresh groundwater and inputs of marine organic matter (Couturier et al., 2017;Kim et al., 2017) as well as by seawater supplied NO 3 − . In combination with coupled nitrification-denitrification, the reduction of NO 3 − to nitrogen gas through denitrification can lead to substantial N-loss as N 2 and N 2 O, which has been observed for permeable intertidal and subtidal sediments (Marchant et al., 2016;Marchant et al., 2018). The net removal of dissolved inorganic nitrogen (DIN) in beaches may therefore affect nutrient budgets in coastal regions where inputs of NO 3 − are high due to anthropogenic activities (Johannsen et al., 2008).
Further, intense Fe and Mn cycling has been observed in intertidal sediments McAllister et al., 2015;O'Connor et al., 2015;Reckhardt et al., 2017;Roy et al., 2010). Pore waters can be enriched in these dissolved metals by orders of magnitude compared to seawater and thus form a potential source for coastal waters.
Redox zonation and transformations in intertidal sediments vary not only in space but also in time. Seasonal variations in hydraulic head lead to changes in seawater residence times and magnitude of fresh groundwater input, which impact pore water geochemistry Michael et al., 2005). Similarly, across the year there are changes in organic matter composition and quantity as well as the quantity of nutrients and metals supplied from the water column (Kowalski et al., 2012;Meier et al., 2015;Seidel et al., 2015).
The present study focuses on redox constituents in sediments influenced by mesotidal conditions at a high energy beach on Spiekeroog Island, Germany, located in the southern North Sea. Previous work at this site has indicated that the whole span of organic matter degradation pathways occurs, from aerobic respiration to sulfate reduction Reckhardt et al., 2017;Waska et al., 2019).
In the comprehensive study by Beck et al. (2017), key processes, affecting beach pore water composition, were identified and subsurface biogeochemistry was linked to sedimentological, hydrogeological, and microbiological data. However, the spatial sampling resolution in the intertidal zone was low and sampling was conducted only once, in spring. Concurrently, samples from these sampling locations were analyzed for trace metals by Reckhardt et al. (2017). More recently, seasonal fluctuations in temperature, organic matter availability, and inland hydraulic head were identified as drivers for the seasonality in pore water concentrations of dissolved O 2 and reduced Fe (Waska et al., 2019). It was hypothesized that the general seasonal patterns are blurred due to continuous sediment redistribution occurring in the intertidal zone, which affects infiltration zones of seawater and exfiltration zones of pore waters. The conclusions were based on salinity, dissolved O 2 , reduced Fe, and fluorescent organic matter analyses at 50 and 100 cm below surface. As benthic processes in intertidal sediments of Spiekeroog Island are linked to pelagic activity , and geochemical data from intertidal pore waters indicate that seasonal effects play a role in determining benthic processes Waska et al., 2019), it is likely that seasonal variations may control electron acceptors and remineralization products and their transport across the sediment-water interface. Here, therefore, the following hypotheses were tested: 1. Advective pore water transport influences the succession of organic matter degradation processes in intertidal sands of a high energy beach by controlling the supply with seawater derived organic matter and electron acceptors.  Röper et al. (2012). The location of the time series station for pelagic water column sampling is indicated by a black star. The light grey colors in the northern part of Spiekeroog Island indicate primary dune distribution. 2. Seasonal variations in coastal seawater composition affect benthic microbial activity and organic matter degradation pathways leading to seasonally changing spatial distributions of electron acceptors and remineralization products in shallow intertidal sands. 3. This seasonal variability in pore water biogeochemistry results in changing net fluxes across the sediment-water interface.
To assess these hypotheses, we monitored the pelagic water composition bimonthly and sampled the shallow intertidal pore waters in autumn, early spring, and summer. In detail, we studied dissolved O 2 , DIN species (nitrite (NO 2 − ), nitrate (NO 3 − ), ammonium (NH 4 + )), dissolved Mn, and dissolved Fe. The spatial sampling resolution was increased within the intertidal zone, compared to previous studies, in order to distinguish spatial from temporal dynamics and to deduce microbial processes from pore water constituent distributions.

Study Site
The island of Spiekeroog is part of a barrier island chain located in the southern North Sea, NW Germany ( Figure 1a). The sampling site is located on the north beach ( Figure 1b). Based on a wave categorization scheme for this region, the near-shore significant wave height is 1.5 m (Dette, 1977). About 10-30 days exhibit storm conditions, annually (Antia, 1995). The site is influenced by mesotidal conditions with a mean tidal range of 2.7 m. During neap and spring tidal cycles, the tidal range changes to 2.4 and 3.3 m, respectively. The beach sediments of the upper meters are composed of fine to medium sand (mean grain size: 170-320 μm; Beck et al., 2017). The western part of the island is underlain by a freshwater lens (Figure 1b), which is confined by clay layer at about 40 m below NHN (standard elevation zero) and exhibits water ages up to 51 years (Röper et al., 2012;Seibert et al., 2018). Due to varying flow paths, fresh groundwater of various ages is delivered to the intertidal subsurface .
Spiekeroog beach exhibits morphological changes on different timescales, including erosional events like storm surges, sediment accumulation during calm conditions, and seasonal ridge migration. From spring to late autumn a pronounced ridge and runnel morphology evolves (Flemming & Davis, 1994;Waska et al., 2019).
According to a previously published hydrogeological model and high-resolution data of the intertidal area of the study site, the intertidal zone can be subdivided into two infiltration zones and two exfiltration zones ( Figure 1c; Beck et al., 2017;Waska et al., 2019): Seawater infiltrates at the high water line (HWL) and discharges in the exfiltration zone in the runnel after weeks to months of residence time within the upper circulation cell. Concurrently, seawater infiltrates at the ridge and drains both landward and seaward. A mixture of upper circulation cell and ridge circulation pore water discharges in the runnel exfiltration zone, whereas ridge circulating pore water and potentially saltwater wedge water discharge at the exfiltration zone at low water line (LWL). Throughout the intertidal zone, but predominantly in low elevation areas, fresh groundwater mixes with circulating seawater, leading to locally restricted patches of brackish discharge (Waska et al., 2019).

Field Sampling
Samples were retrieved during three sampling campaigns in October 2016 (spring tide), March 2017 (neap tide), and August 2017 (spring tide). Six stations were sampled across coast-perpendicular transects extending from HWL to LWL (Figure 1c). Sampling was conducted around low tide in October and March and around high tide in August. The difference in tidal stage during sampling was unavoidable in order to meet the requirements of daylight, intertidal zone accessibility, and fair weather conditions. Additionally, a highresolution transect was sampled in March and August ranging from the HWL to the landward slope of the ridge (Figure 1c). In August, the sampling strategy was further extended and pore waters were extracted using a grid pattern (~200 × 180 m; 192 samples; Waska et al., 2019). Pore water samples were collected at 10, 30, 50, and 100 cm below surface (cmbsf) along transects and at 50 and 100 cmbsf in the grid pattern. Stainless steel push-point samplers were used to sample pore waters. After extracting water with polyethylene syringes, dissolved O 2 was immediately determined using flow through cells and pocket oxygen meters (FireStingGO 2 , Pyro-Science). In October, O 2 depth profiles were sampled on a parallel transect with 2-m distance from regular transect sampling. Oxygen probes were calibrated to 100% air-saturation using on-site seawater. The 0% air saturation calibration solution was prepared by adding the strong reductant sodium dithionite (about 30 g/L) to that seawater sample. Using salinity and temperature data, the O 2 air-saturations were transformed into molar concentrations (Garcia & Gordon, 1992). Temperature and salinity were measured instantly using hand-held conductivity sensors (WTW). During each sampling campaign, discrete surficial seawater samples from the north beach of Spiekeroog were retrieved as well. Additionally, in 2016 and 2017 surficial seawater nutrient samples were taken bimonthly at the permanent time series station close to the western tip of Spiekeroog Island (Figure 1b). Furthermore, freshwater lens monitoring wells were sampled for metal analyses, according to sampling techniques described in Seibert et al. (2018).
To measure sulfide, approximately 1.5 ml of bubble-free extracted pore water was added to 0.6 ml 50 mM Zn acetate solution in preweighted Eppendorf vials. Fe and Mn samples were filtered using 0.45-μm surfactant-free cellulose acetate syringe filters into low-density polyethylene bottles, which were acid-washed and rinsed with ultrapure water prior to use. After sampling, metal samples were acidified to 1% v/v using ultraclean concentrated HNO 3 . Samples were stored cool and dark until analysis.

Laboratory Analyses
Nutrient samples were not stored longer than 6 days. NO x − and NO 2 − were measured photometrically using Griess reagent and vanadium(III) chloride as reduction agent (Schnetger & Lehners, 2014). For determining NH 4 + , a photometric method was adapted (Benesch & Mangelsdorf, 1972). Instead of phenol, we used sodium salicylate to prepare the catalyst reagent. Additionally, time series station surface water samples were analyzed for dissolved Si, applying a spectrophotometric method (Grasshoff et al., 1999). Replicate measurements of in-house reference solutions resulted in precisions of better than 7% (NH 4 + , NO 2 − ), 4% (NO x -), and 9% (Si), respectively. The lowest calibration point was defined as limit of quantification. The limit of detection was therefore defined as one third of the limit of quantification (Table S9 in the supporting information). Dissolved sulfide was determined spectrophotometrically (Cline, 1969).
Trace metal samples from October were analyzed for Mn and Fe using sector field inductively coupled plasma mass spectrometry (SF-ICP-MS) with a resolution of R = 4,500 (Element 2, Thermo Fisher Scientific), whereas samples from March and August were analyzed using triple quadrupole ICP-MS (iCAP TQ ICP-MS, Thermo Fisher Scientific), applying the mode of kinetic energy discrimination to separate the analyte from polyatomic interferences. To measure within the calibration range, the samples were diluted with respect to their Mn and Fe concentration: All samples were diluted with 2% v/v HNO 3 of ultraclean grade at least tenfold and up to 100-fold for high Fe concentrations. Yttrium was used as an internal standard. To monitor trueness and precision, replicates of the two international standards Cass 5 and Nass 6 (National Research Council of Canada) were analyzed during each run. Aliquots of each reference material were spiked with Mn and Fe element standard solutions to concentrations of 0.46 μM Mn, respectively, 0.45 μM Fe as well as 0.91 μM Mn, respectively, 0.90 μM Fe, because pore water samples contained higher Mn and Fe concentration than both seawater reference standards. Replicate measurements of spiked reference standards yielded a trueness better than 2.9% and precision better than 2.6%. The limit of detection was calculated based on five to six blank replicates, which were measured during one run (Hubaux & Vos, 1970). It was multiplied with three to get the limit of quantification (Table S9).

Data Visualization
To suppress effects of vertical exaggeration during spatial interpolation, data points were interpolated on a rectangular grid with depth below surface on y axis. Afterward, these interpolated data were adjusted to  Table 1. *Q seawater balances the sum of Q runnel and Q LWL .
positions relative to standard elevation zero (NHN). The software package GMT 5 (Wessel et al., 2013) was used to compute and visualize the data.

Flux Calculations
Volumetric pore water fluxes to adjacent seawater were calculated for the study site by Beck et al. (2017), based on a numerical density-dependent groundwater flow and transport model. During its calibration, three model parameters (vertical and horizontal hydraulic conductivity and specific yield) were adjusted using groundwater levels measured in three groundwater wells located at supratidal and intertidal sites .
Infiltration zone (in) and exfiltration zone (out) boundaries within the model were operationally set by the half of surface elevation difference between (i) HWL and deepest runnel location, (ii) deepest runnel location and ridge top, and (iii) ridge top and LWL, respectively ( Figure 2). The model provides time-averaged saline and fresh fluxes, which were calculated for distinct hydrogeological areas, representing main flow paths within the intertidal beach ( Figure 2).
According to the model, seawater is delivered to the runnel exfiltration site from both HWL and ridge infiltration zones ( Figure 2). On average, 2.8 m 3 /day per meter shoreline of seawater is infiltrating at HWL and 0.5 m 3 /day per meter shoreline is transported from ridge to runnel. In total, the resulting average saline runnel exfiltration Q runnel is 3.3 m 3 /day per meter shoreline (Table 1). Further, from ridge to LWL an average saline flux Q LWL of 1 m 3 /day per meter shoreline was derived from the model. The total saline efflux must therefore be balanced by a seawater volume influx Q seawater of 4.3 m 3 /day per meter shoreline. Based on groundwater recharge estimations, the average freshwater flux Q FGW was assigned to be 0.75 m 3 /day per meter shoreline on average (Table 1).
The overall error on saline volumetric fluxes has been assumed to range from 0.5 to 2 times the average fluxes (Table 1; Beck et al., 2017). The halving of saline fluxes was assigned to the overestimation of ridge infiltration rates by the model . This overestimation not only led to uncertainty in fluxes but has also been identified to highly affect fresh groundwater flow paths: The model from Beck et al. (2017) revealed that fresh groundwater discharges in a restricted zone within the low lying runnel. In contrast, subsequent salinity mapping and additional hydraulic head measurements of the intertidal zone indicated that fresh groundwater discharges at LWL, too (Figures 1c and 2; Waska et al., 2019). The one fresh groundwater discharge tube concept was attributed to ridge head overestimations in the Beck et al. (2017) model. It produced a nonexistent groundwater divide preventing fresh groundwater from discharging at the LWL and further demonstrated the sensitivity of the flow path system to ridge morphology changes, which occur frequently in the study area (Waska et al., 2019). The fresh groundwater flux is assumed to be unaffected by its flow path through the intertidal beach.
The Beck et al. (2017) model did not account for wave forcing and storm floods. Therefore, the factor 2 has been assigned to saline volumetric fluxes to define the positive error range. Based on a numerical study on effects of tidal and wave forcing on a subterranean estuary, the doubling of saline seawater circulation rates could have been observed by adding wave forcing to a tidal model (Xin et al., 2010).
The intraannual variation of the fresh groundwater flux was estimated from seasonal hydraulic gradient changes along the flow paths from the groundwater divide at the center of the island toward the beach, and ranged from 0.35 to 1.2 m 3 /day per meter shoreline (Table 1; Waska et al., 2019).
To calculate constituent fluxes to the coastal waters, pore water and seawater end-member concentrations were multiplied with volumetric fluxes and summed up to get total efflux J out , influx J in , and net flux J net :

Journal of Geophysical Research: Biogeosciences
Saline volumetric fluxes were multiplied with pore water end-members from intertidal sites. The saline endmember concentrations C runnel and C LWL were chosen among 10-cmbsf samples from runnel and LWL of transect T1 and HRT (Figure 2; salinity >26; O 2 <10 μM). Recent seawater as well as time averages of past seawater compositions up to 8 months were defined as seawater end-members C seawater (Tables S6-S8). Eight months corresponds to the maximum seawater residence time in the upper circulation cell . Fresh groundwater volumetric fluxes, in contrast, were multiplied with an averaged inland fresh groundwater composition to take into account the uncertainties in fresh groundwater flow paths (Waska et al., 2019) and the small number of brackish beach pore water end-members. Fresh groundwater endmember concentrations C FGW were averaged from inland monitoring wells for nutrient (Seibert et al., 2018) as well as Mn and Fe data (Tables S6-S8).
Owing to the large range in possible residence times as well as in pore water end-member concentration, individual constituent fluxes for any combination of seawater age (from zero to eight months) and available pore water end-members were calculated. To consider as well the volumetric flux uncertainty, the average, minimum, and maximum fluxes (Table 1) were used for constituent flux calculations. The resulting individual fluxes were summarized in boxplots. In the following, "fluxes" refer to median fluxes.The applied model assumes that the end-members C runnel and C LWL are representative for exfiltration areas. However, the LWL sampling sites were operationally defined by the lowest reachable saline site (salinity >26) and any subtidal exfiltration areas, which may be connected to intertidal infiltration zones, receive no consideration. The model assumes that the volumetric fluxes remain constant throughout the year.

10.1029/2019JG005399
Journal of Geophysical Research: Biogeosciences from north beach was below detection limit, except for March (0.8 μM).
The highest NH 4 + concentration was found in north beach seawater in October (9 μM), whereas NH 4 + was either at detection limit in March, or below limit of quantification in August (<5.8 μM; Figure 7). Seawater Mn concentrations at the north beach were below 0.1 μM and thus very low in comparison with obtained pore water concentrations ( Figure 6). Similar to Mn, north beach seawater Fe concentrations (<0.03 μM) were orders of magnitude lower than pore water concentrations.

Pore Water 3.2.1. Salinity and Temperature
Although pore water salinities ranged from 12.5 (brackish) to 36.6 (saline), the median salinity of all samples was at 32.0 and 75% of all samples exhibited salinities above 30.6 ( Figure 4a). Confirming the results of Waska et al. (2019), we found brackish discharge patches throughout the intertidal zone especially at the LWL, but also at runnel sites (supporting information, Figures S1, S5, and S6). North beach seawater salinity ranged from 31.6 to 33.2.
Pore water temperatures changed seasonally, with almost no overlap in the temperature ranges between the sampling campaigns.
Corresponding seawater temperatures were below median pore water temperatures in October and March, and above in August (Figure 4b).

O 2 and Sulfide
The highest pore water O 2 concentrations corresponded to concurrent north beach seawater concentrations for October and March campaigns (285 μM in October and 307 μM in March; Figure 5a). In August, the maximum concentration of all samples (200 μM) was notably lower than the corresponding seawater concentration (245 μM) and 75% of the samples contained O 2 concentrations of less than 20 μM (Figure 5a).
In general, dissolved O 2 was found in highest concentrations close to the HWL and gradually decreased toward the runnel and with depth ( Figures 6, S5, and S6). O 2 was elevated on the top of the ridge, depending on season. Close to the LWL, O 2 was absent at all sampling depths. In October and March, the O 2 penetration depth ranged from deeper than 100 cmbsf at the HWL to less than 10 cmbsf at the LWL or at runnel sites (Figures 6, S5, and S6). A sharp horizontal gradient was observed at 35-m distance to the HWL in March, which was mapped precisely by the high-resolution sampling (Figure 6b). In contrast to October and March, O 2 was restricted to the upper beach close to the HWL in August and no O 2 was detected at the ridge. Furthermore, O 2 penetration depth was much lower.
Dissolved sulfide was either close to or below the limit of quantification (5 μM) in all samples (data not shown). However, we found indications of sulfate reduction. In some pore water samples from the LWL or from the runnel, sulfide smell was recognized.

Nitrate, Nitrite, and Ammonium
Within the pore waters, NO 3 − concentrations ranged between 0 and 73 μM over all seasons. The median of the March samples (22 μM) exceeded those of October and August (both below detection limit; Figure 5b). Spatial NO 3 − patterns were similar in all seasons with highest concentrations at elevated parts of the beach, that is, HWL and ridge. Concentrations at these sites exceeded concurrent seawater NO 3 − in many samples.
At the LWL and in runnel sediments, NO 3 − was absent (Figures 7, S7,   NO 2 − was only found at some spots with concentrations generally below 1 μM. Elevated NO 2 − was generally found at depths below high NO 3 − concentrations. Concurrently, in these samples NO 3 − was depleted compared to adjacent sampling depths ( Figures S2, S7, and S8).
Although the entire concentration range was different, October and March transect samples exhibited comparable median (both at 4 μM) and upper quartile concentrations (10 and 16 μM, respectively). In contrast, both median (25 μM) and upper quartile (35 μM) were higher in August (Figure 5b). In October and March, NH 4 + was restricted to the LWL and runnel. In contrast, NH 4 + was present nearly throughout the whole intertidal zone in August with only few exceptions (Figures 7, S7, and S8).

Dissolved Manganese and Iron
Dissolved Mn was detected in most samples at concentrations <20 μM; however, concentrations reached up to 70 μM (Figure 5a). During all campaigns, elevated concentrations were found in ridge pore waters and close to the LWL (Figures 6, S5, and S6). In March, dissolved Mn was additionally detected in the runnel (14 μM). In August, Mn concentrations were highest compared to the other seasons ( Figure 5a) and Mn was patchy distributed across the entire subsurface of the intertidal zone (Figures 6, S5, and S6). Dissolved Fe concentrations ranged up to 163 μM (Figure 5a), and highest concentrations were predominantly found close to the LWL (Figures 6, S5, and S6). The highest Fe concentrations were found in October (up to 163 μM). In general, lower concentrations were observed in March (<102 μM), but

10.1029/2019JG005399
Journal of Geophysical Research: Biogeosciences the spatial distribution was similar to October ( Figure 6). In August, dissolved Fe was spatially more heterogeneously distributed and also found at higher elevation levels across the beach face slope to the runnel as well as in the runnel (Figure 6).

. General Patterns in Flow Paths and Redox Zonation
Generally, flow paths within beaches, in particular, infiltration and exfiltration zones, determine seawater constituent supply (O 2 , NO 3 − , and marine organic matter) and the discharge of remineralization products (Mn, Fe, and NH 4 + ). At our study site, the pore water salinity corresponded with that of the seawater, indicating that seawater circulation was the predominant driver of intertidal pore water geochemistry. This is in line with previous studies at the site .
While O 2 penetrated deeply into the sediment at infiltration zones on the beach, it barely penetrated into the sediment in exfiltration zones, regardless of whether pore waters were saline or brackish. This led to a sharp vertical O 2 boundary located between HWL and runnel at the border of the infiltration and exfiltration zones (Figure 6b). O 2 dynamics were driven by microbial consumption and were closely correlated to N-species: NO 3 − concentrations were consistently elevated in oxic parts of the sediment and were often higher than those in the overlying water. This suggests that remineralized NH 4 + is nitrified to NO 3 − in these zones (Figures 7 and 9). Below oxygenated sediments, NO 3 − was depleted and was likely transformed to NH 4 + via dissimilatory nitrate reduction to ammonium (DNRA), to gaseous N 2 /N 2 O via denitrification, or consumed by microorganisms within the sediments. Consequently, there were zones in the sediment where pore waters had low DIN and low O 2 (Figures 7, 9, and S3). A similar coupling of nitrification and denitrification has been found in intertidal sands from Tolo Harbor, Hongkong (Liu et al., 2017) and in nearby subtidal sediments of the North Sea where N-loss is high. In these permeable sediments, denitrification appears to be favored over anammox due to high organic matter input and rapid fluctuations in O 2 concentrations (Marchant et al., 2016;Marchant et al., 2017).
Additional NH 4 + was delivered to brackish mixing zones by fresh groundwater from the inland's freshwater lens, which carries up to 78 μM NH 4 + (Seibert et al., 2018), but highest NH 4 + concentrations were found in saline anoxic pore waters and most likely produced in situ during anaerobic respiration ( Figure S4). Due to the cooccurrence with reduced substances like dissolved Fe and Mn (Figures 6 and 7), it is likely that NH 4 + Figure 9.

10.1029/2019JG005399
Journal of Geophysical Research: Biogeosciences was released during organic matter degradation via Mn or Fe oxide reduction and is not subject to rapid (re-) oxidation in these environments.
Dissolved Mn and Fe are produced, when particulate Mn and Fe minerals are reduced during organic matter degradation, which occurs after the complete consumption of more favorable electron acceptors like O 2 or NO 3 − . This general pattern was found in our study as well, indicated by the occurrence of Mn and Fe concentrations in zones of O 2 depletion. Mn and Fe concentrations are expected to be highest in exfiltration zones. However, in shallow pore waters of exfiltrating sites (10 cmbsf), we found depletions of Mn and Fe compared to deeper layers, probably due to potential oxidation of both compounds. In intertidal sediments the oxidation of dissolved Mn and Fe by O 2 and NO 3 − is a likely process due to advective pore water transport (Caetano et al., 1997;Luther et al., 1997). Concurrently, dissolved Fe may form solid Fe sulfides, when it reacts with dissolved sulfide (McAllister et al., 2015). Although, Fe-S-interactions are present at the studied sediment depths (<100 cmbsf), they are of less importance as these sediments at the study site are generally characterized by low rates of bacterial sulfate reduction and sulfide reoxidation .

Seasonal Trends in Pore Water Geochemistry
Although the same general geochemical zonation was observed during all three sampling campaigns, substantial seasonal changes occurred. The O 2 penetration depth as well as the spatial distribution of NO 3 − and reduced metabolites, like NH 4 + , Mn, and Fe, varied seasonally (Figures 6 and 7). A changing O 2 penetration depth is an essential factor for biogeochemical processes, because it separates aerobic from anaerobic respiration pathways. O 2 and NO 3 − are the most favorable electron acceptors (Froelich et al., 1979), and their availability mostly suppresses the utilization of other electron acceptors with lower energy yield.
In October and March, deep O 2 penetration and excess NO 3 − indicated aerobic respiration to be the most prominent organic matter degradation pathway in shallow sediments. Concurrently, the runnel site seemed to be a hotspot of DIN-loss where pore waters were found to be low in O 2 , DIN, Mn, and Fe ( Figure 6, 7, and S3). Indications for Mn and Fe reduction co-occurred with highest loads of NH 4 + but were restricted to the LWL.
In contrast to October and March, dissolved O 2 was restricted to the uppermost sampling depths in August, which indicated an increased O 2 consumption rate at that time (see also Waska et al., 2019). There was no pronounced DIN minimum in August (Figures 7 and S3), which does not necessarily mean that denitrification was absent, but intensified remineralization releasing NH 4 + caused excess DIN (Figure 9).
Simultaneously, indications for Mn and Fe reduction were found throughout the intertidal subsurface, including upper beach sites and runnel locations.
The sampling in August was conducted during a different tidal state (shortly after high tide) compared to those in March and October (after low tide). Therefore, we cannot completely rule out an imprint of tidal dynamics on our results, for example, on dissolved O 2 . In August, oxygenated seawater was introduced into the sediment shortly before sampling. This could have increased dissolved O 2 concentrations at HWL locations (Charbonnier et al., 2016) compared to samplings at the same sites during rising tides. Thus, seasonal dynamics would have been rather underestimated than overestimated. Further, tidal dynamics have been shown to even affect pore waters down to 3 mbsf. The observed tidal changes in DIN species were in the range of lower μM, and the overall species distributions did not change significantly (Liu et al., 2017). We cannot exclude that tidal variabilities might have overprinted our results. Nevertheless, we think that the magnitude of the seasonal variability is much higher than that of tidal dynamics.

Pelagic Organic Matter Availability Affects Benthic Rates
O 2 penetration depth is a result of O 2 uptake rates, and transport rates. Assuming no drastic changes in the latter, the mainly microbially mediated O 2 uptake rates are assumed to vary highly with season. Increases in O 2 uptake rates would reduce the O 2 penetration depth, thereby promoting anaerobic degradation pathways in shallow sediments.
Benthic organic matter degradation rates respond to changes in organic matter composition and quantity. Fresh marine-derived organic matter is predominantly available from late spring to early autumn due to 10.1029/2019JG005399

Journal of Geophysical Research: Biogeosciences
intense pelagic activity in our study area (Meier et al., 2015;Seidel et al., 2015). Labile organic matter stimulates benthic carbon degradation and enhances microbial rates (Hedges et al., 1988).
The seasonal changes in primary production in the Wadden Sea area are reflected in Si and NO 3 − concentrations determined in the open water column at a nearby permanently installed time series station (Beck & Brumsack, 2012;Grunwald et al., 2010). Silica decreases in spring are linked to diatom frustule production. Spring Si and NO 3 − decreases are caused predominantly by uptake, but concentrations are also affected by sedimentary recycling processes in sediments and water column (Grunwald et al., 2010). Further NO 3 − inputs from rivers vary seasonally (Johannsen et al., 2008).
Based on nutrient concentrations measured in the water column in 2016 and 2017 (Figure 3), it can be deduced that the March sampling campaign took place during starting bloom conditions, because Si concentrations were already declining. Therefore, we expect a rather low supply of labile marine organic matter in March. Consequently, aerobic respiration rates were assumed to be comparably low, allowing O 2 to penetrate deep (>1 mbsf) into the sediment (Figure 6).
In contrast, the August campaign took place during postbloom conditions, when most nutrients of the coastal water inventory were depleted ( Figure 3). The preceding bloom would have resulted in an enhanced supply of labile organic matter to beach sediments in the months and weeks before. This leads to the complete O 2 consumption in shallow sediment depths and consequently to the shift to other electron acceptors like solid Mn and Fe oxides ( Figure 6). During the October campaign, water column nutrient remineralization exceeded nutrient consumption. Based on pore water O 2 and NO 3 − distributions (Figures 6 and 7), we expect October to be an intermediate state between March and August with respect to labile organic matter availability and O 2 consumption rates.
The supply of labile organic matter is not only related to current pelagic activity. Due to its reduced mobility, particulate organic matter is further trapped in intertidal sediments. This "carbon memory" has been observed to enhance microbial rates for up to 2 months (Kim et al., 2019). It is thus likely that both August and October rates have been affected by the high pelagic activity in prior months.

Seawater Temperature and Electron Acceptor Supply Governs OM Degradation
Due to continuous seawater infiltration, temperatures of saline pore waters were strongly coupled to pelagic water temperatures (Figure 4). O 2 consumption is strongly related to temperature and increases 1.8-fold for a 10°C rise in temperature (Thamdrup et al., 1998). Theoretically, a decrease in O 2 penetration by a factor of about 2 in August compared to March can be solely explained by the pore water temperature increase, assuming all other factors, for example, transport rates, to be constant. The current dataset does not allow to decide whether organic matter or temperature is the main driver for the shift in oxygen penetration depth.
Besides for aerobic degradation rates, a temperature dependence has been observed for nitrate involving degradation pathways (Kaplan et al., 1977;Macfarlane & Herbert, 1984) and Mn and Fe reduction (Lovley et al., 2004;Nevin et al., 2005). However, these pathways are further controlled by oxygen presence or their respective electron acceptor availability (NO 3 − and Mn/Fe oxides).
Besides seasonal variability in electron donors and temperature, the supply of electron acceptors, like O 2 and NO 3 − , to beach sediments also varies seasonally (Figure 3). The infiltration of cold O 2and nutrient-rich seawater from autumn to early spring promotes aerobic respiration and NO 3 − reduction (Froelich et al., 1979).
In March, seawater O 2 concentration was about 20% higher compared to August (Figure 5a) due to higher gas solubility in colder seawater. Additionally, seawater NO 3 − concentration was higher in March compared to October and August (Figure 5b). The availability of NO 3 − is a controlling parameter stimulating nitrate reducing organic matter pathways, like denitrification (Deek et al., 2011). Hence, seawater in March supplied the highest quantity of the electron acceptors O 2 and NO 3 − , supporting aerobic respiration and nitrate reduction.

Effects of Changes in Beach Morphology
The redox zonation is also influenced by pore water flow paths and residence times, which are controlled by beach morphology and aquifer characteristics (McLachlan & Turner, 1994). A recent study at Spiekeroog beach emphasizes the coupling of beach morphology and subsurface pore water geochemistry (Waska et al., 2019). Thereafter, the intertidal morphology governs, where seawater constituents like O 2 , organic matter, and seasonally NO 3 − infiltrate as well as where mineralization products discharge. Further, beach 10.1029/2019JG005399

Journal of Geophysical Research: Biogeosciences
morphology affects the mixing of Fe-rich pore water with sulfidic fresh groundwater, which could lead to Fe precipitation (Waska et al., 2019). It is well known that beach morphology of north beach of Spiekeroog Island exhibits a seasonal dynamic (Waska et al., 2019), namely, a steep summer and a flat winter beach (Flemming & Davis, 1994). A seasonal imprint on pore water flow paths, thus, has to be considered as factor changing subsurface biogeochemistry.

Variability of Constituent Fluxes 4.2.1. Inorganic Nitrogen Fluxes
Flux calculations revealed that inorganic nitrogen species displayed the highest seasonal variability. The intertidal beach sediments behaved as net sink for NO 3 − , which was either transformed to gaseous species like N 2 O or N 2 , to dissolved NH 4 + via DNRA and ammonification, or fixed in biomass. Predominantly from autumn to early spring, NO 3 − was transported into the intertidal subsurface with infiltrating seawater (Figures 3 and 8a). Concurrently, the efflux into surface waters was close to zero throughout the year (Figure 8a).
In contrast, NH 4 + was generally exported from beach pore water to nearshore waters, due to intense microbial organic matter degradation. In October and March lower degradation rates compared to August caused lower net NH 4 + fluxes. Additionally, higher seawater NH 4 + input in March was decreasing the net efflux of NH 4 + (Figure 8c). In August, higher microbial rates promoted anoxic conditions in the subsurface, which led to accumulating NH 4 + in pore waters along the flow path. The August NH 4 + net flux (121 mmol NH 4 + /day per meter shoreline) is comparable to reported May net fluxes at our study site (117 mmol NH 4 + /day per meter shoreline; Beck et al., 2017). Although the May fluxes might have been biased because samples of 50 cmbsf were used as end-members , it confirms that the onset of pelagic activity and rising temperatures in spring stimulated microbial activity in the intertidal subsurface. The trend to higher NH 4 + fluxes in summer has also been observed in a recent study from Tolo Harbor, Hongkong,   (Liu et al., 2018). The authors concluded that the increase in summer net fluxes was stimulated by enhanced NH 4 + production during organic matter degradation.
In a regional comparison with local tidal flats the release of NH 4 + from intertidal beach sands in October (46 mmol NH 4 + /day per meter shoreline) is up to 20 times lower than the release from local backbarrier tidal flats in November (532-829 mmol NH 4 + /day per meter shoreline; Riedel et al., 2011). Combined with the lower spatial extension of beaches (Figure 1), it is obvious that beach sediments may contribute but are not the dominant NH 4 + source to the coastal North Sea.
The high seasonal variability of the net DIN flux results from seasonal fluctuations in seawater NO 3 − inputs and variable effluxes of remineralized NH 4 + . Whereas the first is governed by pelagic algae biomass and other NO 3 − sinks, the latter mainly depends on microbial respiration rates. The temporal variability in DIN fluxes has been shown in prior studies (Couturier et al., 2017;Gonneea & Charette, 2014;Liu et al., 2018) where fluxes were in the same range as at our study site, ranging from 52 mmol DIN/day per meter shoreline (Gonneea & Charette, 2014) to 388 mmol DIN/day per meter shoreline (Couturier et al., 2017). However, a temporary net removal of DIN has not been observed in these environments.  Gao et al., 2012). Due to the smaller spatial extension of sandy beaches compared to intertidal flat margins and subtidal sediments in the southern North Sea, the contribution of sandy beaches to N loss might be only minor in the southern North Sea. Additionally, net N loss is only suggested to be present in cold seasons, when seawater NO 3 − is high and microbial rates are comparably low.
Nevertheless, beach sands contribute to N budgets in a regional context.

Mn and Fe Fluxes
The import of dissolved Mn must be considered for spring and autumn, when dissolved Mn concentrations in adjacent seawater have been reported to exceed 1 μM (Kowalski et al., 2012). However, the high concentrations of Mn have been identified to result from pore water derived Mn release in backbarrier tidal flats, after high organic matter input from algal blooms. It is thus more likely that intertidal beach sediments additionally govern coastal North Sea seawater Mn inventories. The flux of Mn was enhanced during summer when higher microbial rates and O 2 consumption led to more reducing conditions in pore waters (Figure 8b). In August, high Mn concentrations could even be observed in the upper circulation cell ( Figure 6). As a consequence, Mn was exported into coastal surface waters. Oxidation within oxygenated surface sediment is unlikely due to high groundwater velocities of 0.3 to 0.7 m/day  and slow oxidation rates of Mn (Caetano et al., 1997).
Similar to Mn, the maximum Fe flux in August (Figure 8d) can be attributed to more reducing conditions in the beach sediments, especially in the upper circulation cell. The wide range of calculated fluxes is due to the high patchiness of dissolved Fe at 10 cmbsf ( Figure 6). In pore waters dissolved Fe is mainly present as Fe 2+ , but it oxidizes rapidly to particulate Fe when aerated seawater enters the sediment (Caetano et al., 1997). The lower Fe fluxes in comparison with 50-cmbsf-based Fe fluxes from Beck et al. (2017; Figure 8d) is most likely the result of an oxidation effect on 10 cmbsf Fe end-members of the current study. Nevertheless, due to the rapid oxidation of Fe, further oxidation in very surficial sediment depths must be considered and it is likely that potential Fe fluxes to coastal waters are still overestimated. Due to low sulfate reduction rates , Fe is not sequestered in large quantities by dissolved sulfide at this site in contrast to other marine sites with intense sulfate reduction (McAllister et al., 2015;Roy et al., 2010). Therefore, the intertidal beach pore waters exhibit ideal conditions for Fe reduction and for the formation of Fe-rich pore waters, which may be transported across the sediment-water interface, either by advective transport or by sediment erosion. Consequently, beach sediments of Spiekeroog Island must be considered as Fe and Mn source for the adjacent coastal ocean.

Constituent Mass Balances in Dynamic Intertidal Beaches
Constituent net fluxes have been shown to vary significantly on seasonal timescales (Figure 8) due to seasonal changes in subsurface biogeochemistry. The latter as well as pore water flow paths are influenced by steady sediment reallocation, storm surges, spring-neap tide cycles, and potential heterogeneities in sediment permeability. Furthermore, model-related uncertainties in volumetric fluxes have to be considered. Although subsurface flow heterogeneities were assumed to produce long-shore variabilities in salinity and temperature, the resulting biogeochemical patterns were predominantly governed by cross-shore systematics (Waska et al., 2019).
Model-related uncertainties in volumetric flux magnitudes have been integrated to flux calculations, which also accounted for different volumetric ratios of saline runnel and LWL exfiltration. Generally, such uncertainties will not change the ubiquitous mechanisms of the beach bioreactor based on a mass balance model of influx of seawater and the efflux of modified pore waters, which have been overprinted by organic matter degradation. At our study site, fluxes in the upper circulation cell are the main driver of total seasonal flux variations. It exhibits the highest volumetric exchange rates  and greatest variability in pore water biogeochemistry (Figures 6 and 7).
Uncertainties related to changes in flow paths have not been considered explicitly. Drastic changes, for example, the sole exfiltration of all infiltrating seawater at the LWL, principally could lead to wrong combinations of end-member and volumetric flux, but we can exclude such a flow path pattern based on intertidal head measurements Waska et al., 2019). Additionally, pore water data of all transect replicates showed a comparable succession of electron acceptors, such as deep O 2 penetration at HWL, intermediate O 2 penetration at ridge sites, and very low oxygen penetration at runnel and LWL sites (Figures 6, 7, S5, S6, S7, and S8), which supports the underlain hydrogeological concept.
Flow path variabilities might increase the complexity of the organic matter degradation in intertidal sediments. Although potentially prone to errors, the presented constituent flux variability seems to be a robust estimation of how seasonal biogeochemical changes affect the fate of inorganic N, Fe, and Mn in intertidal high energy beach sediments.

Conclusions
The presented pore water data reveals that the interplay of benthic remineralization processes in intertidal pore waters and biological activity in adjacent coastal seawater affects organic matter degradation in intertidal permeable sediments. This, in turn, affects the composition of discharging fluids and thus constituent fluxes across the sediment-water interface. The main conclusions are visualized in Figure 10.
In the upper circulation cell, distinct vertical redox boundaries were found separating infiltration and exfiltration zones of seawater. Seasonal changes in temperature, organic matter, and electron acceptor availability (O 2 , NO 3 − ) affect microbial rates and respective electron acceptor utilization ( Figure 10). The spatial distribution of organic matter degradation pathways, predominantly in the upper seawater circulation cell, affects pore water species and finally their flux across the sediment-water interface.
Net fluxes calculation of DIN species, Mn, and Fe showed seasonal variations by orders of magnitude. But fluxes may not only vary in magnitude. Beaches may even switch from sink to source, as shown for DIN. From autumn to early spring, seawater supplies high amounts of NO 3 − , low organic matter quantity, and quality, and thus microbial rates are considered to be low ( Figure 10). During this time, aerobic and nitrate reducing organic matter degradation are most dominant and intertidal sediments can be an overall sink for seawater DIN. In contrast, seawater supplies very low amounts of NO 3 − and concurrently a high quantity and quality of fresh organic matter in summer. In combination with increased temperatures microbial rates are higher in summer. This creates spatially more extended anoxic conditions, promotes Mn and Fe reduction and enhances the release of NH 4 + ( Figure 10). As a consequence, pore waters provide recycled DIN, as NH 4 + , to coastal seawater in biologically active seasons and may thus refuel surface bioproductivity with 10.1029/2019JG005399

Journal of Geophysical Research: Biogeosciences
potentially limiting DIN species. In this respect beaches may be regarded as a buffer for DIN supply of the nearshore.
In the regional context, the imprint of DIN fluxes from barrier island beaches on nutrient budgets might be limited due to the large spatial extension of subtidal sediments and tidal flats, which release NH 4 + as well. In contrast to DIN, barrier islands must be considered as Fe source to coastal seawater, although there are still uncertainties in the sediment-water interface Fe sequestration.
The results have implications for temporal constituent flux extrapolations of subterranean estuaries, which receive organic matter and DIN from overlying waters and which are affected by seasonal changes in pelagic activity. The overall impact of such subterranean estuaries on adjacent surface waters must be carefully assessed, considering flux variations or even the inversion of source or sink characteristics.
To date, high energy beach biogeochemistry has received only minor attention, regarding nutrient and metal fluxes although these beach types are widespread along global coastlines. For example, the effect of steady beach sediment redistribution is not yet constrained with respect to dissolved constituent and suspended matter export and should be subject to future studies.