Compound-specific stable isotope analysis of amino acid nitrogen reveals detrital support of microphytobenthos in the Dutch Wadden Sea benthic food web

The Wadden Sea is the world’s largest intertidal ecosystem and provides vital food resources for a large number of migratory bird and fish species during seasonal stopovers. Previous work using bulk stable isotope analysis of carbon found that microphytobenthos (MPB) was the dominant resource fueling the food web with particulate organic matter making up the remainder. However, this work was unable to account for the trophic structure of the food web or the considerable increase in δ15N values of bulk tissue throughout the benthic food web occurring in the Eastern regions of the Dutch Wadden Sea. Here, we combine compound-specific and bulk analytical stable isotope techniques to further resolve the trophic structure and resource use throughout the benthic food web in the Wadden Sea. Analysis of δ15N for trophic and source amino acids allowed for better identification of trophic relationships due to the integration of underlying variation in the nitrogen resources supporting the food web. Baseline-integrated trophic position estimates using glutamic acid (Glu) and phenylalanine (Phe) allow for disentanglement of baseline variations in underlying δ15N sources supporting the ecosystem and trophic shifts resulting from changes in ecological relationships. Through this application, we further confirmed the dominant ecosystem support by MPB-derived resources, although to a lesser extent than previously estimated. In addition to phytoplankton-derived particulate, organic matter and MPB supported from nutrients from the overlying water column there appears to be an additional resource supporting the benthic community. From the stable isotope mixing models, a subset of species appears to focus on MPB supported off recycled (porewater) N and/or detrital organic matter mainly driven by increased phenylalanine δ15N values. This additional resource within MPB may play a role in subsidizing the exceptional benthic productivity observed within the Wadden Sea ecosystem and reflect division in MPB support along green (herbivory) and brown (recycled/detrital) food web pathways.


Introduction 28
The Wadden Sea is the world's largest intertidal ecosystem (Wolff 1983) stretching 29 from The Netherlands to Denmark formed behind a barrier island chain with connection to 30 the North Sea (Postma 1996). This intertidal ecosystem has been designated as a culturally 31 significant UNESCO World Heritage site due to considerable biodiversity within the system 32 as benthic productivity supports an estimated 10 -12 million migratory birds across each 33 year (Reise, et al. 2010). The Wadden Sea has a long history of multiple direct impacts from 34 human activity (Wolff 2013, Eriksson, et al. 2010) that includes land reclamation, partial 35 damming and hydraulic changes, eutrophication, overfishing, and extensive dredging for 36 shellfish. These impacts have resulted in a shift from a benthos dominated by seagrass, 37 extensive bivalve reefs, and supporting apex predators towards one dominated by 38 polychaetes, with minimal fringing bivalve reefs, and largely devoid of apex predators 39 (Philippart, et al. 2007). Despite historical and recent anthropogenic impacts, the Wadden Sea 40 remains a very productive intertidal ecosystem. However, changes in higher trophic level 41 species have coincided with shifts in the benthic community as inferred from bird species 42 shifting from primarily bivalve carnivores towards polychaete carnivores (Eriksson,et al. 43 2010, Van Roomen, et al. 2005) and the drivers of these shifts remain unclear. 44 Analysis of the stable isotope composition of carbon (δ 13 C) and nitrogen (δ 15 N) in 45 animal tissues are routinely used to identify underlying resource use and trophic relationships 46 within ecosystems (Fry 1988, Minagawa andWada 1984). Through the application of species 47 specific trophic discrimination factors (TDFs) for carbon and nitrogen (Δ 13 C and Δ 15 N) 48 stepwise isotopic increases that occur between a consumer's diet and their tissues during 49 metabolism can be accounted for (McCutchan, et al. 2003, Post 2002, allowing for 50 identification of the animal's trophic level. Accounting for isotopic changes across trophic 51 levels allows for the application of both simple and complex stable isotope mixing models 52 (SIMMs) to be used to resolve the relative use of food resources by consumers within a food 53 web (Stock, et al. 2018, Fernandes, et al. 2014). For the Dutch part of the Wadden Sea, 54 Christianen, et al. (2017) identified that the majority of the biomass is supported by both 55 pelagic phytoplankton and microphytobenthos. Their study used a two-source mixing model 56 with a single isotopic tracer (δ 13 C) and clearly identified benthic productivity from 57 microphytobenthos (MPB) as an important resource supporting the benthic community. 58 However, they may have overestimated the amount of support due to compression of the 59 MPB endmember through the use of an animal proxy instead of directly measured resources 60 (Post 2002). 61 Thus far, it has not been possible to assess the trophic structure within the Dutch 62 Wadden Sea benthic community due to the large variability in δ 15 N values observed in 63 multiple benthic species (Christianen, et al. 2015). This work found clear trends of increased along with longer exposure times in the east than in West (Compton, et al. 2013, Otto, et al. 79 1990). This combination of unique physical and biogeochemical settings may contribute to 80 increased underlying δ 15 N values but identifying any use of denitrification-affected N 81 porewater resources by MPB and the food web has remained largely intractable solely using 82 traditional bulk isotope techniques.

83
MPB thrive in unvegetated sandy sediments where they can contribute significantly to 84 the primary production supporting an ecosystem (Miller, et al. 1996) as benthic diatoms fix 85 carbon and take up nitrogen from the overlying water column and porewaters (Cook,et al. 86 2007, , Oakes, et al. 2012). Much of the fixed carbon is excreted as 87 extracellular polymeric substances (EPS) which are sticky, sugar rich substrates that help to 88 stabilize sediment and facilitate diatom motility throughout sand (Goto, et al. 1999, Stal 89 2010), but also serve as a labile carbon source for heterotrophic bacteria (Taylor, et al. 2013).

90
Diatom motility in sandy sediments is an adaptation to the physiochemical variations that 91 occur within intertidal settings. Diel vertical migration of diatoms coincides with tidal cycles and available sunlight to support favorable conditions for photosynthesis at the surface 93 (Barnett, et al. 2020) and to maximize nutrient availability for cell growth and division in the   In this study, we apply natural abundance stable isotope analysis of amino acid      (Table 1).  associated with the mouth of the river Ems). Despite application of a regional adjustment, 274 ranges for bulk δ 13 C and δ 15 N values remained large for consumers ( Fig. 1a and 1b)

294
Calculated trophic levels for primary producers ranged from 0.3 for Fucus vesiculosus 295 to 1.7 for C. rubrum (Fig. 4). Primary consumers ranged from 1.2 for Ensis leii to 2.3 for 296 Littorina littorea, secondary consumers from 2.6 for Hediste diversicolor to 3.5 for Crangon 297 crangon, and tertiary consumers from 3.6 for Gobius sp. to 4 for D. labrax (Fig. 3). No amino 298 acid measurements were possible for POM or MPB as part of this study as the sample 299 weights remaining after the initial bulk analysis between 2012 and 2015 were below 300 detection limits for this analysis. Consumer δ 13 C values fell between the measured values for 301 POM and MPB (-21.2‰ and -13.4‰). We therefore used the trophic level estimates 302 calculated using system-specific TDF and β to account for fractionation while estimating 303 values for δ 13 CBase and δ 15 NPhe-base. 304 We then used δ 13 CBase values from POM and MPB (Table 2) and each consumer to 305 calculate dietary contributions for species averages (Fig. 5a) and for all individuals within  (Table 2b), to calculate dietary contributions from species average values (Fig. 5b). Both benthic species prior to analysis with SIMMs. Through the application of source amino acids, we were able to construct a trophic structure within the Dutch Wadden Sea despite the 345 considerable underlying variability present in bulk δ 15 N values across the area (Fig. 1a). This 346 variability was likely caused by widespread biogeochemical processing unique to regions of 347 intense tidal flushing within intertidal sands (e.g. denitrification). Analysis of trophic and 348 source amino acid δ 15 N values allowed for the development and application of a system 349 specific TDF and β to determine the trophic structure of the Wadden Sea benthic food web 350 that was previously impossible using solely bulk isotopic techniques. 351 We confirm that MPB is a major source of productivity that widely supports the food  Table 1). Despite this adjustment within the wider data set, there 396 remains considerable variability in δ 15 N values (Fig. 1a) across the basin that has made 397 identifying a trophic structure using bulk δ 15 N values alone difficult without a method to 398 integrate underlying shifts in baseline δ 15 N values across the Dutch Wadden Sea.

399
The TDF value from analysis of the difference between Glu and Phe throughout the 400 food web is lower in this study than the most often applied canonical value (4.9‰ compared  The lower TDF value reflects an ecosystem wide application to species ranging from 404 macroalgae to teleosts and represents a compromise between application of multiple species-405 specific TDFs as determined by feeding studies (28×, 1 per species) or feeding group  readily available due to tidal pumping (Fig. 6). Food webs are often split into herbivory 457 (green) and detrital (brown) pathways depending on whether newly fixed productivity or 458 detrital reworking are the basal resource supporting them (Middelburg 2014, Odum 1969.

459
The separation observed between these two groups of species appears to be along the