Deep Into Oceanic N2 Fixation

Deep into oceanic N2 fixation Benavides Gorostegui, Mar; Bonnet, Sophie; Berman-Frank, Ilana; Riemann, Lasse Published in: Frontiers in Marine Science DOI: 10.3389/fmars.2018.00108 Publication date: 2018 Document Version Publisher's PDF, also known as Version of record Citation for published version (APA): Benavides Gorostegui, M., Bonnet, S., Berman-Frank, I., & Riemann, L. (2018). Deep into oceanic N2 fixation. Frontiers in Marine Science, 5, [108]. https://doi.org/10.3389/fmars.2018.00108

The biological fixation of dinitrogen (N 2 ) by marine prokaryotes called diazotrophs is the major source of nitrogen to the ocean, estimated at ∼106-120 Tg N y −1 (Gruber, 2004;Gruber and Galloway, 2008). This process contributes importantly to sustain primary production and maintain the global nitrogen inventory. The nitrogen reservoir is further controlled by fixed nitrogen loss processes including sediment burial, denitrification, and anammox (Falkowski, 1997), which exceed fixed nitrogen gains through N 2 fixation, leading to an imbalanced global nitrogen budget (Codispoti et al., 2001;Codispoti, 2007;Eugster and Gruber, 2012). Since the early 1970s, diazotrophic activity has been attributed to autotrophic cyanobacteria constrained to the sunlit and oligotrophic layer of the tropical and subtropical oceans (Zehr, 2011). Yet substantial evidence indicates a high diversity and wide distribution of non-cyanobacterial diazotrophs (bacteria and archaea) in the oceans (Zehr et al., 1998(Zehr et al., , 2000Farnelid et al., 2011;Bombar et al., 2016;Moisander et al., 2017). These diazotrophs are potentially not constrained by light as are their cyanobacterial counterparts, and have been detected in wide-ranging environments such as nutrient-rich, cold, and/or dark ecosystems including coastal upwelling regions (Sohm et al., 2011), temperate coastal zones (Bentzon-Tilia et al., 2015), and the deep ocean (Hewson et al., 2007;Hamersley et al., 2011).
Stretching the environmental boundaries, beyond those traditionally thought to constrain N2 fixation, will likely impact current estimates of nitrogen input to the global ocean. Extending the latitudinal limits from the tropics and subtropics to temperate waters would already represent a considerable increase in the potentially active N 2 fixation area, but spreading this area vertically to the mesopelagic (200-1,000 m) and bathypelagic (1,000-4,000 m) ocean would be immense. Aphotic N 2 fixation rates are usually low when compared to surface activity (<1 nmol N L −1 d −1 ; see Table 1 in Moisander et al., 2017) but the volume of the deep ocean is enormous. Consequently, studies comprising both photic and aphotic N 2 fixation measurements report depth-integrated aphotic rates representing 40-95% of the whole water column activity (Bonnet et al., 2013;Rahav et al., 2013;Benavides et al., 2015). Hence, aphotic fixation can account for a significant or even predominant fraction of water column N 2 fixation.
With the mere purpose of illustrating the potential budgetary relevance of the aphotic N 2 fixation to the global fixed nitrogen input, a back-of-the-envelope calculation can be carried out. If we consider a scenario for the mesopelagic zone (where the great majority of published aphotic N 2 fixation measurements were obtained from): taking the lower-end range of aphotic N 2 fixation rates available in the literature (0.01-0.1 nmol N L −1 d −1 ; Table 1 in Moisander et al., 2017), and the estimated volume of the mesopelagic zone (2.63 × 10 17 m 3 ; Arístegui et al., 2005), mesopelagic N 2 fixation would range between 13 and 134 Tg N y −1 . Fixed nitrogen inputs to the ocean include fluvial inputs, atmospheric deposition and biological N 2 fixation, which add up to 187-279 Tg N y −1 ( Table 1). Combining denitrification (including sediment burial) and anammox, fixed nitrogen losses add up to 260-475 Tg N y −1 ( Table 1). Adding mesopelagic N 2 fixation to fixed nitrogen inputs and subtracting losses from gains, we obtain differences ranging from a loss of 183 to a surplus of 114 Tg N y −1 ( Table 1). Despite this extrapolation may be questionable given that data on aphotic N 2 fixation are so sparse that the spatial distribution of mesopelagic N 2 fixation is unknown, it does illustrate that aphotic N 2 fixation could be important to global nitrogen budget considerations, and thus deep N 2 fixation should be further explored. Considering the stock of fixed nitrogen in the mesopelagic zone (Gruber, 2008) and the range of mesopelagic N 2 fixation rates estimated here (i.e., 13-134 Tg N y −1 ), N 2 fixed and eventually remineralized to nitrate in the mesopelagic zone would turn over in 4 to 43 y. The currently available dataset (Table 1 from Moisander et al., 2017, this issue) lacks robustness because (i) the number of measurements is limited and geographically sparse, and (ii) methodological difficulties are entailed in the detection of low N 2 fixation rates. While aphotic N 2 fixation has been consistently reported in several tropical and temperate waters ( Table 1; Moisander et al., 2017, this issue), it is unknown whether it occurs homogeneously throughout the dark water column or only in micro-niches where suitable conditions are found. Such hospitable niches may comprise aggregates, or organic matter accumulation zones like ecotones, fronts or water mass boundaries (Benavides et al., 2015;Bombar et al., 2016). Only a few studies have documented nifH gene expression in aphotic waters (e.g., Jayakumar et al., 2012), and it is debated whether reported abundances of non-cyanobacterial diazotrophs can account for measured rates of N 2 fixation when considering published cell specific rates of cultivated strains (Turk-Kubo et al., 2014;Bentzon-Tilia et al., 2015). This introduces uncertainty to the reliability of measuring especially low N 2 fixation rates (Gradoville et al., 2017), and emphasizes the need for continued refinement of the 15 N 2 incorporation method .
In this context, it is pertinent to consider the methodological difficulties encompassed in the detection of low N 2 fixation rates using 15 N 2 as a tracer. The precision of N 2 fixation rates may be affected by (i) a slower than theoretically assumed dissolution of the 15 N 2 bubble in seawater (Mohr et al., 2010;Großkopf et al., 2012), (ii) the contamination of 15 N 2 gas stocks with nitrogenous species other than N 2 (Dabundo et al., 2014), and (iii) failure to measure time zero δ 15 N values of the particulate nitrogen pool. As any other tracer method, 15 N 2based N 2 fixation rates are subject to a number of other sources of error, including variability in incubation and/or filtration time among replicates, sample particle size and its retention in filters varying with filter pore size (Bombar et al., 2018), as well as heterogeneous distribution of particles in Niskin bottles (Suter et al., 2017). Moreover, the vast majority of 15 N 2 -based published N 2 fixation measurements report net rates, whereas the leakage of 15 N-labeled dissolved organic nitrogen and/or ammonium can be significant in certain cases (e.g., Berthelot et al., 2017).
Most of the compiled aphotic rates (Moisander et al., 2017, this issue) were measured using the bubble method (Montoya et al., 1996), and should be considered as minimum estimates, despite the fact that they were performed in cold waters (typically ∼10 • C), which enhances gas dissolution and hence optimizes isotopic equilibrium in seawater samples enriched with 15 N 2 gas. Moreover, the majority of the studies i) used an isotope brand that provides high purity 15 N 2 gas, affecting aphotic N 2 fixation rates by <1% when 15 N-labeled nitrogen molecules other than N 2 are taken up (Benavides et al., 2015) and/or ii) provided time zero δ 15 N values of the particulate nitrogen pool at each sampling depth, making their results robust (Bonnet et al., 2013;Rahav et al., 2013;Benavides et al., 2015Benavides et al., , 2016. Finally, the variability between replicates in all terms included in the N 2 fixation calculation equation (as outlined in Montoya et al., 1996) may throw back minimum quantifiable rates values below estimated aphotic N 2 fixation rates (Gradoville et al., 2017). Propagating errors of the data (Birge, 1940), in five out of the nine aphotic N 2 fixation studies currently available, results in minimum quantifiable rates ranging from 0.01 to 2.7 nmol N L −1 d −1 (Table S1), suggesting that most of the aphotic N 2 fixation rates published are significant.
The potentially high budgetary significance of aphotic N 2 fixation to the global nitrogen budget calls for further studies that will establish the geographical and temporal distribution of aphotic N 2 fixation and consolidate the volumetric rates published thus far. In future studies, we encourage researchers in the field of marine nitrogen cycling to place emphasis on documenting N 2 fixation in the aphotic ocean and identifying environmental drivers of aphotic N 2 fixation: including oxygen, dissolved organic matter availability and particle colonization (Riemann et al., 2010;Benavides et al., 2015;Bombar et al., 2016). The availability of more data is essential to facilitate modeling and assessment of the distribution and magnitude of aphotic N 2 fixation in the global ocean (i.e., association with water masses, ecotones or density fronts). Eventually, a more comprehensive understanding of the ecophysiology of aphotic N 2 fixers and their contribution to global nitrogen input, will reveal their ecological importance and may help answer such question as what are the evolutionary advantages of the energetically-expensive process of N 2 fixation in an environment rich in dissolved inorganic nitrogen, and how does it affect oceanic carbon sequestration.

AUTHOR CONTRIBUTIONS
MB gathered N 2 fixation rates and made the calculations shown in the tables. MB, IB-F, SB, and LR wrote the article.

ACKNOWLEDGMENTS
MB and LR were supported by grant 6108-00013 from the Danish Council for independent research and by the BONUS BLUEPRINT project receiving funding from BONUS (Art 185) funded jointly from the European Union's Seventh Programme for research, technological development and demonstration and from The Danish Council for Strategic Research (LR).

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmars. 2018.00108/full#supplementary-material Table S1 | Error propagation analysis of aphotic N2 fixation rates from various published and unpublished studies.