Nitrogen loss processes in response to upwelling in a Peruvian coastal setting dominated by denitrification – a mesocosm approach

Upwelling of nutrient-rich deep waters make eastern boundary upwelling systems (EBUSs), such as the Humboldt Current system, hot spots of marine productivity. Associated settling of organic matter to depth and consecutive aerobic decomposition results in large subsurface water volumes being oxygen depleted. Under these circumstances, organic matter remineralisation can continue via denitrification, which represents a major loss pathway for bioavailable nitrogen. Additionally, anaerobic ammonium oxidation can remove significant amounts of nitrogen in these areas. Here we assess the interplay of suboxic water upwelling and nitrogen cycling in a manipulative offshore mesocosm experiment. Measured denitrification rates in incubations with water from the oxygen-depleted bottom layer of the mesocosms (via ¹⁵N label incubations) mostly ranged between 5.5 and 20 nmol N₂ L⁻¹ h⁻¹ (interquartile range), reaching up to 80 nmol N₂ L⁻¹ h⁻¹. However, actual in situ rates in the mesocosms, estimated via Michaelis–Menten kinetic scaling, did most likely not exceed 0.2–4.2 nmol N₂ L⁻¹ h⁻¹ (interquartile range) due to substrate limitation. In the surrounding Pacific, measured denitrification rates were similar, although indications of substrate limitation were detected only once. In contrast, anammox (anaerobic ammonium oxidation) made only a minor contribution to the overall nitrogen loss when encountered in both the mesocosms and the Pacific Ocean. This was potentially related to organic matter C $/$ N stoichiometry and/or process-specific oxygen and hydrogen sulfide sensitivities. Over the first 38 d of the experiment, total nitrogen loss calculated from in situ rates of denitrification and anammox was comparable to estimates from a full nitrogen budget in the mesocosms and ranged between ∼ 1 and 5.5 µmol N L⁻¹. This represents up to ∼  20 % of the initially bioavailable inorganic and organic nitrogen standing stocks. Interestingly, this loss is comparable to the total amount of particulate organic nitrogen that was exported into the sediment traps at the bottom of the mesocosms at about 20 m depth. Altogether, this suggests that a significant portion, if not the majority of nitrogen that could be exported to depth, is already lost, i.e. converted to N₂ in a relatively shallow layer of the surface ocean, provided that there are oxygen-deficient conditions like those during coastal upwelling in our study. Published data for primary productivity and nitrogen loss in all EBUSs reinforce such conclusion.

1 Introduction

Amongst the most productive marine ecosystems are eastern boundary upwelling systems (EBUSs), which are mainly fuelled by the wind-driven upwelling of dissolved inorganic nutrient-rich deep waters to the surface ocean, stimulating primary and associated higher trophic level productivity (Chavez and Messié, 2009; Kämpf and Chapman, 2016; FAO, 2018). This is particularly true for the Humboldt Current system off Peru (Montecino and Lange, 2009). High productivity and the eventual export of organic matter to depth result in marked oxygen consumption by aerobic respiration, leading to so-called subsurface oxygen-depleted zones and virtually anoxic oxygen minimum zones (ODZs and OMZs, respectively) at depth (e.g. Cline and Richards, 1972; Paulmier and Ruiz-Pino, 2009). In the absence of oxygen, heterotrophic organic matter decomposition can continue with alternative electron acceptors, such as nitrate (${\mathrm{NO}}_{\mathrm{3}}^{-}$) or nitrite (${\mathrm{NO}}_{\mathrm{2}}^{-}$) via nitric oxide (NO) and nitrous oxide (N₂O) to molecular nitrogen (N₂), in a series of separate steps carried out by a variety of bacteria (Zumft, 1997). In their entirety, these processes are summarised as denitrification. A by-product of denitrification is ammonium (${\mathrm{NH}}_{\mathrm{4}}^{+}$) from remineralised organic matter, but overall denitrification constitutes a net loss of bioavailable nitrogen (Paulmier et al., 2009). A second prevalent nitrogen loss pathway in ODZs and OMZs is autotrophic anaerobic ammonium oxidation (anammox) which utilises ${\mathrm{NH}}_{\mathrm{4}}^{+}$ and ${\mathrm{NO}}_{\mathrm{2}}^{-}$ produced by heterotrophic processes to produce energy for carbon dioxide (CO₂) fixation and organic matter production (Thamdrup et al., 2006; Brandes et al., 2007).

Observational and modelling studies have estimated the loss of bioavailable nitrogen via water column denitrification and anammox in suboxic ODZs and anoxic OMZs to amount to about 20 %–35 % of the total nitrogen losses ocean-wide (for reviews, see Bianchi et al., 2012; Zhang et al., 2020). In a context of climate/ocean change, which has been projected to enhance the temperature gradient between land and ocean and, hence, alongshore winds (Di Lorenzo, 2015), upwelling intensity, and the frequency of deep waters would subsequently increase (Hauri et al., 2013; Wang et al., 2015). Such a scenario has been put forward for the Humboldt Current system off Peru but also other EBUSs (Bakun and Weeks, 2008; Varela et al., 2015). Furthermore, due to increasing temperatures, the ocean loses oxygen (O₂), and OMZs are expanding (e.g. Bopp et al., 2002; Bograd et al., 2008; Stramma et al., 2008; Oschlies et al., 2017). Together with changes in microbial activity, this modifies the biogeochemical properties of upwelled waters, including, next to O₂, carbonate chemistry speciation, i.e. ongoing ocean acidification further decreases already low deep water pH levels (e.g. Feely et al., 2008; Franco et al., 2018; Schulz et al., 2019). Changes in the upwelling frequency and/or intensity, oxygen availability, temperature, and pH could influence planktonic food web functioning in EBUSs, with repercussions for nitrogen loss processes.

To better understand the events following the coastal upwelling of oxygen- and nitrogen-depleted deep waters, we make use of an offshore mesocosm set-up. This approach allows the simulating of upwelling and the tracing of biogeochemical element cycling, as well as associated trophic interactions. We were specifically aiming to address the question of nitrogen cycling, i.e. the build-up and turnover of organic nitrogen pools, their export from the surface to depth, and, most importantly, potential loss processes. Because such an approach enables budgeting of the various pools, it will be an alternative and independent assessment of the nitrogen balance in coastal ODZs next to classical shipboard transects.

2 Methods

The current experiment started on 22 February 2017 with the deployment of eight Kiel Off-Shore Mesocosms for Future Ocean Simulations (KOSMOS) about 4 nautical miles (nm) off the Peruvian coast, close to San Lorenzo Island at 12.0555^∘ S and 77.2348^∘ W, and ended on 16 April 2017 after 50 d of sampling. Full details on the experimental set-up, the sampling, manipulation, and maintenance schedule can be found in Bach et al. (2020a).

2.1 Mesocosm set-up and sampling

The KOSMOS system comprised eight mesocosms consisting of polyurethane bags 2 m in diameter extending to ∼ 18.7 m depth, with the last 2 m being a funnel-shaped sediment trap. After 2 d of thorough flushing and soaking of the bags, which had been pulled under the surface and were open to water exchange through a 3 mm mesh on both ends, the bags were secured above the water line, and the sediment traps were attached, enclosing about 54 m³ of a natural plankton assemblage. Sampling commenced on 26 February, i.e. day 1. Sampling days typically started in the morning between 06:30–07:00 PET (Peru time) by removing the material that had accumulated in the sediment traps with a pump, followed by CTD (conductivity, temperature, and depth) casts. The CTD (Sea & Sun Technology 90M memory probe) was additionally equipped with an optical oxygen sensor and a calibrated AMT (Analysenmesstechnik GmbH, Germany) amperometric hydrogen sulfide (H₂S) sensor (for details on the operation, the other sensors, and calibrations/corrections, see Schulz and Riebesell, 2013). CTD casts were followed by sampling with an integrating water sampler (IWS; Hydro-Bios). Because strong thermal stratification resulted in two distinct water masses at the surface (high oxygen and pH) and at the bottom (low oxygen and pH) of the mesocosms (Fig. 2a, b and c), two separate depth-integrated water samples were taken. According to changes in stratification, the depths were adjusted over time (days 1–2 – 0–5 and 5–17 m, surface and bottom, respectively; days 3–28 – 0–10 and 10–17 m, surface and bottom, respectively; days 30–50 – 0–12.5 and 12.5–17 m, surface and bottom, respectively).

Furthermore, to stabilise and maintain bottom water characteristics (as in the surrounding Pacific), a NaCl brine solution was homogeneously injected below 10 m on day 13 and below 12.5 m on day 33, increasing salinity, and, hence, stratification, by about 0.7 and 0.5 psu (practical salinity unit), respectively (see Bach et al., 2020a, for details).

While all seawater bulk parameters, such as particulate and dissolved organic matter or dissolved inorganic nutrients, were measured on the two depth-integrated samples (see Sect. 2.5 below for details), samples for N loss process incubations (see Sect. 2.3 below for details) were taken with a Niskin sampler from 15.25 m depth and treated as gas sensitive, i.e. filled into two 100 mL glass-stoppered DURAN bottles with at least one bottle volume of overflow, closed without headspace, and kept cold and in the dark (∼ 2–4 h) until processing. A direct comparison of oxygen concentrations in these samples by means of Winkler titrations, following the recommendations of Carritt and Carpenter (1966); Bryan et al. (1976), and Grasshoff et al. (1983) with CTD–oxygen–optode measurements, revealed an offset by +13 µmol L⁻¹ in the CTD data, although response time hysteresis had been corrected for τ=1 s, as described in Fiedler et al. (2013). Hence, oxygen concentrations at depth, as shown here from CTD casts, are likely to have been significantly lower. The most likely explanation for this offset is that the response time of the sensor was actually slower, i.e. of the order of 2–2.5 s.

Table 1 Concentrations of nitrate, nitrite, ammonium, dissolved organic nitrogen (DON), and dissolved organic phosphorus (DOP), and silicate (micromoles per litre, hereafter µmol L⁻¹) in the two deep water batches and associated inorganic N $/$ P $/$ Si in comparison to a Redfield ratio of $\mathrm{16}/\mathrm{1}/\mathrm{16}$ (see the text for details), and the inorganic N deficit calculated as ${\mathrm{N}}^{\ast }=\mathrm{DIN}-\mathrm{16}\times {\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$, with DIN denoting the sum of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ ${\mathrm{NO}}_{\mathrm{2}}^{-}$ and ${\mathrm{NH}}_{\mathrm{4}}^{+}$.

2.2 Deep water collection and addition

To simulate upwelling with two distinct OMZ signatures, in terms of N deficit, deep water was collected from 90 m at 12.044333^∘ S, 77.377583^∘ W, and 30 m at 12.028323^∘ S, 77.223603^∘ W, on days 5 and 10, respectively. However, both waters had quite a strong N deficit (N^∗) in comparison to a typical N $/$ P of $\mathrm{16}/\mathrm{1}$ required for phytoplankton growth (Redfield et al., 1963; Brzezinski, 1985), and will be referred to as “low N $/$ P” and “very low N $/$ P” treatments in the following (compare Table 1). The deep waters were collected into 100 m³ bags, without headspace, at the respective depths and sealed once brought back to the surface. Deep water was added by first removing about 20 m³ from each mesocosm and replacing it with the respective deep water that was injected into the bottom layer, between 14 and 17 m, on day 11, and the surface layer, between 1 and 9 m, on day 12. To minimise changes to deep water gas concentrations during injection, water was pumped from 2 m depth out of the deep water bags. For further details on collection and addition, see Bach et al. (2020a).

Figure 1 Schematic reaction diagram of major marine nitrogen cycling processes, with (1) hydroxylamine oxidation, (2) nitrifier denitrification, (3) denitrification, (4) anammox, (5) nitrification, (6) nitrate ammonification (DNRA), (7) anaerobic nitrite oxidation, and (8) nitrogen fixation. While processes 1–4 are considered nitrogen loss processes, 5–7 constitute neither a loss nor a gain, with (8) being the latter. Blue denotes oxic and red suboxic and/or anoxic processes. Please note that, while the reactions have been chemically balanced for electro-neutrality, the exact amount of H⁺, e⁻, or water produced and/or consumed will depend on the actual organism or enzyme and reaction pathway. See Bourbonnais et al. (2017), Codispoti et al. (2005), and Zumft (1997), and references therein, for details.

Figure 2 Temporal development of upwelling and stratification in the surrounding Pacific waters at the mooring site (water depth between 25 and 30 m), as evidenced by changes in CTD-derived (a) temperature (degrees Celsius), (b) salinity (practical salinity unit – psu), (c) oxygen saturation (percent) and (d) pH_T (total scale) depth profiles. Note that dips in salinity at the surface correspond with El-Niño-related torrential rain events further inland and a increased discharge of freshwater by the nearby river Rímac.

2.3 N loss processes incubations, measurements, and calculations

The two main N loss processes in oxygen-deficient waters off the Peruvian coast, i.e. denitrification and anammox (Fig. 1), were assayed with incubations of water from 15.25 m depth, using labelled ¹⁵NH₄Cl (≥ 98 atom %) or Na¹⁵NO₂ (98 atom %), i.e. by an addition of 3 µmol L⁻¹ each. During incubations with the former, anammox will be traced by the production of ²⁹N₂ (¹⁵N¹⁴N) and ³⁰N₂ (¹⁵N¹⁵N) if coupled to nitrification (although unlikely as it is an oxic process; compare Fig. 1). During incubations with ${}^{\mathrm{15}}{\mathrm{NO}}_{\mathrm{2}}^{-}$, anammox would again produce ²⁹N₂, while denitrification would produce both ²⁹N₂ and ³⁰N₂ (Fig. 1). Here, a further complication for rate calculations would be coupled nitrate ammonification, i.e. dissimilatory nitrate or nitrite reduction to ammonium (DNRA), also leading to ³⁰N₂ production via anammox (compare Fig. 1 and Holtappels et al., 2011). Incubations were in 12 mL glass double wadded exetainers (Labco Ltd.) in duplicates for each of the four time points, i.e. 0, ∼ 2, ∼ 7 and ∼ 20 h after label addition (in order to avoid potential bottle effects), in the dark, and inverted at a fairly constant 17±0.2 ^∘C, close to respective in situ conditions (similar to the protocols described in Dalsgaard et al. (2003), Ward et al. (2009), and Bourbonnais et al. (2021)). All sample handling, such as filling of the exetainers with 8 mL of labelled sample water using an Eppendorf Multipette E3, was carried out in a glove box that had been evacuated three times with a vacuum pump, followed by flushing with N₂ 4.5 gas (this was also done to the open exetainers). To reduce large background ²⁸N₂ levels and facilitate the detection of the small isotopic signal of labelled N₂ being produced during incubations, the exetainers were capped and sparged with Helium 5.0 at 3 psi for 6 min on a manifold that could hold all 16 exetainers of a single mesocosm. Based on previous calculations and measurements, such a set-up will replace about 20 times the volume of each exetainer (unpublished data). This is lower than the 24 times reported to ensure that the reduction in O₂ concentration is by less than 20 % compared to in situ conditions (Holtappels et al., 2011). A similar reduction would also be observed for most other gases (Wanninkhof, 1992), except those buffered by conjugate acid–base pairs, such as H₂S or CO₂, for which the reduction would have been even less. Incubations were stopped at each time point by the injection of 50 µL of a ZnCl₂ solution (50 % $w/v$), followed by thorough mixing.

Exetainer vials were stored upside down at room temperature in the dark. Prior to analysis, the headspace and water were equilibrated at room temperature by placing the exetainers on a platform mixer at 100 rpm (revolutions per minute) overnight. For measurements, 100 µL of headspace from each exetainer was injected, using an autosampler (Carvalho and Murray, 2018), into a PLOT (porous layer open tubular) gas chromatography (GC) column, at 2 mL min⁻¹, housed in a TRACE GC oven and interfaced with a Thermo Scientific DELTA V Plus mass spectrometer via a GC Combustion III unit, followed by a liquid nitrogen trap. The latter minimises interference by CO (a constituent in air) and NO (a secondary product during ionisation of water and/or oxygen, followed by the production of reactive oxygen species and recombination with nitrogen) due to imperfect GC column peak separation. The mass spectrometer was calibrated for N₂ concentrations by injections of known amounts of air.

Rate calculations were straightforward, as in most ${}^{\mathrm{15}}{\mathrm{NH}}_{\mathrm{4}}^{+}$ incubations no enrichment over time in ²⁹N₂ was detected, indicating the absence of anammox. Hence, denitrification in the ${}^{\mathrm{15}}{\mathrm{NO}}_{\mathrm{2}}^{-}$ incubations (and because ³⁰N₂ was quite noisy) was calculated from the increase in ²⁹N₂ and the known ratio of labelled to unlabelled ${\mathrm{NO}}_{\mathrm{2}}^{-}$ (in the rare cases where anammox was detected in the ${}^{\mathrm{15}}{\mathrm{NH}}_{\mathrm{4}}^{+}$ incubations, denitrification was corrected for by subtraction). Rates were calculated from the linear regression slopes (see Fig. S1 in the Supplement) of the increase in the overall amount of N removed (as measured by areas ²⁸A and ²⁹A and associated δ¹⁵N), which was determined by the following conversions (modified from Thamdrup et al., 2006; Holtappels et al., 2011):

$$\begin{array}{}\text{(1)}& \begin{array}{rl}& {{\mathrm{N}}_{\mathrm{removal}}}_{t\mathrm{1}-\mathrm{3}}\left(\mathrm{\mu }\mathrm{mol}{\mathrm{N}}_{\mathrm{2}}{\mathrm{L}}^{-\mathrm{1}}\right)\\ & =\left[{\left({}^{\mathrm{15}}r{\displaystyle \frac{\left({}^{\mathrm{28}}\mathrm{A}+{}^{\mathrm{29}}\mathrm{A}\right)c{f}_{\mathrm{MS}}\frac{V_{\mathrm{h}}}{V_{\mathrm{a}}}}{f_{{{\mathrm{N}}_{\mathrm{2}}}_{\mathrm{air}}}c{f}_{{}^{\mathrm{15}}\mathrm{N}}}}\right)}_{t\mathrm{1}-\mathrm{3}}\right.\\ & \left.-{\left({}^{\mathrm{15}}r{\displaystyle \frac{\left({}^{\mathrm{28}}\mathrm{A}+{}^{\mathrm{29}}\mathrm{A}\right)c{f}_{\mathrm{MS}}\frac{V_{\mathrm{h}}}{V_{\mathrm{a}}}}{f_{{{\mathrm{N}}_{\mathrm{2}}}_{\mathrm{air}}}c{f}_{{}^{\mathrm{15}}\mathrm{N}}}}\right)}_{t\mathrm{0}}\right]{\displaystyle \frac{\mathrm{1000}}{V_{\mathrm{i}}}},\end{array}\end{array}$$

with

$$\begin{array}{}\text{(2)}& {}^{\mathrm{15}}r={\displaystyle \frac{{}^{\mathrm{15}}R}{\mathrm{1}{+}^{\mathrm{15}}R}},\end{array}$$

and

$$\begin{array}{}\text{(3)}& {}^{\mathrm{15}}R={\displaystyle \frac{{\mathit{\delta }}^{\mathrm{15}}\mathrm{N}}{\mathrm{1000}}}{}^{\mathrm{15}}{R}_{\mathrm{air}}+\mathrm{1},\end{array}$$

where ¹⁵R and ¹⁵R_air denote the ratios of ${}^{\mathrm{15}}\mathrm{N}{/}^{\mathrm{14}}\mathrm{N}$ in a sample gas or air, respectively (the latter was determined from measured ²⁸N₂ and ²⁹N₂ by Junk and Svec, 1958, i.e. 0.00367647, and recommended by Coplen et al., 2002). δ¹⁵N is the resulting isotopic composition (‰), V_h, V_i, and V_a are the volumes of headspace, sample incubated, and analysed (millilitres), respectively. cf_MS is the determined calibration factor for each measurement run to convert mass spectrometer peak area (²⁸A+²⁹A) into abundance (µmol), ${f_{{\mathrm{N}}_{\mathrm{2}}}}_{\mathrm{air}}$ is the fraction of N₂ found in equilibrium in the 4 mL of headspace (V_h) in relation to the total amount, including the 8 mL of sample water (V_i), in an Exetainer, and t1–3 or t0 refer to the respective incubation sampling times. The conversion factor ${{f_{\mathrm{N}}}_{\mathrm{2}}}_{\mathrm{air}}$ was calculated from N₂ solubility (Hamme and Emerson, 2004) for a lab temperature of 21 ^∘C and a salinity of 35. To extrapolate from measured removal of ¹⁵N to total N a conversion factor, $c{f}_{{}^{\mathrm{15}}\mathrm{N}}$ was calculated, taking into account the availability of labelled and overall substrate(s) together with the likelihood of ²⁹N₂ production, i.e. 2F_N(1−F_N), for denitrification incubations (binomial), and F_N, for anammox, with F_N denoting the ratio of labelled to total [${\mathrm{NO}}_{\mathrm{2}}^{-}$] or [${\mathrm{NH}}_{\mathrm{4}}^{+}$], respectively.

It is noted that there have been studies which found discrepancies between denitrification calculated using ²⁹N₂ as above, and ³⁰N₂, due to non-binomial distributions (De Brabandere et al., 2014; Chang et al., 2014). This has been attributed to so-called intracellular ${\mathrm{NO}}_{\mathrm{3}}^{-}$ $/$ ${\mathrm{NO}}_{\mathrm{2}}^{-}$ shunting, which leads to an error in the calculations of labelled to unlabelled substrate, based on known additions and measured seawater concentrations. With respect to noisy ³⁰N₂ data, we cannot check if that was an issue here; yet, it would lead to an underestimation of denitrification rates in both cases. Given the good agreement between our rate measurements and a full nitrogen budget (see Sect. 3.2 for details), it, however, appears that potential ${\mathrm{NO}}_{\mathrm{3}}^{-}$ $/$ ${\mathrm{NO}}_{\mathrm{2}}^{-}$ shunting did not affect our rate measurements significantly.

2.4 In situ substrate limitation of denitrification, total N loss calculations, and orni-eutrophication

In more than half of the denitrification incubations, measured N loss over a period of about 20 h was higher than the combined concentrations of in situ ${\mathrm{NO}}_{\mathrm{3}}^{-}$ and ${\mathrm{NO}}_{\mathrm{2}}^{-}$. This indicates that the 3 µmol L⁻¹ addition of Na¹⁵NO₂ alleviated substrate limitation, and rates in the incubations were higher than theoretically sustainable at in situ conditions. Furthermore, acknowledging that rate measurements at different substrate concentrations in comparison to in situ conditions are always potential rates, we estimated in situ rates from a Michaelis–Menten kinetic by first calculating the maximum rate, V_max, as follows:

$$\begin{array}{}\text{(4)}& V_{\max }={\displaystyle \frac{{\mathrm{Rate}}_{\mathrm{meas}.}\left([S{]}_{\text{in situ}}+\left[{}^{\mathrm{15}}{\mathrm{NO}}_{\mathrm{2}}^{-}\right]+K_{\mathrm{1}/\mathrm{2}}\right)}{[S{]}_{\text{in situ}}+\left[{}^{\mathrm{15}}{\mathrm{NO}}_{\mathrm{2}}^{-}\right]}},\end{array}$$

which was followed by the realised rates at in situ substrate concentrations:

$$\begin{array}{}\text{(5)}& {\mathrm{Rate}}_{\text{in situ}}={\displaystyle \frac{[S{]}_{\text{in situ}}{V}_{\max }}{[S{]}_{\text{in situ}}+K_{\mathrm{1}/\mathrm{2}}}},\end{array}$$

with [S]_{in situ} referring to in situ substrate concentrations of ${\mathrm{NO}}_{\mathrm{2}}^{-}$ and ${\mathrm{NO}}_{\mathrm{3}}^{-}$, $\left[{}^{\mathrm{15}}{\mathrm{NO}}_{\mathrm{2}}^{-}\right]$ referring to the 3 µmol L⁻¹ label addition, and $K_{\mathrm{1}/\mathrm{2}}$ referring to the reaction half-saturation constant. As we have not performed any substrate vs. rate essays ourselves, and since there is only little information on the kinetics of water column denitrification, we adopted a conservative approach and chose the highest published $K_{\mathrm{1}/\mathrm{2}}$ of 5 µmol L⁻¹ (Michiels et al., 2019), over the 2.9 and 2.5 µmol L⁻¹ of Jensen et al. (2009) and Dalsgaard et al. (2013), respectively.

Total N loss by denitrification and anammox was calculated only for the first 38 d of the experiment. The reason was to keep the estimates by rate measurements comparable to estimates from a full N budget of the mesocosms, as, after day 40, massive perturbations in the latter became obvious. This was caused by birds aggregating on the roofs of the mesocosms and depositing nitrogen-rich faeces (orni-eutrophication). The onset of this perturbation was estimated by the sudden increase in total particulate phosphorus deposition rates in the sediment traps, most likely from excrement, as phosphate concentrations during this time were relatively constant (Fig. S2; for details see Bach et al., 2020a). Total N loss was then estimated by summing up hourly in situ rate estimates (Eq. 5) for each mesocosm (in case anammox was also detected, the measured rate was added to that of denitrification), factoring in the varying measurement intervals and, thus, filling in days without measurements. This hourly N₂ loss was then multiplied by 24 h d¹⁻ × 2 (conversion between N₂ to N) and divided by three (average contribution of bottom layer water to overall mesocosm volume), converting it to total N loss over the first 38 d of the experiment.

2.5 Ancillary measurement parameters

Dissolved inorganic nutrient concentrations in seawater, i.e. ${\mathrm{NO}}_{\mathrm{3}}^{-}$, ${\mathrm{NO}}_{\mathrm{2}}^{-}$, and ${\mathrm{NH}}_{\mathrm{4}}^{+}$ (together referred to as DIN) and ${\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ and Si(OH)₄ were measured colourimetrically on a segmented flow analyser (QuAAtro, SEAL Analytical) on site (for details, see Bach et al., 2020a).

Dissolved organic nitrogen (DON) was calculated from a mass balance by subtracting measured DIN from total dissolved nitrogen (TDN) concentrations. The latter was also determined by segmented flow analysis after an oxidising step with Oxisolv (Merck) for 30 min in an autoclave.

Particulate organic nitrogen and carbon (PON and POC, respectively) concentrations in the water column were determined by filtering known amounts of seawater over precombusted glass fibre filters (GF/Fs) that were stored frozen until analysis on an elemental analyser (EA; EuroVector EA 3000; for details, see Bach et al., 2020a).

The removal of nitrogen from the water column via sedimentation (N_sed) was determined on material collected every other day from the sediment traps which was quantitatively precipitated, freeze-dried, weighed, and then an aliquot was measured on an elemental analyser (EuroVector EA 3000; for details, see Bach et al., 2020a and Boxhammer et al., 2016 again).

2.6 Statistical analyses

In order to assess what drives denitrification rates, stepwise multiple linear regressions (MLRs) with interaction terms were carried out. Prior to the regressions, outliers in estimated in situ denitrification rates (see Sect. 2.4 and Table 2 for details) were identified in a box-and-whisker plot and removed, i.e. five values with rates higher than 10 nmol L⁻¹ h⁻¹. In order to avoid overfitting and find a balance between model complexity and explanatory power, we followed a backward elimination process, starting with the full seven potential measured predictors, i.e. PON, PON_sed, DON, ${\mathrm{NO}}_{\mathrm{3}}^{-}$, ${\mathrm{NO}}_{\mathrm{2}}^{-}$, O₂, and H₂S, all measured in the bottom layer of the mesocosms (with the exception of sedimenting PON_sed). Note that we have opted to not include POC, which is highly co-correlated with PON, and DOC, which was not measured in the bottom layer (but probably would have a similar issue). In a next step, MLRs with all possible combinations of six potential predictors out of the overall seven and their resulting R² were calculated, which was followed by further MLRs, subsequently reducing the number of predictors each time by one (a total of 119 MLRs were fitted). Calculations were performed using the functions “boxplot”, “stepwisefit”, and “plotEffects” in MATLAB.

Table 2 Calculated in situ denitrification rates (nanomole dinotrogen per litre per hour; hereafter nmol N₂ L⁻¹ h⁻¹) in the low N $/$ P and very low N $/$ P deep water addition mesocosms and the surrounding Pacific (see Sect. 2.4 for details), together with those for anaerobic ammonium oxidation (anammox^∗) (nmol N₂ L⁻¹ h⁻¹), when encountered, at various days. N loss refers to the total amount of nitrogen (micromole nitrogen per litre; hereafter µmol N L⁻¹) being lost through these processes for the first 38 d, for which an N budget can be calculated (compare Fig. 4). N loss was calculated for each mesocosm until day 38 by taking into account the varying measurement intervals and assuming an average contribution of bottom water to the overall mesocosm volume of one-third. For details on calculations, see Sect. 2. Note: SD – standard deviation. Note: anammox rates in the table below are shown in italics and with an asterisk to distinguish them from denitrification rates.

3 Results

The experiment took place during the 2017 coastal El Niño, which was characterised by three significant surface ocean warming events throughout January to April (Garreaud, 2018), with the last two, at the end of February and mid-March, clearly evident by the water surface temperatures above 22^∘ C at our mooring site (Fig. 2a). The El Niño was accompanied by torrential rains further inland, which was reflected by periods of significant reductions in surface ocean salinity (Fig. 2b), coinciding with water discharge of more than twice the typical rates of the nearby river Rímac (Fig. S3). During our experiments, there were, however, also periods of deep water upwelling, as evidenced by colder surface ocean temperatures, and reduced oxygen saturation states and pH levels reaching down to ∼ 30 % and 7.5 (total scale), respectively (Fig. 2c, d).

3.1 Temporal changes in oxygen, inorganic nitrogen, and organic nitrogen, as well as hydrogen sulfide, in the bottom layer of the mesocosms

Thermal stratification of the Pacific Ocean and the mesocosm waters during the initial phase of the experiment meant strict separation of well-oxygenated surface (∼ 0–10 m) from oxygen-depleted bottom waters (Fig. 2c). And while in the surrounding Pacific bottom waters oxygen remained depleted throughout, corresponding oxygen levels in the mesocosms started to increase (Figs. 2 and 3e). This was caused by the cooling of the surface waters due to upwelling in the surrounding Pacific and the resultant mixing of surface and bottom waters in the mesocosms. Such artificial behaviour was mitigated by increasing bottom layer salinity and, therefore, stratification on days 13 and 33, which brought oxygen concentrations at depth down to typical Pacific levels again (Fig. 3e).

Figure 3 Temporal evolution of depth-integrated bottom layer (a) PON, (b) DON, (c) nitrate (${\mathrm{NO}}_{\mathrm{3}}^{-}$), and (d) nitrite (${\mathrm{NO}}_{\mathrm{2}}^{-}$), together with CTD-derived (e) oxygen (O₂) and (f) hydrogen sulfide (H₂S) concentrations at the Niskin sampling depth in the mesocosms (M1–M8) and the surrounding Pacific. Blue and red denote the low N $/$ P and very low N $/$ P deep water additions, respectively (see Sect. 2.2 for details). The dashed green lines denote deep water additions on days 11 and 12, while the dotted black lines denote additions of a brine solution to the bottom layer to increase salinity, strengthen stratification, and reduce mixing. See Sect. 2 for details.

All mesocosms started with significant amounts of dissolved inorganic nitrogen present as ${\mathrm{NO}}_{\mathrm{3}}^{-}$ and ${\mathrm{NO}}_{\mathrm{2}}^{-}$, which, however, were quickly depleted within the first 2 weeks (Fig. 3c, d). The deep water addition significantly increased ${\mathrm{NO}}_{\mathrm{2}}^{-}$ concentrations in the low N $/$ P treatments, although it made only a minor contribution to the overall nitrogen budget (see the following sections for details).

Initial particulate and dissolved organic nitrogen concentrations (PON and DON) at depth were similar in all mesocosms (between 6 and 10 µmol L⁻¹), and while there was no clear temporal trend for PON, DON saw a steady decline by about 30 % until day 40 (Fig. 3a, b).

H₂S concentrations at depth were in the micromolar range in all mesocosms and the surrounding Pacific and mostly oscillated between 3 and 10 µmol L⁻¹ (Fig. 3f), equivalent to 0.1–0.3 ppm. In contrast, concentrations in surface waters were mostly in the high to low nanomolar range (data not shown).

3.2 Rates of denitrification and anammox and overall nitrogen loss

Measured denitrification rates in the 24 h incubations were similar in all mesocosms, ranging between less than 1 to up to ∼ 80 nmol L⁻¹ h⁻¹ (Table S1). More importantly, however, towards the end, measured rates exceeded those theoretically sustainable by substrate availability, i.e. combined in situ ${\mathrm{NO}}_{\mathrm{3}}^{-}$ and ${\mathrm{NO}}_{\mathrm{2}}^{-}$ concentrations. In comparison, measured denitrification rates in samples from the surrounding Pacific were comparable, although more variable, between consecutive measurement days.

Estimates for in situ denitrification (Eq. 5) were similar to measured rates at the beginning of the experiment when substrate availability was higher than or close to the half-saturation constant for denitrification. Towards the end, at high to low nanomolar substrate concentrations, in situ estimates were significantly lower (compare Tables 2 and S1).

Anammox was only detected on day 12 in mesocosms 7 and 8 at rates significantly smaller than those typical for denitrification (Table 2). In contrast, anammox was occasionally detected in the surrounding Pacific Ocean, i.e. on days 8, 12, and 46.

When comparing potential nitrogen loss, calculated as the sum of in situ estimates of denitrification and anammox over the first 38 d prior to the onset of orni-eutrophication (see Sect. 2.4 and Table 2 for details), with estimates of total nitrogen losses in each mesocosm from a nitrogen budget approach (compare Table 2 and Fig. 4 and see the next section for details), they were similar, with a mean of 2.69±1.18 and 3.64±2.12 µmol L⁻¹, respectively. It is acknowledged, however, that there was no statistically significant correlation between the two approaches.

Figure 4 Representative example of a total nitrogen budget from mesocosm 7, considering all relevant pools such as particulate organic nitrogen (PON), dissolved organic nitrogen (DON), and dissolved inorganic nitrogen in the form of nitrate (${\mathrm{NO}}_{\mathrm{3}}^{-}$), nitrite (${\mathrm{NO}}_{\mathrm{2}}^{-}$), and ammonium (${\mathrm{NH}}_{\mathrm{4}}^{+}$), cumulative particulate organic nitrogen exported to the sediment trap (N_sed), and the net change to all the above-mentioned nitrogen species (with the exception of PON, for which there was no deep water data) by deep water addition (N_DW). Black horizontal markers denote the total amount of nitrogen in these pools, calculated as an average of 3 consecutive sampling days at the start of the experiment (days 1–3) and towards the end (days 36–40), prior to the onset of orni-eutrophication. The deficit in this N budget comprises all nitrogen loss processes, dominated by denitrification (compare Table 2). Please note that the initial dip in nitrogen inventory is probably the result of a lag phase of nitrogen settling into the sediment trap.

3.3 Nitrogen budget

Summing up all organic and inorganic nitrogen species (excluding dissolved gases) in the water column, i.e. PON, DON, ${\mathrm{NO}}_{\mathrm{3}}^{-}$, ${\mathrm{NO}}_{\mathrm{2}}^{-}$, ${\mathrm{NH}}_{\mathrm{4}}^{+}$, N_sed (nitrogen exported through the sediment traps), and N_DW (overall deep water addition changes to the various nitrogen pools), revealed a first phase in all mesocosms for which the budget did not seem to be closed, as evidenced by a constant decline in total N during the first 2 weeks (Fig. 4). This was followed by a steady increase in total N until day 40, which was still below starting levels in all mesocosms. This net loss reflects, as we will argue below, the combined nitrogen loss processes, such as denitrification and anammox, exceeding nitrogen fixation, which was less than 0.1 µmol L⁻¹ over the entire water column in this period (Leila Kittu, personal communication, 2020). Finally, the last 10 d were characterised by a rapid increase in total N, mostly driven by rising surface PON concentrations, following continuous biomass production fuelled by external orni-eutrophication, i.e. the introduction of nitrogen-rich bird faeces to the mesocosms (see Bach et al., 2020a, for details).

3.4 Stepwise multiple linear regressions

In terms of R², there was a turning point in the number of potential main variables to explain in situ denitrification rates in stepwise multiple linear regressions (MLRs) with interactions. Above four to five independent variables, R² only increased slightly, while below it decreased relatively quickly (Fig. 5a). Both models with the highest R², for four and five main variables, identified ${\mathrm{NO}}_{\mathrm{2}}^{-}$, ${\mathrm{NO}}_{\mathrm{3}}^{-}$, and PON as those with the highest positive main effect sizes, with the remaining variables having relatively low effect sizes (Fig. 5b, d; Table 3). Finally, O₂ had a small negative main effect size in both models.

Table 3 Multiple linear regression statistics (standard error – SE; t and p values), describing denitrification in response to various environmental variables, for the best five-variable model (in terms of R²), with interactions (compare Fig. 5c), and the best four-variable model.

The resulting linear regression of in situ (compare Table 2) versus predicted denitrification rates slightly deviated from an ideal 1:1 relationship and had a small, yet positive, y axis intercept, meaning that lower denitrification rates would be over, while higher ones underestimated (Fig. 5c).

4 Discussion

A discussion on the unusual situation of a coastal El Niño during our study period, i.e. significant warming of surface waters, increasing stratification, and reducing upwelling intensity and/or frequency can be found elsewhere (Bach et al., 2020a). Most importantly, however, there were also multiple upwelling events in the surrounding Pacific during this time which were of similar magnitude to our experimental upwelling (Chen et al., 2021), providing the natural context.

Figure 5 Stepwise multiple linear regressions (MLRs) of in situ denitrification rates in the mesocosms against up to seven potential predictors and their interactions (PON, PON_sed, DON, ${\mathrm{NO}}_{\mathrm{3}}^{-}$, ${\mathrm{NO}}_{\mathrm{2}}^{-}$, O₂, and H₂S), with the (a) reducing numbers of measured variables and resulting maximum R². The red arrows denote the relatively small increase in maximum R² beyond five main variables and the relatively large decrease below, indicating this to be the sweet spot in terms of balancing the model complexity with predictive power and avoiding overfitting (see Sect. 2 for details). Numbers in parentheses denote the number of possible predictor combinations, i.e. MLRs fitted. (b) Main effect sizes of the stepwise MLRs with five main variables and the highest (0.8109) R² (compare Table 3). (c) Linear fit through in situ and predicted denitrification rates (nanomole dinitrogen per litre per hour; hereafter nmol N₂ L⁻¹ h⁻¹) by the MLR with five main variables. (d) Main effect sizes of the stepwise MLR with four main variables and the highest (0.7666) R² (compare Table 3).

4.1 Denitrification rates and nitrogen budgets

Denitrification rates measured in the incubations of water from the mesocosms and the surrounding Pacific of up to ∼ 80 nmol N₂ L⁻¹ h⁻¹ (although mostly well below, at a median of 12.4) were within the range of reported rates in Peruvian ODZ and OMZ waters (Farías et al., 2009; Dalsgaard et al., 2012). This was despite significantly higher oxygen concentrations here than typical for suboxic ODZs, i.e. less than ∼ 6 µmol L⁻¹ (Tyson and Pearson, 1991; Yang et al., 2017), although denitrification can be relatively insensitive to oxygen concentrations of up to at least 30–40 µmol L⁻¹ (Farías et al., 2009), which is higher than levels in our study for most of the time (see Fig. 3 but also Sect. 4.2.2).

Most interestingly, measured denitrification rates in incubations of mesocosm waters during the second half of the experiment exceeded those theoretically sustainable at in situ substrate availability (compare Sect. 2.4 and Table 2). Nitrification, supplying additional nitrite for denitrification, can operate at low micromolar (and even nanomolar) levels such as those found in our study (Bristow et al., 2016). However, a significant contribution to measured denitrification rates is unlikely as there and elsewhere (Peng et al., 2016; Santoro et al., 2021) nitrification rates were usually at least an order of magnitude lower than our measured denitrification rates. Hence, the reason for higher measured rates is most likely found in the fact that both ${\mathrm{NO}}_{\mathrm{3}}^{-}$ and ${\mathrm{NO}}_{\mathrm{2}}^{-}$ concentrations had dropped below 0.1 or even 0.01 µmol L⁻¹ during this stage, indicating that the denitrifying community was actually substrate limited. Such an observation was also made by Michiels et al. (2019) in a temperate fjord system. In comparison, denitrification rates in the surrounding Pacific waters had a higher day-to-day variability, and substrate limitation only occurred once. This is to be expected in a much more variable system characterised by frequent upwelling events of oxygen-depleted waters (Fig. 2) with relatively higher ${\mathrm{NO}}_{\mathrm{3}}^{-}$ and ${\mathrm{NO}}_{\mathrm{2}}^{-}$ concentrations (Fig. 3c, d).

Calculated Michaelis–Menten kinetic scaled estimates for in situ denitrification in the mesocosms and the surrounding Pacific were lower than measured rates during the incubations. This was particularly the case during the second half of the experiment, which reinforces the notion of substrate limitation.

The primary drivers of in situ denitrification rates appeared to be the availability of ${\mathrm{NO}}_{\mathrm{2}}^{-}$ and ${\mathrm{NO}}_{\mathrm{3}}^{-}$, followed by particulate organic matter (nitrogen), as indicated by multiple linear regression and effect size analysis (Fig. 5). This suggests that heterotrophic rather than chemolithotrophic denitrification was dominant (compare Sect. 4.2.2), as all three are substrates for the former process and eventually limit rates. The reason for ${\mathrm{NO}}_{\mathrm{2}}^{-}$ rather than ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations explaining the rates of denitrification in one of the MLRs could be found in the following. As denitrification from ${\mathrm{NO}}_{\mathrm{3}}^{-}$ to N₂ involves multiple and independent steps and organisms, the correlation between the N₂ production rate and a substrate concentration should become better the closer one comes to the end of this chain (Fig. 1). For example, there should be a perfect correlation between N₂O concentrations and N₂ production rate, and ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations and their turnover to ${\mathrm{NO}}_{\mathrm{2}}^{-}$ are meaningless if the intermediate steps to nitric and nitrous oxide are blocked or constitute a bottleneck. This would also contribute to the finding that denitrification rate measurements based on ${}^{\mathrm{15}}{\mathrm{NO}}_{\mathrm{3}}^{-}$ can be lower than those based on ${}^{\mathrm{15}}{\mathrm{NO}}_{\mathrm{2}}^{-}$ (Hamersley et al., 2007).

Overall nitrogen loss, calculated from in situ denitrification rates over the first 38 d of the experiment prior to the onset of orni-eutrophication, was comparable to an alternative estimate that was based on a full nitrogen budget in each mesocosm (Fig. 4; Table 2). The means of both methods, i.e. 2.69 ± 1.18 and 3.64 ± 2.12, were insignificantly different, although there was no statistically significant correlation when all mesocosms were compared. However, given that the nitrogen budget calculations involved a mass balance of seven entities with individual measurement uncertainties, this probably does not come as a surprise. There were also no statistically significant differences between the two deep water addition treatments, which was most likely due to relatively small differences in dissolved inorganic nitrogen in both waters in relation to the overall nitrogen pool in the mesocosms and a similar N deficit in both deep water batches (Table 1). However, N loss estimated from in situ denitrification rates in the low N $/$ P mesocosms was significantly lower than the budget estimate. It is noted, however, that in situ denitrification rates are a conservative estimate that potentially underestimate the true loss. For instance, adopting a lower half-saturation rate constant of 2 µmol L⁻¹ (Eqs. 4 and 5) would remove any statistically significant difference between the two approaches. Furthermore, ${\mathrm{NO}}_{\mathrm{3}}^{-}$ and ${\mathrm{NO}}_{\mathrm{2}}^{-}$ standing stocks most likely underestimate availability and, hence, in situ denitrification rates calculated here because of hidden turnover.

Finally, modelling exercises suggest that nitrogen loss in the water column is linked with mature El Niño/La Niña periods, with up to 70 % reduced rates during the former, most likely linked to increased water column oxygen concentrations at reduced upwelling and/or enhanced stratification (Deutsch et al., 2011; Yang et al., 2017), and increasing mesoscale turbulence and associated offshore nutrient export by eddies (Espinoza-Morriberón et al., 2017). This suggests that the rates observed here could actually be significantly higher in the more frequent La Niña periods.

4.2 Rates of anammox and the lack thereof in the mesocosms

Anammox and denitrification were first reported in the eastern tropical South Pacific (ETSP) in 2006 (Thamdrup et al., 2006), and the ETSP remains the most frequently sampled and most thoroughly characterised open ocean OMZ region. Anammox was the dominant N loss process in several studies, where denitrification was either not (Hamersley et al., 2007) or only sporadically (Kalvelage et al., 2013) detected. Subsequent studies in the ETSP detected both processes, and anammox was a substantial and mostly dominant part of the total N loss. There was an average of 78 % anammox along a Chilean coastal transect (Dalsgaard et al., 2012), 82 %–90 % at two high-resolution stations (De Brabandere et al., 2014), and 49 ± 20 % over four high-resolution stations (Babbin et al., 2020).

Anammox requires a source of ${\mathrm{NH}}_{\mathrm{4}}^{+}$ and ${\mathrm{NO}}_{\mathrm{2}}^{-}$, which must either be produced in situ by the remineralisation of organic matter in the absence of oxygen (as, for instance, by denitrification) or must be transported into the system from elsewhere, e.g. from adjacent sediments (Ward, 2013). When both processes are supported by organic matter remineralisation, then there is theoretical (Paulmier et al., 2009; Koeve and Kähler, 2010) and experimental evidence (Babbin et al., 2014) that the ratio of denitrification to anammox should be connected to the elemental composition of organic matter being decomposed, at least on larger spatial and temporal scales. The reasoning behind this connection is that complete anaerobic organic matter decomposition by denitrification produces ${\mathrm{NH}}_{\mathrm{4}}^{+}$ and N₂ in quantities mostly dependent on the carbon to nitrogen ratio (C $/$ N) of the organic matter being decomposed. Hence, this ratio dictates the denitrification (only utilising ${\mathrm{NO}}_{\mathrm{3}}^{-}$ and/or ${\mathrm{NO}}_{\mathrm{2}}^{-}$) to anammox (utilising both ${\mathrm{NO}}_{\mathrm{2}}^{-}$ and ${\mathrm{NH}}_{\mathrm{4}}^{+}$) ratio in a steady state (Babbin et al., 2014). However, DNRA could also supply ${\mathrm{NH}}_{\mathrm{4}}^{+}$ for anammox. Although there have been reports of DNRA in the ETSP (Lam et al., 2009), it is usually negligible in the water column (De Brabandere et al., 2014; Kalvelage et al., 2013) and rather restricted to shallow coastal systems dominated by sediments (Jensen et al., 2011). In the absence of DNRA, complete anaerobic decomposition of average phytoplankton-derived organic matter, i.e. C₁₀₆H₁₇₅O₄₂N₁₆P (Anderson, 1995), would require a ∼ 28 % contribution of anammox to overall N loss via these two processes (Babbin et al., 2014).

The reason for an anammox dominance in several studies mentioned above, despite above outlined stoichiometric constraints, could be partial denitrification of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ to ${\mathrm{NO}}_{\mathrm{2}}^{-}$ without the following steps leading to N₂ production, supplying both ${\mathrm{NO}}_{\mathrm{2}}^{-}$ and ${\mathrm{NH}}_{\mathrm{4}}^{+}$ for anammox (Lam et al., 2009; Jensen et al., 2011; Kalvelage et al., 2013; Peters et al., 2016). Furthermore, the ratio of ${\mathrm{NO}}_{\mathrm{2}}^{-}$ to ${\mathrm{NH}}_{\mathrm{4}}^{+}$ produced by organic matter decomposition of a certain C $/$ N ratio could also be influenced by anaerobic chemolithotrophic nitrite oxidation (compare Fig. 1), looping ${\mathrm{NO}}_{\mathrm{2}}^{-}$ back to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ for further partial denitrification and associated ${\mathrm{NH}}_{\mathrm{4}}^{+}$ production (Babbin et al., 2020).

In summary, anammox, in the absence of denitrification, is difficult to explain (although partial denitrification is not easily detected), while denitrification in the absences of anammox is rarely observed in the ETSP but presents no stoichiometric conflicts.

4.2.1 Organic matter C $/$ N in mesocosms

For the first 38 d of the experiment, particulate organic matter in the bottom layer of the mesocosms, where the N loss process incubation samples were taken from, had a C $/$ N of ∼ 7–8 (compare Bach et al., 2020a), which would not change the theoretical anammox contribution of ∼ 28 % by much. However, this ratio could be misleading as the particulate organic matter will be comprised by more and less labile fractions. A good proxy for that should be to just consider the freshly produced organic matter, i.e. the particulate and dissolved matter in the sunlit surface layer of the mesocosms in our case. Interestingly, just looking at C $/$ N of standing stocks for particulate organic matter, it was initially lower than in the bottom layer, indicative of preferential nitrogen remineralisation at depth, but increased to values of ∼ 10 in all mesocosms but one (no bloom of the dinoflagellate Akashiwo sanguinea; see Bach et al. (2020a), for details) after the deep water addition (Fig. S4). Similarly, C $/$ N in dissolved organic matter started already at about 10, increasing to up to 30 after the deep water addition, and when summed up to total organic matter, C $/$ N levels of up to 20 were reached (Fig. S4). Increasing C $/$ N from 6.625 to 20 would reduce the theoretical anammox contribution to only about 10 %. Furthermore, as already mentioned above, it should not be the standing stocks that are considered but the actual fresh production. Unfortunately, neither C nor N production were directly measured, but gauging the change in total organic carbon (TOC) concentrations after the deep water addition suggests that at least ∼ 100 µmol L⁻¹ were produced (Fig. S4). In contrast, total organic nitrogen (TON) concentrations hardly changed at all, suggesting even higher C $/$ N in freshly produced organic matter than in measured standing stocks. This could be the result of general dissolved inorganic nitrogen limitation and, hence, carbon overconsumption (Toggweiler, 1993) but also changes in the phytoplankton community composition (see Bach et al., 2020a, for details). It is noted, though, that changes in organic matter standing stocks over time are not necessarily a good indicator of the quantity and quality of fresh production due to unknown and potentially different turnover times for each element, although it is probably better than using standing stocks themselves. In this sense, the ODZ/OMZ signature of upwelled water, i.e. its nitrogen deficit, should influence the ratio of denitrification to anammox that is subsequently measured. Hence, regions with either one or the other process dominating could be indicative of the upwelling source water history.

4.2.2 Oxygen and hydrogen sulfide in mesocosms

Varying ratios of anammox to denitrification could also be the result of different sensitivities of both processes to prevailing oxygen concentrations. However, while there are a few experiments that have directly addressed the oxygen sensitivity of various N cycling processes, there appears to be no straightforward answer. For instance, in manipulative experiments, Jensen et al. (2008) found anammox at oxygen concentrations of up to ∼ 12.5 µmol L⁻¹, with denitrification not encountered at all. This is similar to Kalvelage et al. (2011), who observed anammox to cease above ∼ 20 µmol L⁻¹. And although no full denitrification was observed to at least N₂O or N₂O, the first step from ${\mathrm{NO}}_{\mathrm{3}}^{-}$ to ${\mathrm{NO}}_{\mathrm{2}}^{-}$ was found to occur even at the highest oxygen concentration of ∼ 25 µmol L⁻¹. This is, in turn, consistent with the nitrate reduction observed to at least N₂O at oxygen concentrations of up to ∼ 30–40 µmol L⁻¹ (Frey et al., 2020). In contrast, both anammox and denitrification can already be significantly or completely inhibited by oxygen additions of 3 or 8 µmol L⁻¹ (Babbin et al., 2014) or even lower (Dalsgaard et al., 2014). In essence, there is no clear indication of one process being more sensitive to oxygen than the other, which is potentially also related to variability in oxygen concentrations on small scales, i.e. in microenvironments around particles which are not captured by bulk seawater oxygen concentration measurements. And, indeed, there is evidence that microbial diversity in the OMZ is critically linked to particles, as are rates of denitrification and anammox (Ganesh et al., 2014, 2015).

The situation appears to be similar for H₂S, which, furthermore, has been observed to reach 10 µmol L⁻¹ and even higher in the coastal subsurface ETSP close to our study location (Callbeck et al., 2018, 2019). H₂S has been reported to completely inhibit anammox at 3 µmol L⁻¹ (Jensen et al., 2008), a concentration similar to that at the bottom layer of our mesocosms, while Dalsgaard et al. (2014) found no effect on N₂ production by anammox, or denitrification, at slightly lower levels of 1 µmol L⁻¹. Concerning denitrification, Dalsgaard et al. (2013) reported no effects of H₂S at even higher concentrations of up to 10 µmol L⁻¹, which is in the range of maximum concentrations found in our study, and there was stimulation in some instances, although this is most likely related to chemolithotrophic as opposed to heterotrophic denitrification (see Dalsgaard et al., 2013; Bonaglia et al., 2016, and references therein). Chemolithotrophic denitrification coupled to H₂S oxidation has also been hypothesised for coastal high H₂S stations off Peru by Kalvelage et al. (2013). Interestingly, by day 16 and onwards, the anammox functional marker gene hzo (Schmid et al., 2008) was not detectable anymore in any of the mesocosms (data not shown), which could be linked to relatively high H₂S concentrations, explaining the lack of anammox – at least for most parts of the experiment.

5 Conclusions: ODZ/OMZ nitrogen budget implications

The loss of, on average, 3–4 µmol L⁻¹ of nitrogen in our mesocosms fits well to zonal estimates by DeVries et al. (2012) for the ODZ in the eastern South Pacific between 75–100 m depth, where oxygen levels were similar to those encountered in our study (Chang et al., 2010), although the latter study reported a 2–3 times higher nitrogen deficit. Furthermore, the overall amount of nitrogen loss measured in the Pacific was at the upper end of ranges encountered in the mesocosms, potentially connected to substrate limitation in the mesocosms during the second half of the experiment (Table 2).

In contrast to shipboard measurements, the mesocosms offer the unique opportunity to put the various nitrogen pathways, i.e. loss as opposed to initial bioavailable standing stocks and export, into context. Between ∼ 1 and 5.5 µmol L⁻¹ of nitrogen were lost as N₂, representing up to 20 % of the initially bioavailable inorganic and organic nitrogen until day 38 of the experiment (compare Table 2 and Fig. 4). Interestingly, the amount of particulate organic nitrogen being exported below 20 m, i.e. the approximate depth of the sediment traps at the bottom of the mesocosms (compare Fig. 1 in Bach et al., 2020a), was in the same range. This indicates that, in the Peruvian EBUS, about half of the nitrogen that could be exported to depth would already be lost, i.e. converted to N₂, in a relatively shallow layer of the surface ocean, provided that there were oxygen-deficient conditions during the coastal upwelling as in our study (Fig. 2). Furthermore, over the entire pelagic water column, nitrogen loss of exported organic matter is likely to be even higher, suggesting that the majority of the dissolved inorganic nitrogen assimilated during new production (equalling export production on larger scales) should actually be lost in EBUSs.

Similar conclusions can also be reached by alternative means, i.e. by starting with global export production. Recent estimates based on observations and models range between ∼ 5 and 15 Pg C yr⁻¹, also including, next to the gravitation-driven biological pump, those by particle injection (see Boyd et al., 2019, and references therein). Assuming a Redfield molar C $/$ N of 6.625 (Redfield et al., 1963) would translate to 0.75–2.26 Pg N yr⁻¹ being exported. Furthermore, considering that about 5 % of global primary production and, hence, potential export production is located in the surface ocean of the four major EBUSs (Carr, 2002), the Arabian Sea (Vijayaraghavan and Krishna Kumari, 1989) and the Bay of Bengal (Poornima et al., 2020) above the ODZs and OMZs, about 38–113 Tg N yr⁻¹ could potentially be lost. In comparison, estimates of water column denitrification and/or anammox, i.e. 20 %–35 % of total marine losses of about 260 ± 100 Tg N yr⁻¹ (see Zhang et al., 2020, and references therein), then range between 52 ± 20 and 91 ± 35 Tg N yr⁻¹, indicating that a significant portion, if not the majority, of the exported nitrogen is indeed lost in ODZs and OMZs.

Nitrogen cycling in ODZs and OMZs currently plays a very important role in the overall marine nitrogen budget. However, the magnitude and direction of change in the actual nitrogen loss term in response to ongoing climate and ocean change (e.g. ocean stratification, acidification, and/or changes in temperature and oxygen levels) is uncertain. This issue is further complicated by uncertainties in future primary productivity and organic matter export estimates. For instance, depending on the representative concentration pathway, future export production could decrease as a result of changes to community structure (see Bindoff et al., 2019, for details and references therein). In summary, future changes in upwelling intensity and frequency, as well as the other potential biotic and abiotic factors mentioned above, could change the nitrogen (im)balance in ODZs and OMZs, having a significant impact on the overall marine nitrogen budget.