Heterotrophic denitrification vs . autotrophic anammox – quantifying collateral e ff ects on the oceanic carbon cycle

Heterotrophic denitrification vs. autotrophic anammox – quantifying collateral effects on the oceanic carbon cycle W. Koeve and P. Kähler IFM-GEOMAR, Leibniz-Institut für Meereskunde, Düsternbrooker Weg 20, 24105 Kiel, Germany Received: 8 February 2010 – Accepted: 14 March 2010 – Published: 16 March 2010 Correspondence to: W. Koeve (wkoeve@ifm-geomar.de) Published by Copernicus Publications on behalf of the European Geosciences Union.


Introduction
The importance and relative proportion of processes removing combined nitrogen from the marine environment is currently under discussion.There is evidence supporting the long standing view that heterotrophic denitrification dominates oceanic N loss, but also autotrophic anaerobic ammonium oxidation (anammox) has been reported to make up for large shares, or even the bulk, in certain waters (e.g.Thamdrup et al., 2006;Ward et al., 2009).Both processes convert fixed nitrogen into N 2 (Ward et al., 2007;Devol, 2008) and reduce the oceanic nutrient inventory in this way.Temporal changes of the nitrogen removal flux in the past (on glacial/interglacial timescales), or from present to future, are thought to influence the level of oceanic production and associated CO 2 fluxes (Altabet et al., 1995;Ganeshram et al., 1995;Codispoti, 1995) by tightening or relaxing Nlimitation of oceanic primary production and export.There are other aspects in which both processes differ (collateral effects, Voss and Montoya, 2009).One example is the formation of climate reactive gases, namely N 2 O (Jin and Gruber, 2003), which is an intermediate of denitrification (Yoshinari and Knowles, 1976) but not known as one of anammox.Here we focus on collateral effects of the trophic status of nitrogen loss processes on the carbon cycle, as recently proposed by Voss and Montoya (2009).
Their argument is the following.Denitrification is a heterotrophic process during which organic matter is consumed and CO 2 is released to ambient waters.Pelagic denitrification thus effects a potential short-circuit in the biological pump by producing CO 2 from organic matter which otherwise might descend deeper into the ocean to be stored there for longer.In contrast, anammox is an autotrophic process potentially increasing the efficiency of the biological pump by fixing additional carbon in intermediate waters and thus reducing net CO 2 production in the water column.It appears to be of importance to the carbon budget whether it Table 1.Stoichiometric equations for (1) dissimilatory nitrate reduction to nitrite (DNRN), (2) denitrification, (3) anammox, and (4) dissimilatory nitrate reduction to ammonium (DNRA) for bulk organic matter with an average composition of C a H b O c N d P e S f .For simplicity and following Paulmier et al. (2009) we give the stoichiometric equations in non-ionic forms.We assume reaction of NH 3 and CO 2 with water and subsequent dissociation as well as dissociation of HNO 3 , HNO 2 , H 3 PO 4 , and H 2 SO 4 according to seawater pH.For a more detailed discussion of the derivation of Eqs.(1), (2), and (3) see Paulmier et al. (2009) a The energy gain from the anammox reaction (3a) is used to drive the fixation of CO 2 into organic matter (biomass of anammox bacteria).Here we follow the suggestion of Strous et al. (1998) and Kuenen (2008) that nitrite is used in this reaction as the electron donor of CO 2 fixation (Eq.3b).Since the combined system of equations (3a, 3b) is underdetermined we are unable to provide a generic solution for x, y, z, and w.In our computations we use instead empirical values taken from the experimental work of Strous et al. (1998) is a heterotrophic process or an autotrophic one which dominates nitrogen loss processes in the ocean's water column.
In view of projected increases in the extent of oxygen minimum zones (Matear and Hirst, 2003;Oschlies et al., 2008;Hofmann and Schellnhuber, 2009), heterotrophy or autotrophy in relation to nitrogen losses taking place there would be of increasing importance, potentially providing a positive or negative feedback on the carbon cycle, respectively.In this short note we analyse the stoichiometries of suboxic nitrogen conversions and their effect on the carbon balance.

Background and definitions
Nitrogen in the ocean occurs in seven oxidation states and there are transformations between all, oxidations and reductions.Nitrogen serves both as a constituent of organic matter and nitrogen compounds are used as oxidants and reductants in dissimilatory reactions.Historically, a number of terms, and varieties of definitions of some, have been in use for many of these reactions.We will in the following use only four reactions, all relevant to nitrogen loss in suboxic environments: (1) dissimilatory nitrate reduction to nitrite (DNRN); (2) denitrification, the production of N 2 from nitrite (denitrification sensu strictu; Zumft, 1997), this is a heterotrophic process consuming organic carbon; (3) anammox, the combination of nitrite and ammonia to produce N 2 , which is an autotrophic process consuming CO 2 ; (4) dissimilatory nitrate reduction to ammonia (DNRA).Both DNRN and DNRA are heterotrophic.Formulas describing the bulk stoichiometries of these processes are given in Table 1.We use only these four definitions of suboxic nitrogen transformations to develop our points.There are numerous variations to these (incomplete reactions, shortcuts, combinations, byreactions) which can be of interest in special environments.
We confine the treatment to oxygen minimum zones (OMZ) which are the only pelagic realms in which nitrogen loss occurs (at [O 2 ]<5 mmol m −3 ; Devol, 2008).In the cores of OMZs, N transformations are based on the N inventory present plus anything which reaches them by sedimentation.
It is these that we start with (Sects.2.1 and 2.2).Later we shall consider the allochthonous supply of additional substrates by diffusion from the fringes of the OMZ, and other special situations encountered in the sea (Sects.2.3 and 3).The largest oxygen minimum zones (OMZ) meeting these low oxygen conditions are the intermediate to deep waters of the Arabian Sea and the Eastern Tropical South and North Pacific.Additional sites of suboxic nitrogen removal are enclosed seas like the Black Sea, the Baltic Sea and some fjords.While until recently all suboxic N 2 -production in the ocean has been ascribed to denitrification, it is now known that a number of biotic and abiotic nitrogen transformations contribute to nitrogen loss (Hulth et al., 2005).At present denitrification and anammox are considered the most important ones for N 2 production (e.g.Thamdrup et al., 2006;Ward et al., 2009).
Already during early work on denitrification, it had been observed that this process cannot account for all observed nitrogen loss.Ammonia liberated from organic matter during its heterotrophic consumption by denitrification and DNRN should accumulate in an oxygen-free environment, but it does not (Thomas, 1966;Cline and Richards, 1972;Codipoti and Christensen, 1985).Therefore a reaction involving the combination of NO − 3 and NH + 4 to produce N 2 has been invoked (Richards, 1965;Sen Gupta and Koroloff, 1973;Stumm and Morgan, 1996) and deduced from evolutionary and thermodynamical knowledge (Broda, 1977).Finally, a similar reaction has been observed in nature (Mulder et al., 1995;Thamdrup and Dalsgaard, 2002;Kuypers et al., 2003), the combination of NO − 2 and NH + 4 to form N 2 , which was called anaerobic ammonium oxidation (anammox).
During anammox NH + 4 and NO − 2 react in an approximately equimolar ratio (Table 1).Since oceanic OMZs are extensive lenses of oxygen free water surrounded by oxygen rich waters above, below and at least towards the open sea, and since NH + 4 and NO − 2 are usually scarce in these surrounding oxic waters (Zafiriou et al., 1992;Brzezinski, 1988), the major sources of the reactants of anammox must be autochthonous, i.e.NH + 4 and NO − 2 must be produced in the suboxic water body itself.Anammox therefore depends on nutrient regeneration for the supply of both its substrates (NH + 4 and NO − 2 ) (Ward et al., 2009).In principle, NO − 2 can be supplied by DNRN (Table 1) and NH + 4 may be liberated from organic matter broken down during DNRN or denitrification.The low production ratios of NH + 4 :NO − 2 of these reactions (compare Fig. 1b), however, allow only for a limited quantitative importance of anammox for N 2 production (see Sect. 2.2 for details).An alternative and additional autochthonous source of NH + 4 may be dissimilatory nitrate reduction to ammonium (DNRA; Kartal et al., 2007;Lam et al., 2009) which is associated with heterotrophy as well.
In this paper, we will refer to the conversion of fixed nitrogen (i.e. the sum of NO − 3 , NO − 2 , NH + 4 , and organic nitrogen) to nitrogen gas (N 2 ) in suboxic waters as "suboxic N 2 -production", irrespective of the pathways or agents (organisms) involved.Different stoichiometries of suboxic nitrogen conversions have been discussed in the literature, differing by the composition of the organic matter utilized and the fate of remineralised nitrogen (e.g.Richards, 1965;Canfield, 2006;Paulmier et al., 2009).In the following section we will present the bulk stoichiometries of two possible systems, one consisting of combinations of DNRN, denitrification and anammox (i.e. a system where heterotrophic denitrification necessarily dominates N 2 production) and an alternative system where DNRN, DNRA and anammox co-exist (i.e. a system where autotrophic anammox is the exclusive process forming gaseous nitrogen).We will also briefly discuss to what extent and under which specific conditions allochthonous sources of substrates can be relevant and evaluate their maximum effect on the trophic state of the suboxic 4 but variable fractions of NO − 2 accumulate.On the x-axes we plot the property "1 -NO − 2 accumulated : NO − 3 −consumed ".We interpret this property as the efficiency of the overall N-conversion process where the value of one represents the condition of a fully efficient conversion of NO − 3 to N 2 (i.e.all NO − 2 is used up).Solid lines are for a mean composition of respired organic matter of C 106 H 175 O 42 N 16 P (Anderson, 1995), dashed lines for respiration of pure proteins (C 3.83 H 6.05 O 1.25 N, Laws, 1991;Anderson, 1995).(a) Fraction (in percent) of total N 2 -production which is due to anammox.In the combined reactions of scenario I the remainder to 100 percent is due to denitrification.(b) Ratio of production rates of NH + 4 and NO − 2 (mol:mol) during the coupled reactions of DNRN (providing NH + 4 and NO − 2 ) and denitrification (providing NH + 4 only) for the given boundary conditions (no NH + 4 accumulation) and the respective efficiencies of the overall N-conversion process (x-axes).Note that this ratio is always well below one, the stoichiometric ratio of NH + 4 and NO − 2 in anammox, indicating NH + layer.Our general subject will be to quantify the net ratio of CO 2 produced to molecular nitrogen formed ( CO 2 : N 2 ) given various combinations of the processes involved in suboxic N-conversions.

Stoichiometric constraints
First, let us consider the simple case that organic matter of standard oceanic composition (C 106 H 175 O 42 N 16 P; Anderson, 1995) is completely oxidized with nitrate to form CO 2 , N 2 and water according to Reaction (R1) (Canfield, 2006).
Complete oxidation here refers to the boundary condition that neither NH + 4 nor NO − 2 accumulate.This yields a ratio of organic carbon oxidized to nitrate consumed of close to 1 (106 C:104 NO − 3 ) and a gross ratio of CO 2 produced to molecular nitrogen formed ( CO 2 : N 2 ) of +1.77 (106 C:60 N 2 ).In suboxic waters no NH + 4 accumulates (Richards, 1965) and here we assume that the oxidation of NH + 4 is due to anammox.In this reaction 1 mol of NH + 4 combines with 1 mol of NO − 2 to form 1 mol of N 2 and water (Eq.3a in Table 1).Each mol of NH + 4 consumed supports the autotrophic fixation of about 0.07 mol of CO 2 (Strous et al., 1998;Tijhuis et al., 1993) yielding a molar CO 2 : N 2 ratio of anammox of 0.07.The electron donor required for the reduction of CO 2 is not well known.In aerobic ammonium oxidation, NH + 4 is the only reductant.In anammox, NO − 2 has been proposed as the electron donor resulting in NO − 3 as a product of CO 2 fixation (van de Graaf et al., 1996;Strous et al., 1998;Eq. 3b in Table 1).During experiments in a sequencing batch reactor of these authors the ratio NO − 2 consumed : NH + 4 consumed differed significantly from the 1:1 ratio, which is usually assumed for marine anaerobic ammonium oxidation (e.g.Kuypers et al., 2003).About 20% of the nitrite was converted to nitrate and the NO − 2 : NH + 4 ratio of the combined reaction (3a, 3b) was about 1.3:1.Under marine conditions, with substrate concentrations several orders of magnitude smaller than in the batch reactor experiments, a smaller CO 2 : NH + 4 is expected because of the energy requirements for maintainance.This results in a lower NO − 2 : NH + 4 ratio of the combined reaction (3a+3b).Even when assuming the published NO − 2 : NH + 4 ratio from batch reactor experiments to be valid for marine anammox, the effect on the nitrogen budget of the Nconversions is small.Consider the oxidation of a one mole P-equivalent of organic matter according to Reaction (R1) by DNRN+denitrification+anammox which consumes 104 mol of nitrate and implies the oxidation of 16 mol of NH + 4 with 16 mol NO − 2 due to anammox.The associated CO 2 fixation should consume another 4 mol of nitrite and yield 4 mol nitrate, thus replenishing only about 4% of the nitrate consumed during DNRN+denitrification.
Using generic stoichiometric equations to describe the possible reactions which contribute to suboxic N 2production (Table 1) we can quantify the proportions in which the individual reactions involved (DNRN, denitrification, anammox) are required for a variety of bulk organic matter compositions (Table 2) and for a range of boundary conditions (fraction of accumulating intermediate NO − 2 ).For the mean organic matter composition given above, the condition of complete conversion of fixed nitrogen to N 2 , is met if 1 mol P-equivalent of organic matter is remineralised through DNRN, 1.27 mol P equivalents of organic matter through denitrification and if the 2.27•16 mol NH + 4 produced in these heterotrophic reactions are oxidized with NO − 2 to form N 2 via anammox.In this scenario about 73% of the N 2 produced is by denitrification and 27% by anammox (Table 2).The respective autotrophic CO 2 fixation is 2.54 (0.07•2.27•16) mol and the bulk CO 2 : N 2 ratio for the combined heterotrophic and autotrophic processes changes to +1.75.This is, for all practical purposes, indistinguishable from the gross ratio (+1.77) which does not account for the autotrophic carbon fixation.The net CO 2 : N 2 ratio for the complete conversion of fixed nitrogen to N 2 may vary between 1.58 and 1.90, depending on the composition of organic matter (Table 2).
Significantly higher contributions of anammox to N 2 production of up to 100% have been suggested from tracer experiments (Kuypers et al., 2005;Thamdrup et al., 2006;Hamersley et al., 2007).With a combination of DNRN, denitrification and anammox (scenario I, Figs.1-3) this can be achieved if nitrite accumulates (Fig. 1a).Nitrite accumulation is a characteristic of the upper margin of oxygen minimum zones (Cline and Richards, 1972;Sen Gupta and Naqvi, 1984;Codispoti and Christensen, 1985).The ratio of nitrite accumulating to nitrate consumed denotes the inefficiency of suboxic N 2 -production.We use the term "1 -NO − 2 accumulated : NO − 3 consumed ", i.e. the efficiency of suboxic N 2 -production, as the independent variable (x-axes) in Figs.1-5.Contrary to expectations, a higher contribution of anammox to total N 2 production goes along with an increase (and not a decrease or even turn in sign) of the ratio of CO 2 produced to N 2 formed ( CO 2 : N 2 , Fig. 2a).In the most extreme case (no denitrification, 100% anammox; high NO − 2 accumulation) the ratio is about +6.5, i.e. almost four times as high as for 100 percent efficient N 2 -production (Fig. 2a).This effect is due to the increased contribution of organic nitrogen to produced N 2 (Fig. 2b).The higher the contribution from anammox the more inefficient the suboxic N-removal becomes.
Direct and indirect effects of autotrophic CO 2 -fixation have a small impact on the integrated CO 2 : N 2 -ratio (Fig. 3).The direct effect (from CO 2 -uptake) is largest where the contribution of anmmox is highest.The indirect effect (i.e. the effect of NO − 3 -production on the x-value) is largest at moderatly low x-axes values (Fig. 3a).Taken together, the combined effect is within ±1% of the uncorrected CO 2 : N 2 (Fig. 3b).
Alternatively, OMZs may function as systems in which dissimilatory nitrate reduction to ammonium (DNRA) supplements the respiratory pathways of DNRN and denitrification in the production of ammonium to supply substrates to anammox (Lam et al., 2009;Eq. 4 in Table 1).In this case high shares of anammox in total N 2 -production may be achieved even with no or little nitrite accumulation, i.e. with highly efficient nitrogen removal.Here (scenario II, Fig. 4) we assume combinations of DNRA (major NH + 4 source), DNRN (prime source of NO − 2 and minor NH + 4 source), and anammox as the only process producing N 2 .Combining DNRA and DNRN in variable ratios yields a range of efficiencies of N 2 -production (x-axes) accompanied by varying NO − 2 -accumulation (again using the boundary condition that no NH + 4 should accumulate).Both DNRA and DNRN are heterotrophic.Figure 4a shows their relative contribution along the efficiency gradient expressed as the fraction of NH + 4 provided via DNRA, to the total flux of NH + 4 to anammox.High contributions of DNRA allow for highly efficient N-conversion while low efficiencies are found where NH + 4 provision from DNRA falls below 50%.Although in this scenario 100 percent of N 2 production is from the autotrophic anammox reaction for all possible efficiencies, the overall process (i.e. the combined net effects of DNRA, DNRN, and anammox) is clearly heterotrophic (Fig. 4b), with CO 2 : N 2 ratios almost indistinguishable from those given in Fig. 2a where DNRN, denitrification, and anammox co-exist.
Differences occur related to the quality of organic matter consumed during the N-conversions.Using protein instead of mean bulk organic matter, the CO 2 : N 2 ratio is somewhat lower (Figs.2a, 4b) and the yield of N 2 -N produced per nitrate molecule consumed is larger (Fig. 5b) with maximum values of 2 in the case of very inefficient N-conversion.The major difference, however, is in the molar PO 3− 4 : N 2 yield (Fig. 5a).For mean bulk organic matter of a composi-tion commonly used in global biogeochemical models (Paulmier et al., 2009), the PO 3− 4 : N 2 yield increases from about 0.02 mol P:mol N 2 (efficient N-conversion) to about 0.06 (highly inefficient N-conversion).If, however, mainly proteins were preferentially respired in OMZs as indicated by recent particle-flux and decay studies (van Mooy et al., 2002), the PO 3− 4 : N 2 yield should be much smaller and even approach zero (Fig. 5a).
Assuming that autochthonous substrates to the anammox reaction dominate in typical open ocean OMZs, we find that although anammox itself is autotrophic, the sum of processes providing substrates for anammox and/or denitrification in all possible combinations of DNRN, denitrification, DNRA and anammox is heterotrophic.The degree of this heterotrophy depends on the efficiency of N 2 -production.In a combination of DNRN, denitrification, and anammox it is actually positively correlated with the importance of anammox for N 2 production (Fig. 6).

Allochthonous substrate sources
So far we addressed a typical open-ocean OMZ bounded by oxic waters where substrates to anammox are autochthonous, i.e. produced within the OMZ.This is in particular relevant for NH + 4 , which appears to be limiting to anammox in a system characterized by DNRN, denitrification and anammox.Potential external sources of NH + 4 are anoxic waters or sediments located below suboxic zones and the primary ammonia maximum at the base of the euphotic zone.In this section we discuss the potential effects of allochthonous substrate sources for CO 2 : N 2 ratios.
In sediments or enclosed seas like the Black Sea, suboxic waters may sit on top of fully anoxic systems in which NH + 4 has accumulated which has been produced from organic matter remineralised by sulphate reduction (Codispoti et al., 1991).Here, diffusive flux provides for additional NH + 4 available to anammox in adjacent suboxic waters (Murray et al., 2005) with nitrate diffusing downwards from overlying oxic waters may provide additional nitrite or ammonium (Konovalov et al., 2008) to support anammox and/or denitrification.In a system like the Black Sea such allochtonous sources of substrates may dominate (Fuchsman et al., 2008).Assuming DNRN as the sole NO − 2 source and diffusive NH + 4 fluxes as the major NH + 4 supply of anammox in the suboxic layers of the Black Sea, the net CO 2 : N 2 ratio may be as low as 0.38 inside the suboxic layer.This is still heterotrophic, but to a much lesser degree than under the conditions discussed above.Heterotrophy may become even smaller when assum-  3. Effects of autotrophic CO 2 fixation on CO 2 : N 2 ratio (for scenario I).(a) Absolute anomalies (mol:mol).Dash-dotted line shows the direct effect of CO 2 fixation on CO 2 : N 2 as difference between CO 2 : N 2 (corrected for CO 2 -fixation) and CO 2 : N 2 (without this correction).Dashed line shows the indirect effect from NO − 3 production during anammox on CO 2 : N 2 as difference between CO 2 : N 2 (corrected for NO − 3 -production) and CO 2 : N 2 (without this correction).This indirect effect acts on values of the x-axes, not the CO 2 : N 2 itself.The solid line gives the combination of both effects.(b) Plot shows the relative anomaly (%), i.e. the combined anomaly due to CO 2fixation and NO − 3 -production from (a) devided by the fully corrected CO 2 : N 2 ratio times 100.
ing HS − to diffuse upward to combine with nitrate (Konovalov et al., 2008) producing NO − 2 by an autotrophic process.Under such conditions it is possible that all substrates for the anammox reaction are produced autotrophically.Also HS − may combine with nitrate producing N 2 (chemolithotrophic denitrification;Hannig et al., 2007;Brettar and  1991).Hence suboxic N 2 production, supplied with substrates from outside, may locally become fully autotrophic.However, diffusion of reduced substrates is accompanied by diffusive CO 2 -fluxes from the remote heterotrophic decomposition of organic matter by sulphate reduction, which drive the overall CO 2 : N 2 back into the positive range.
While sulphate reduction can supply NH + 4 to the suboxic layer from below, there is also the possibility of NH + 4 entering from above.The primary NH + 4 maximum at the base of the euphotic zone is a characteristic feature of open-ocean NH + 4 distribution (Brzezinski, 1988).Where surface pro- duction and carbon turnover are high like in upwelling regions, NH + 4 concentrations as high as 0.5 µmol/L have been observed in this layer (Gibb et al., 1999;Molina et al., 2005;Molina and Farías, 2009).It is under such conditions that also the lower slope of the primary NH + 4 maximum and the oxycline coincide, and diffusive fluxes of NH + 4 across the upper fringe of the OMZ may occur.Whether this is a significant NH + 4 source for suboxic anammox may, however, be debated.On thermodynamic grounds (e.g.Brewer and Peltzer, 2009) it can be argued that, assuming similar energy yields for (oxic) nitrification (to NO − 2 ) and (suboxic) anammox, nitrite concentrations larger than its oxygen equivalent (i.e. about 3/2* [O 2 ]) are needed for anammox to be more effective in oxidising NH + 4 than nitrification.However, kinetics will matter as well.CO 2 : N 2 vs. percent fraction of N 2 produced by the anammox reaction for scenario I (DNRN+denitrification+anammox).Solid line is for bulk organic matter compositions, dashed line for proteins.constants of aerobic ammonia oxidation and surge-uptake of substrate pulses have recently been observed in nitrifying archaea (Martens-Habbena et al., 2009).Unfortunately, 15N-isotope experiments of anammox studies have usually applied micro-molar tracer additions, often larger than the ambient substrate concentration, and therefore provide potential rather than in situ substrate uptake rates (Hamersley et al., 2007).It is therefore difficult to compare in situ kinetics of aearobic and anaerobic ammonium oxidation.From the thermodynamic argument given above we conclude that it appears more likely that low-oxygen nitrification stops at the NO − 2 level, providing NO − 2 rather than NH + 4 to anammox (e.g.Schmidt et al., 2002) via diffusion of substrates into suboxic layers.Anyway, the NH + 4 invading suboxic waters from above is of heterotrophic origin from the oxic remineralisation of organic matter and hence should be accompanied by diffusive fluxes of respiratory CO 2 , similar as in an anoxic system underlying suboxic zones discussed above.This should drive the CO 2 : N 2 ratio of the upper margin of the OMZ back towards values computed for autochthonous substrate sources of anammox.

Discussion
Considering autochthonous sources of NH + 4 and NO − 2 to anammox and a coupled system with DNRN, denitrification and anammox, we find the somewhat counterintuitive relationship that the higher the contribution of autotrophic anammox to pelagic N 2 -production, the more heterotrophic the system is (Fig. 6).Hence the feedback switch proposed by Voss and Montoya (2009) to the effect that expending OMZs (Stramma et al., 2008;Oschlies et al., 2008) will either act as positive or negative feedbacks in the carbon cycle depending on whether anammox or denitrification dominate N 2production in OMZs does not exist.Including additional autochthonous NH + 4 sources from DNRA does not change the picture significantly.Even when combining DNRA, DNRN, and anammox in scenarios with anammox always contributing 100 percent to N 2 production, the coupled system is always heterotrophic.What appears to be variable in both systems is the degree of heterotrophy, however, depending on the efficiency of N 2 -production.
Allochthonous supply of NH + 4 (or NH + 4 and NO − 2 ) may contribute to the substrate needs of anammox, as has been observed in the Black Sea (Murray et al., 2005;Fuchsman et al., 2008;Konovalov et al., 2008).In such a situation, CO 2 : N 2 ratios in the suboxic layer are much lower than with autochthonous substrate supply, and hence the degree of heterotrophy is lower.However, the NH + 4 diffusing from anoxic waters underlying a suboxic system is from organic matter remineralised via heterotrophic sulphate reduction, which has a concomitant CO 2 production.Hence NH + 4 fluxes go along with CO 2 fluxes.NH + 4 and total dissolved sulfide (S T =H 2 S+HS − +S 2− ) as well as S T and total dissolved inorganic carbon (C T ) co-vary linearly over much of the anoxic water body of the Black Sea (Volkov and Rozanov, 2006).Averaging over anoxic waters from the upper 2000 m Volkov and Rozanov (2006) find S T -NH + 4 slopes of 4.29 and C T -S T slopes of 2.01, indicating an average C:N ratio of remineralisation of 8.6 which is close to that of bulk standard organic matter.Just below the suboxic layer, however, the HS − to NH + 4 slope is less (about 2) which if combined with the average C T -S T plot yields a C:N ratio of only 4.2.There is the possibility that this reduction in the apparent C:N remineralisation ratio can be explained as due to nitrogen-rich material (proteins) preferentially remineralised in the upper part of the anoxic layer.This has been suggested for other low oxygen waters by van Mooy et al. (2002).Alternatively, this difference in the apparent C:N ratio can be taken as another indication of the quantitative importance of anammox in close-by suboxic waters, providing a significant sink for NH + 4 but not for CO 2 , as evident from the observed low CO 2 :NH + 4 efficiency of the anammox reaction (Strous et al., 1998;Tijhuis et al., 1993).Though details will depend on the respective NH + 4 supplies by diffusion or autochthonous sources, respectively, the overall CO 2 : N 2 ratio should be larger than in the most extreme case computed above ( CO 2 : N 2 =+0.38) and approach the autochthonous ratio ( CO 2 : N 2 =+1.75).
Summarizing the above discussion, we find no simple relationship between the contribution of anammox to total N 2 -production and the degree of heterotrophy.In particular, where autotrophic anammox contributes 100 percent to suboxic N 2 -production, we find CO 2 : N 2 yields varying between about +2 and +6 for open ocean OMZs.oxic conditions.Hence there is no significant difference between suboxic and oxic systems of the aphotic zone of the ocean concerning their trophic state.

Fig
Fig. 6.CO 2 : N 2 vs. percent fraction of N 2 produced by the anammox reaction for scenario I (DNRN+denitrification+anammox).Solid line is for bulk organic matter compositions, dashed line for proteins.

Table 2 .
Bulk ratios for complete conversion of fixed nitrogen to N 2 (i.e.no accumulation of NO − 2 or NH + 4 ) for different compositions of organic matter.Bulk CO 2 : N 2 ratios include the effect of autotrophic CO 2 fixation (data for scenario I, with DNRN, denitrification and anammox, only).Ratio of denitrification to DNRN, in mol:mol of organic matter oxidized, respectively. a