Marine Chemistry 88 (2004) 1 – 20 www.elsevier.com/locate/marchem Coupling the 15 N/14N and 18O/16O of nitrate as a constraint on benthic nitrogen cycling Moritz F. Lehmann a,*, Daniel M. Sigman a, William M. Berelson b a Department of Geosciences, Princeton University, Guyot Hall, Princeton, NJ 08544, USA b Department of Earth Sciences, U.S.C., Los Angeles, CA 90089-0740, USA Received 19 September 2003; received in revised form 19 February 2004; accepted 19 February 2004 Abstract We report 15N/14N and 18O/16O ratios of nitrate in benthic chamber incubations in the continental shelf sediments of the Santa Monica Bay (SMB) to deconvolve the effects of nitrification and denitrification. Estimates of denitrification rate from benthic flux stoichiometry range from 0.9 to 2.5 mmol N m 2 d 1. Between 46% and 100% of the total denitrification can be explained by nitrate or nitrite from nitrification. In general and independent of the denitrification rate and the fraction of remineralized N being denitrified, nitrate 15N/14N and 18O/16O ratios do not change significantly with progressive nitrate depletion in the chambers. A lack of change in nitrate 15N/14N could be due to either the lack of effective N isotope fractionation associated with sedimentary denitrification or the balancing of a denitrification isotope effect by the addition of low-15N/14N nitrate from nitrification. However, the lack of an increase in nitrate 18O/16O indicates that the isotopic fractionation specifically associated with sedimentary denitrification is, in fact, negligible. The coupled N and O isotope measurements also indicate that there is no significant gross efflux of 15N-depleted nitrate from nitrification, leading to the conclusion that nitrification is closely coupled to denitrification, even in the bioturbated sediments of the SMB. D 2004 Elsevier B.V. All rights reserved. Keywords: Nitrate isotopes; Sedimentary denitrification; Nitrification; Isotope effects; Benthic flux chamber; Santa Monica Bay 1. Introduction Denitrification, the bacterially mediated reduction of NO3 and NO2 to N2O and N2, is the major mechanism of fixed nitrogen loss from the ocean (Codispoti and Christensen, 1985; Seitzinger, 1988). It occurs in both the water column and sediments * Corresponding author. Tel.: +1-609-258-7544; fax: +1-609258-0796. E-mail address: [email protected] (M.F. Lehmann). 0304-4203/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.marchem.2004.02.001 where O2 concentration is f 5 Amol/l and lower (Codispoti et al., in press). Denitrification in continental margin sediments probably accounts for more than half of global ocean denitrification (Christensen, 1994; Brandes and Devol, 2002). Yet, the absolute rate of sedimentary denitrification is poorly constrained, with recent estimates ranging from 7.1 to 32.1 Tmol N yr 1 (100 to 450 Tg N yr 1) (Middelburg et al., 1996; Devol et al., 1997; Codispoti et al., 2001; Brandes and Devol, 2002). Denitrification, like other microbially mediated biogeochemical reactions, exhibits a significant bi- 2 M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 ological N and O isotope fractionation; that is, 14Nand 16O-bearing nitrate is preferentially consumed by the denitrifying organism, leaving the residual nitrate pool enriched in 15N (Cline and Kaplan, 1975; Liu and Kaplan, 1989) and 18O (Lehmann et al., 2003 and references therein). It appears that the N isotope effect, e, for denitrification in the ocean water column, where isotope fractionation should be nearly fully expressed, is more or less invariant in the global ocean, close to 25x(Brandes et al., 1998; Altabet et al., 1999; Voss et al., 2001; Sigman et al., 2003) (e={15k/14k 1}1000, where 15k and 14k are reaction rates heavier and for the lighter nitrate molecules, respectively). However, Bender (1990) predicted that the ‘‘effective’’ isotope fractionation (community fractionation) during microbial reactions in sediments should be markedly reduced with respect to the ‘‘intrinsic’’ (biological) fractionation, due to the fact that the oxidant (e.g., nitrate) is not consumed from a mixed pool, that is, partial substrate consumption occurs in deeper parts of the sediments where it is already enriched in the heavier isotope. Indeed, it was subsequently observed that sedimentary denitrification does not produce any significant apparent N isotope fractionation during nitrate loss (Brandes and Devol, 1997, 2002; Sebilo et al., 2003). Based on previous data from Puget Sound (Brandes and Devol, 1995), Brandes and Devol (1997) hypothesized that the lack of isotope effect is due to the limitation of denitrification by the diffusion of nitrate into reactive microsites. If both the N isotope effects associated with water-column and sedimentary denitrification are known, the 15N/14N of water-column nitrate can potentially be used to infer the relative importance of benthic versus pelagic denitrification, globally (Brandes and Devol, 2002) or in a specific environment (Sigman et al., 2003). These estimates are highly sensitive to the overall isotope effect of Fig. 1. Schematic of the benthic nitrogen cycle with sources and sinks of DIN, and intrinsic isotope effects (e) associated with N-cycling reactions (large arrows). Mineralization ( FMin) usually causes only a small fractionation between degrading organic N and NH+4 (Kendall, 1998), whereas nitrification and denitrification occur with significant intrinsic isotope effects. Yet, relative to intrinsic isotope effects, apparent isotope effects (as determined in the overlying water) may be highly reduced as a result of diffusion limitation. Large letters refer to selected fluxes of DIN and are used in the text. The ammonium – efflux ratio QNH4 defines the partitioning of remineralized ammonium (efflux versus nitrification). GCND is the fraction of newly produced nitrate being denitrified within the sediments, and hence represents a constraint on the degree of coupling between nitrification and denitrification in the sediments. The simple box model described in the text is based on a similar benthic DIN flux scheme. M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 benthic nitrogen cycling on the marine dissolved inorganic nitrogen (DIN) pool. Yet, there has been only limited study of the typical effective e for sedimentary denitrification and the isotopic impact of other benthic N cycle reactions on the overlying water column. Fig. 1 compiles the most important sources and sinks of nitrate in shallow marine sediments, which include the microbial oxidation (nitrification) of ammonium (Flux B) that is generated from organic matter mineralization (Flux FMin), the net diffusive flux of nitrate across the sediment – water interface (Flux Db Flux C), and the loss of nitrate due to microbial denitrification ( FDenit = Fluxes Dn + Db). Very often, nitrification rather than the diffusive flux from overlying waters is the major source of nitrate for denitrification within the sediments (Christensen et al., 1987; Seitzinger, 1988; Devol and Christensen, 1993; Middelburg et al., 1996; Jahnke and Jahnke, 2000), and nitrification can produce a net nitrate flux out of the sediments, even in actively denitrifying sedimentary environments. Because partial nitrification occurs with a marked isotope effect (e.g., Casciotti et al., 2002), newly produced nitrate can be significantly depleted in 15 N with respect to substrate ammonium. Indeed, Hartnett (1998) reported that, although nitrate was being consumed by denitrification in Washington and Mexican continental margin sediments, nitrate in overlying water became strongly depleted during incubation experiments, due to the efflux of low-15N/14N nitrate from nitrification within the sediments. Here, we report nitrate-isotope data from in situ incubation experiments conducted in the Santa Monica Bay (SMB). The main objective of this study is to use the combined measurement of 15N/ 14N and 18 16 O/ O (dual isotope approach) in benthic-chamber nitrate as a constraint on the role of microbial nitrate production within sediments, an uncertainty that has been recognized but not directly addressed in the estimation of the N isotope effect of sedimentary denitrification. Using a simple model, we explore the effect of variations in the relative importance of nitrification and denitrification on the benthic nitrate N and O isotope dynamics. Using the dual-isotope approach, we find that SMB sediments, despite active bio-irrigation, are ‘‘tight’’ with respect to nitrate; that is, a given nitrate ion that is transported in from bottom water or produced by nitrification within the 3 sediment will most likely be consumed before it has the opportunity to exit the sediment. 2. Sampling and methods Benthic chamber devices were deployed in the Santa Monica Bay in August 2000 and March 2001 at four sites, all of which are shallower than 65 m water depth (Fig. 2). The sediments on the SMB shelf are characterized by relatively high total organic carbon contents (5% to 8%). In general, bioturbation plays an important role in the SMB sediments and the solute transport by bio-irrigation may be a widespread phenomenon (Fraser and Berelson, unpublished data). Oxygen (O2) concentration within sediments were not determined, but modeled bioirrigation depths between 2 and 15 cm (Fraser and Berelson, unpublished data) may indicate O2 penetration depths of several centimeters. The O2 concentration in the bottom water ranged between 110 and 150 AM. Benthic fluxes were measured using techniques described in Berelson and Hammond (1986). Each chamber was sampled six times during deployment periods of 12 to 21 h. The O2 concentration within the chamber was monitored by a pulsed electrode with readings made every 6 min. In all experiments, chamber water remained oxic during the incubation period. At each site, bottomwater samples were collected with a 5 l Niskin bottle. Fig. 2. Benthic chamber deployment sites (Malibu, Topanga, Marina del Rey, Hyperion) in the Santa Monica Bay (California). 4 M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 Water samples were filtered (0.45 Am) and analyzed within 24 h or stored frozen for later analyses. Ammonium, nitrate, nitrite (in 2001 only), phosphate, and silicic acid were analyzed by colorimetric autoanalyzer techniques with replicate precision of 0.2, 0.1, 0.01, 0.05, and 1 Amol/l, respectively. Alkalinity (TA) was determined by titration using 5 ml sample aliquots, with a precision of 10 Amol/l. Total dissolved inorganic carbon (TCO2) was calculated from TA and pH. In addition, it was measured directly using a coulometer (UIC 5012). Most of the TCO2 fluxes reported here represent mean values combining the two methods, and the uncertainty in the mean was calculated two ways: (a) as the standard error of the mean and (b) as the uncertainty expected in the mean of a group considering the uncertainty of each value. The larger of these two uncertainties is reported. Oxygen electrode measurements were calibrated using Winkler titration. All chamber concentrations were corrected for dilution due to sample withdrawal from the chamber and subsequent volume replenishment with ambient water. Benthic fluxes were estimated from the derivative of the linear fit to the concentrations through time and the chamber volume, as determined from dilution of a Br-tracer injected early in the incubation. Solute concentration change through time was not always linear. Flux calculations were made using the initial linear portions of the concentration versus time plots. The uncertainty in the flux includes analytical precision of concentration measurements, the uncertainty associated with the linear regression, and the uncertainty in the determination of the benthic chamber volume. The N and O isotopic composition of dissolved nitrate + nitrite was determined with the ‘‘denitrifier’’ method of Sigman et al. (2001) and Casciotti et al. (2002) (henceforth, when reporting isotope ratios, the term ‘‘nitrate’’ includes the very small contribution, generally < 2%, from nitrite). In brief, sample nitrate was converted to nitrous oxide (N2O) by denitrifying bacteria that lack N2O-reductase activity, Pseudomonas chlororaphis (ATCC# 43928) or Pseudomonas aureofaciens (ATCC# 13985). N2O was stripped from the sample vial using helium as carrier gas, purified, and analyzed for its N and O isotopic composition. Oxygen atom exchange with H2O and blank size were measured for each sample run and corrected for following the procedure described in Casciotti et al. (2002). Oxygen exchange was always < 3% (for P. aureofaciens) and the blank was generally less than 0.3 nmol N2O (as compared to 10 –20 nmol in the sample). The use of P. aureofaciens allowed for the simultaneous measurement of nitrate 15N/14N and 18 16 O/ O ratios, whereas P. chlororaphis was only used for N isotope analysis (Sigman et al., 2001; Casciotti et al., 2002). There was good agreement between d15N measurements using P. aureofaciens and those using P. chlororaphis. Nitrogen and oxygen isotope ratios are reported in the conventional d-notation with respect to atmospheric N2 (AIR) and V-SMOW, respectively: Rsample dsampleðxÞ ¼ 1 1000 ð1Þ Rstandard where R represents the 15N/14N or 18O/16O ratios, respectively. Isotope values were calibrated using IAEA-N3, an international KNO3 reference material with an assigned d15N of + 4.7x(Gonfiantini et al., 1995) and reported d18O of + 22.7xto + 25.6x (Revesz et al., 1997; Silva et al., 2000; Lehmann et al., 2003; Böhlke et al., 2003). In this study, we adopt a d18O of + 22.7x . Based on replicate measurements of standards and samples, the analytical precision for d15N and d18O was generally better than F 0.2xand F 0.3x(1 S.D.), respectively. 3. Results and discussion 3.1. Benthic flux stoichiometry and denitrification The overall trends in concentration –time data are consistent with metabolic oxidation of sedimentary organic matter during early diagenesis and the associated remineralization of organic carbon, nitrogen, and phosphorous (Fig. 3). The sediments represent a sink for dissolved O2 and a source of TCO2, ammonium, and phosphate (Table 1). Measured nitrate fluxes suggest that, in general, the SMB sediments are a net nitrate sink (negative fluxes in Table 1). In some cases, however, the nitrate concentration within the benthic chamber increased slightly during benthic chamber incubation or did not change significantly. Whether nitrate is released from or consumed within the sediments indicates the relative importance of N-producing and N-consuming processes. Nitrite fluxes out of M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 5 Fig. 3. Examples of flux-chamber time-series from Station 1 at the Hyperion site in March 2001. Bottom-water solute concentrations are shown on the vertical axes (t = 0). The different symbols represent different chambers deployed at the same station. Shown TCO2 data were determined coulometrically. the sediment generally contribute less than 3% to the overall DIN flux; hence, they are not included in the subsequent discussion of the benthic N budget. The rate at which organic N is remineralized during organic matter decomposition in the sediments ( FMin) is not measured directly. We use diagenetic stoichiometry to estimate this flux, with the relationship between TCO2, PO43 , and ammonium fluxes being set by assuming ‘Redfield’ behavior (C:N:P = 106:16:1; Redfield, 1958). As outlined below, we found that using TCO2 fluxes is generally the most viable approach to predict the flux of N from organic matter decomposition and subsequently derive denitrification rates. To calculate the latter, we adapted the concept of N* (Gruber and Sarmiento, 1997; Deutsch et al., 2001). Here, N* quantifies the deviation of the [DIN]:[TCO2] 6 M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 Table 1 Summary of benthic flux measurements in the Santa Monica Bay Station/chamber August 2000 Hyperion 1-B 1-R Malibu 2-B 2-R 7-B 7-R Marina del Rey 8-B March 2001 Hyperion 1-B 1-R 1-Y 5-Y Topanga 3-B 3-R 3-Y Marina del Rey 4-B O2 (mmol m 2 d 1) TCO2 (mmol m 2 d 1) ALK (mmol m 2 d 1) PO34 (mmol m 2 d 1) NO 3 (mmol m 2 d 1) NH+4 (mmol m 2 d 1) Si(OH)4 (mmol m 2 d 1) 6.6 F 1.4 7.0 F 1.4 14.8 F 6.3* 13.3 F 5.8* 11.0 F 5.2 10.5 F 5.7 0.07 F 0.048 0.05 F 0.04 0.10 F 0.21 0.06 F 0.06 0.21 F 0.17 0.31 F 0.19 2.3 F 0.3 2.4 F 0.2 10.3 F 2.1 11.4 F 2.4 18.7 F 3.8 8.9 F 1.9 9.5 F 3.7 12.1 F 3.0 15.3 F 2.7 9.7 F 1.7 4.1 F 3.8 4.7 F 3.7 6.5 F 3.2 2.4 F 2.1 0.04 F 0.05 0.07 F 0.03 0.08 F 0.02 0.13 F 0.02 0.14 F 0.06 0.06 F 0.05 0.11 F 0.05 0.94 F 0.32 0.15 F 0.12 0.26 F 0.12 0.59 F 0.15 0.26 F 0.07 3.5 F 0.4 3.8 F 0.5 4.1 F 0.6 3.8 F 0.4 9.9 F 2.3 10.3 F 2.4 5.2 F 2.5 0.065 F 0.18 0.01 F 0.1 0.08 F 0.08 0.51 F 0.09 6.0 F 1.2 4.9 F 1.1 6.8 F 1.5 7.8 F 2.0 13.5 F 2.4 13.0 F 4.1 11.2 F 2.5 17.9 F 6.8 8.9 F 2.4 8.5 F 2.3 6.3 F 3.3 13.8 F 9.1 0.164 F 0.036 0.155 F 0.036 0.118 F 0.035 0.306 F 0.09 0.81 F 0.08 0.72 F 0.20 0.63 F 0.10 1.09 F 0.22 0.86 F 0.09 1.15 F 0.16 0.76 F 0.13 1.75 F 0.35 2.30 F 0.20 1.97 F 0.32 2.26 F 0.30 3.78 F 0.68 2.3 F 2.3 3.6 F 2.6 3.1 F 2.7 0.164 F 0.035 0.09 F 0.031 0.143 F 0.039 0.36 F 0.07 0.33 F 0.10 0.37 F 0.07 0.52 F 14 0.15 F 0.05 1.10 F 0.27 2.30 F 0.14 1.90 F 0.27 2.48 F 0.30 12.6 F 5.8 0.152 F 0.051 0.4 F 0.13 0.58 F 0.09 2.49 F 0.38 5.3 F 1.0 5.9 F 1.3 6 F 1.3 6.0 F 1.2 2.7 F 3.0* 5.2 F 3.4 4.4 F 6.1 12.4 F 5.7 Negative fluxes are from the overlying water into the sediments. Unless indicated by an asterisk, TCO2 fluxes represent mean values combining measured and calculated (based on TA and pH) TCO2 concentrations. TCO2 fluxes with an asterisk were derived using only calculated TCO2 concentrations. ratio from ‘Redfield’ N:C stoichiometry. The flux of N*TCO2 (FNc*) is defined as: FNc* ¼ FNO3 þNO2 þ FNH4 ð16=106 FTCO2 Þ ð2Þ Negative N*TCO2 fluxes (Table 2) correspond to lower than expected DIN:TCO2 ratios and suggest that the SMB sediments are a net sink of N. FNc* can be directly translated into the total denitrification rate FDenit (assuming that denitrification is the dominant N2-producing process) and (16/106) FTCO2 equals the total remineralization of organic N ( FMin). Accordingly, the nitrification rate ( FNit) equals the difference between FMin and FNH4. TCO2 flux would overestimate organic C oxidation in the case of carbonate dissolution. A relatively high TA flux with respect to TCO2 flux could be taken to indicate carbonate dissolution in the SMB. However, from similar depositional systems, it has been reported that carbonate dissolution plays a relatively unimportant role (less than 5% of the TCO2 flux; McNichol et al., 1988; Hammond et al., 1999; Hopkinson et al., 2001), suggesting that the TA flux in the SMB is primarily driven by net sulfate reduction rather than carbonate dissolution (ammonium production and nitrate consumption also contribute to the total TA flux but only at a low level). In an alternative approach to calculate total denitrification rates, we use phosphate fluxes according to Eq. (3): FNp* ¼ FNO3 þNO2 þ FNH4 ð16 FPO4 Þ ð3Þ If the PO43 flux is a conservative tracer of benthic organic matter decomposition (that is, P displays Redfield-stoichiometric behavior during mineralization and is not produced or consumed by additional processes), FNp* should, again, be a direct measure of the total denitrification rate FDenit. In some cases, the denitrification rate based upon phosphate flux agrees M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 Table 2 N* fluxes Station/chamber August 2000 Hyperion 1-B 1-R Malibu 2-B 2-R 7-B 7-R Marina del Rey 8-B March 2001 Hyperion 1-B 1-R 1-Y 5-Y Topanga 3-B 3-R 3-Y Marina del Rey 4-B FNc* (mmol m 2 d 1) FNp* (mmol m 2 d 1) 2.12 F 1.03 1.76 F 0.94 1.01 F 0.83 1.05 F 0.68 1.15 F 0.61 1.50 F 0.53 1.83 F 0.54 2.14 F 0.46 0.93 F 0.82 0.80 F 0.52 0.80 F 0.41 2.76 F 0.55 1.46 F 0.44 0.95 F 2.89 1.93 F 0.47 1.52 F 0.72 1.52 F 0.47 2.01 F 1.16 2.56 F 0.72 2.04 F 0.75 1.75 F 0.66 4.23 F 1.70 0.24 F 0.73 0.95 F 0.53 0.07 F 0.97 2.47 F 0.72 1.62 F 0.56 1.56 F 0.80 1.64 F 0.91 2.25 F 0.91 N* fluxes represent a direct measure of the total denitrification rate. Discrepancies between TCO2-based ( FNc* ) and PO34 -based ( FNp* ) N* fluxes are in most cases most likely due to the nonconservative behavior of phosphate during early sedimentary diagenesis (PO34 retention in sediments). well with that calculated using TCO2 flux (Table 2), supporting our assumption that marine phytoplankton is the primary organic matter source in the SMB, undergoing diagenesis according to near-Redfield stoichiometry. However, for some locations, and particularly in August 2000, when DIN and phosphate fluxes were generally lower and afflicted with a high relative uncertainty, FNp* values do not correspond well to FNc* values. This imbalance is probably primarily due to the fact that phosphate is not always a conservative indicator of benthic metabolism (that is, P remineralization also involves authigenic and Fesorbed P). Hence, we conclude that using phosphate fluxes is not a viable alternative approach to derive total denitrification rates in the SMB. The consumption of O2 and the efflux of Si(OH)4 can give additional constraints on organic matter diagenesis. Rates of O2 consumption were often 7 significantly lower than the TCO2 flux out of the sediment (Table 1), suggesting that anaerobic organic matter decomposition (denitrification, sulfate reduction) accounts for a large portion of total organic carbon oxidation at these sites, and that reduced end products of anaerobic respiration remained stored in the sediments or were released to the overlying water prior to being reoxidized. Thus, O2-based calculations of N and P fluxes in the SMB would lead to underestimates. The fact that ‘‘for about half the chamber experiments’’ Si(OH)4:TCO2 flux ratios are relatively close to an elemental ratio typical for diatoms (0.13 F 0.5; Brzezinski, 1985) suggests that fresh pelagic diatoms represent the major biogenic component in SMB sediments. In turn, the Si(OH)4 data support the TCO2 flux as generally a good measure of benthic metabolic rates in this environment. Hence, we favor the use of TCO2 flux as a basis for calculating ammonium production and total denitrification rates. Calculated total denitrification rates are mostly between 1.0 and 2.0 mmol N m 2 d 1 (Table 3). The overall mean of 1.74 F 0.33 mmol N m 2 d 1 falls into the range of denitrification rates reported for other coastal marine environments (compilation in Seitzinger, 1988; Hopkinson et al., 2001; Laursen and Seitzinger, 2002 and references therein). There was not a consistent variation in total denitrification rate among the different locations or between seasonal regimes. The CO2 flux that is due to denitrification can be estimated according to the equation: ðCH2 OÞ106 ðNH3 Þ16 ðH3 PO4 Þ þ 84:8 HNO3 ! 106 CO2 þ 42:4 N2 þ 16 NH3 þ H3 PO4 þ 148:4 H2 O ð4Þ (Froelich et al., 1979). In the SMB sediments, 12% to 22% (on average 15%) of the total carbon mineralization can be attributed to microbial denitrification. 3.2. Benthic nitrogen pathways The fraction of FMin that is denitrified in the sediments depends on two branching points: (1) the fraction of remineralized N that is nitrified within the sediments (B versus A in Fig. 1) and (2) the fraction of nitrified N that is denitrified within the sediments (Dn versus C in Fig. 1). The ratio of the ammonium flux 8 M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 Table 3 Total organic N remineralization rates FMin, nitrification rates FNit, denitrification rates FDenit, and ammonium – efflux ratios QNH4 in the SMB Station/chamber August 2000 Hyperion 1-B 1-R Malibu 2-B 2-R 7-B 7-R Marina del Rey 8-B March 2001 Hyperion 1-B 1-R 1-Y 5-Y Topanga 3-B 3-R 3-Y Marina del Rey 4-B FMin (mmol N m 2 d 1) FNit (mmol N m 2 d 1) FDenit (mmol N m 2 d 1) QNH4 (%) 2.23 2.01 2.02 F 1.03 1.70 F 0.94 2.12 F 1.03 1.76 F 0.94 9.4 15.4 95 97 1.43 1.83 2.31 1.46 1.29 F 0.61 1.56 F 0.53 1.83 F 0.54 1.20 F 0.46 1.15 F 0.61 1.50 F 0.53 1.83 F 0.54 2.14 F 0.46 20.1 17.6 25.5 17.8 112 104 100 56 1.55 1.47 F 0.44 1.46 F 0.44 5.8 101 2.04 1.96 1.69 2.70 1.12 F 0.46 0.78 F 0.69 0.89 F 0.46 0.92 F 1.14 1.93 F 0.47 1.52 F 0.72 1.52 F 0.47 2.01 F 1.16 44.9 59.9 47.3 65.9 58 52 59 46 2.62 1.44 2.29 2.10 F 0.72 1.28 F 0.56 1.18 F 0.80 2.47 F 0.72 1.62 F 0.56 1.56 F 0.80 19.9 11.6 48.5 85 79 76 1.87 1.24 F 0.91 1.64 F 0.91 33.5 76 FNit/FDenit (%) FMin is predicted from ‘Redfield’ stoichiometry. The nitrification rate FNit is, in essence, the difference between the predicted FMin and the observed ammonium flux out of the sediments. The denitrification rate FDenit is the difference between FNit and the observed nitrate flux (where nitrate fluxes into the sediments have a negative sign). The ammonium – efflux ratio QNH4 is calculated by dividing the net flux of ammonium out of the sediments by FMin. (1 QNH4) can be translated into an upper limit of the percentage of remineralized N that is eventually denitrified (i.e., ‘‘denitrification efficiency’’, Berelson et al., 1998). The ratio FNit/FDenit defines the maximum portion of total denitrification that may be fueled by nitrification within the sediments. FMin and dependent parameters are mostly calculated using TCO2 fluxes. Only at Station 3 (Topanga) in March 2001, phosphate fluxes are used to predict total organic N remineralization. The validity of thus derived values for FNit, FDenit, and QNH4 at this station is corroborated by P:Si flux ratios close to the expected value of 1:15. out of the sediments (Flux A) over the total amount of remineralized nitrogen ( FMin), from now on referred to as the ammonium – efflux ratio QNH4, characterizes the first branching point: the lower QNH4 is, the higher is the fraction of FMin that is available for denitrification. This branching point can be estimated from the benthic flux data alone ( FNH4/[(16/106)FTCO2]). The ‘‘coupling of nitrification to denitrification’’ describes the second branching point. GCND, which denotes the degree of this coupling, is the ratio of the fraction of newly produced nitrate that is denitrified in the sediment over the total amount of nitrified nitrate (Flux Dn/(Flux C + Flux Dn) in Fig. 1). The closer the coupling between nitrification and denitrification (or the higher GCND), the lower is the portion of newly produced nitrate fluxing out of the sediments. As described next, GCND is not constrained by the benthic flux data alone but does influence isotope data. The net nitrate flux into the sediments determined from the benthic chamber data is the net inward flux of nitrate that is consumed by denitrification (Db in Fig. 1) minus the net outward flux of nitrate driven by nitrification that is not coupled to denitrification (C in Fig. 1). Total denitrification ( FDenit) is, in a sense, composed of two components: the denitrification fed by the inward flux of bottom-water nitrate (Db) and that fed directly by nitrification in the sediments (Dn). In addition, a fraction of the inward flux of bottomwater nitrate driven by Db may be countered by an outward flux of nitrate from nitrification in the sediments. For both of these reasons, we expect and observe that the denitrification rate estimated from M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 benthic flux stoichiometry is significantly higher than the rate of nitrate loss from benthic chamber water (Tables 1 and 2). The two alternative explanations for the difference between FDenit and bottom-water nitrate disappearance (the denitrification term coupled to nitrification and the efflux of newly produced nitrate into the water column) cannot be distinguished based on the benthic flux data alone. That is, for a given total denitrification rate FDenit, the benthic flux data cannot discern a case of higher C, higher Db, and lower Dn from a case of lower C, lower Db, and higher Dn. We will see that the nitrate isotopes can distinguish between these two alternatives. In a similar vein, the isotope data also provide an indication of how much exchange of bottom water-derived nitrate occurs across the sediment –water interface (E in Fig. 1), a process that, like the interfering terms C and Db, cannot be extracted from the benthic flux data alone. Like denitrification rates FDenit, ammonium – efflux ratios QNH4 and nitrification rates (listed as FNit and equivalent to flux B in Fig. 1 are estimated from benthic flux data at each of the chamber sites (Table 3). In essence, FNit is the difference between denitrification FDenit and the net change in chamber water [NO3]. FNit has often been referred to as ‘‘the coupled nitrification –denitrification rate’’, neglecting the uncertainty of gross fluxes of newly produced nitrate out of the sediment (e.g., Berelson et al., 1998; Laursen and Seitzinger, 2002). However, in net denitrifying systems, FNit is equivalent to Dn in Fig. 1 only in the case where the efflux of newly nitrified nitrate (Flux C in Fig. 1) is zero. Because we cannot neglect a possible efflux of newly nitrified nitrate, we restrict the term ‘‘coupled nitrification – denitrification’’ for flux Dn in Fig. 1. For the same reason, the nitrification:denitrification ratio in Table 3 defines the maximum portion of total denitrification that may be fueled by nitrification within the sediments, but does not necessarily indicate the actual degree of coupling between nitrification and denitrification. The highest ammonium –efflux ratio QNH4 (i.e., the highest net outward NH4+ flux relative to FMin) was observed for Station 5-Y (Table 3). Elevated radon fluxes (data not shown) measured in the same chamber suggest the enhanced exchange of solutes across the sediment– water interface due to nondiffusive flux, most likely as a result of bioirrigation. It is very likely that irrigation rates have a 9 strong effect on benthic N cycling (e.g., Blackburn and Henriksen, 1983; Mayer et al., 1995; Berelson et al., 1998). Because they represent an important control on the exchange between NH4+-rich sediment pore waters and NH4+-poor bottom waters, irrigation rates are likely to have a significant impact on QNH4 , and thus on the sediment nitrification-denitrification rate as well. For example, at Station 5-Y, nondiffusive processes enhanced the escape of remineralized N from the sediments as NH4+ prior to its oxidation to nitrate, resulting in a comparatively high value for QNH4 (66%). The bay-wide ammonium – efflux is lower, with generally less than 50% of the remineralized NH4+ escaping from the sediments. Bioirrigation will also influence the gross flux of nitrate into the sediment. Thus, the effect of bioirrigation on total denitrification rates will depend on the relative potential of bottom-water nitrate and pore water ammonium to serve as nitrate sources for denitrification. In the SMB, as observed in other coastal marine systems (Seitzinger, 1988; Laursen and Seitzinger, 2002), the rate of nitrification is >50% of the total denitrification rate. If the sediment efflux of newly nitrified nitrate is not significant, then this would imply that coupled nitrification – denitrification within the sediments is greater than that of ‘‘direct’’ denitrification (where nitrate in sediments is supplied by diffusive flux from overlying water). For example, in March 2001, although near-bottom water nitrate concentrations were relatively high (20 F 2 Amol/l), nitrate from nitrification within the sediment had the potential to contribute more than 50% of the nitrate for denitrification. In August 2000, when near-bottom nitrate concentrations were significantly lower, nitrification may have played an even more important role, potentially accounting for 90% to 100% (mean 91.7%) of the total denitrification. On average, nitrification could account for 78 F 18% of the nitrate required by the total denitrification rate, attesting to the potential importance of this process on the role of continental shelf sediments as effective sinks for fixed nitrogen (Devol, 1991; Christensen, 1994). However, it must again be pointed out that, by looking at the benthic flux stoichiometry alone, it is not possible to address how much newly nitrified nitrate may actually be evading to the bottom water. 10 M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 3.3. Effects of benthic N cycling on nitrate d15N and d18O The N and O isotopic composition of ‘initial’ nitrate in the SMB chambers as determined in Niskin bottom-water samples was more or less constant for all incubation experiments and averaged + 7.8 F 0.45xand + 1.4 F 0.7x , respectively. Only in August 2000 at Station 8 (Marina del Rey), where near-bottom water was strongly nitrate-depleted ([NO3] < 2 Amol/l), d15N and d18O values were significantly higher ( + 9.3xand + 5.0x , respectively). As generally lower nitrate concentrations in August ( f 10 Amol/l compared to f 20 Amol/l in March) are due to phytoplankton assimilation, the same explanation presumably applies to the elevation of nitrate d15N and d18O. Even under high-[NO3] conditions in March 2001, the nitrate d15N and d18O in the SMB are high relative to the global ocean mean d15N of 5x(Sigman et al., 2000) and d18O of f 0x (a global database for marine nitrate d18O does not yet exist but deep North Pacific nitrate has a d18O close to that of seawater ( f 0x); Casciotti et al., 2002; A. Knapp, M.F. Lehmann, and D.M. Sigman, unpublished data). The elevated nitrate d15N in subsurface SMB nitrate during winter fit well into the general N – S trend along the coast of Mexico and California, with a continuously increasing nitrate d15N toward the eastern tropical North Pacific (ETNP; Altabet et al., 1999; Sigman et al., 2003). Therefore, it indicates the impact of a northward undercurrent of denitrificationinfluenced water from the ETNP (Altabet et al., 1999), which has been recognized as an important region of water-column denitrification and, hence, as a source of 15N-enriched nitrate (Cline and Kaplan, 1975; Brandes et al., 1998; Sigman et al., 2000; Voss et al., 2001). Bacterial denitrification is associated with a substantial intrinsic isotope fractionation of 25 F 5x (e.g., Mariotti et al., 1981; Brandes et al., 1998; Barford et al., 1999). Although a significant amount of nitrate was lost from most of the chambers through denitrification within the sediments, the nitrate N isotope ratios at the end of the deployment period were not markedly different from those in the beginning (Fig. 4). This basic observation is consistent with observations made by Brandes and Devol (1997, 2002) during incubation experiments in the Puget Sound and along the Washington and Mexican continental margins. There are two possible explanations for the lack of N isotope enrichment observed in the SMB. Firstly, as proposed by Brandes and Devol (1997), the effective N isotope fractionation is indeed close to zero because denitrification is limited by diffusion through sediments. Secondly, a nonzero isotope effect of sedimentary denitrification on the N isotopic composition of nitrate in overlying water may be effectively balanced by a 15N-depleted nitrate source, i.e., nitrate from the oxidation of ammonium. The latter option represents an alternative that has not been addressed in detail by previous work. In fact, in the chamber experiment 5Y, where the net DIN flux out of the sediment was greatest, we indeed observe a subtle but continuous decrease in d15N from + 7.8xto + 6.8xduring deployment time (Fig. 4). This suggests that some fraction of the 15N-depleted nitrate produced via nitrification of ammonium in oxic sediments is not consumed by denitrifying bacteria and ends up in the benthic chamber, counteracting any N isotope enrichment of chamber nitrate due to denitrification. This observation is consistent with results reported from lander-incubation and whole-core incubation experiments on the Mexican and Washington continental margins (Hartnett, 1998), where decreasing nitrate d15N trends corresponded to net nitrate consumption within the sediments, and nitrification occurring with a significant isotope fractionation was suggested to be the source of 15N-depleted nitrate. The partitioning between ammonium and newly produced nitrate fluxing out of the sediment and ammonium being nitrified and then denitrified within the sediments not only controls the amount of fixed nitrogen escaping from the sediments but also affects its 15N/14N. Thus, the ammonium – efflux ratio QNH4 and the degree of coupling between nitrification and denitrification are important constraints on the total isotope effect of benthic N cycling in the overlying water. Three specific scenarios will elucidate this point. In all of these scenarios, NH4+ with a d15N of + 6xis produced during mineralization of organic matter, given that the d15N of surface-sedimentary organic matter in SMB is f 6x(data not shown) and assuming that eMineralization c 0x(Wada et al., 1980; Kendall, 1998). The intrinsic N isotope effect M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 11 Fig. 4. Nitrate d15N and d18O during benthic chamber incubations in March 2001. Different symbols represent chambers deployed at different stations (1-B, 1-Y, 5-Y: Hyperion; 3-B, 3-R: Topanga; 4-B: Marina del Rey). Error bars represent standard deviations (r1) calculated from duplicate or replicate isotope analyses. For samples that were analyzed only once, the length of the error bars are based upon estimates on instrument precision ( F 0.2xfor d15N and F 0.3xfor d18O, respectively). for nitrification can vary substantially with the molecular nitrification enzymes but is generally large ( 12xto 38x ; Mariotti et al., 1981; Horrigan et al., 1990; Casciotti et al., 2003). Here, we adopt an average enrichment factor of 17x , derived from experiments with two marine nitrifier isolates (Casciotti et al., 2003). Under steady-state conditions in the SMB sediments, if the ammonium from organic matter degradation is c ompletely nitrified ( QNH4 = 0%), then newly produced nitrite +nitrate will, by mass balance, have the same d15N as the source organic matter ( + 6x). At the other extreme, if almost all of the remineralized ammonium fluxes out of the sediments before being oxidized to nitrate ( QNH4 c 100%), its d15N would be + 6x . In this case, the small fraction of the ammonium that is nitrified will yield nitrate that is extremely depleted in 15N ( 11x ). In an intermediate scenario, where 50% of the ammonium is lost by diffusion out of the sediments and the other 50% is oxidized by nitrifying bacteria ( QNH4 = 50%; e.g., March 2001 Stations 1-B, 1-Y, 3-Y; Table 3), at steady state, the d15N of newly produced nitrate will be 2.5xand that of effluxing ammonium will be + 14.5x(similar d15N values for effluxing ammonium from Puget Sound sediments have been reported by Brandes and Devol, 1997). In essence, the lower the fraction of NH4+ being nitrified within the sediments (or the higher the QNH4, which equals Flux B/(Flux A + Flux B) in Fig. 1), the lower is the d15N of product nitrate that potentially contributes to the nitrate d15N signal in the benthic chamber. However, there is one additional consideration: nitrate from nitrification can only be a ‘‘light’’ nitrate source to overlying water if it is not readily utilized by denitrifying bacteria before entering the benthic chamber, that is, if nitrification is not closely coupled to denitrification within the sediments. The greater the fraction of newly produced nitrate that is denitrified within the sediments (i.e., high GCND and close coupling of nitrification and denitrification), the smaller is the potential flux of nitrate into the overlying water. Moreover, this denitrification would also transform the pore water nitrate from 15N-depleted to 15 N-enriched. From the nitrate d15N data alone, it is not possible to infer a priori whether the apparent N isotope effect associated with denitrification in SMB sediments is indeed close to 0xor whether it is significantly higher but effectively balanced by the gross flux of low-15N nitrate from nitrification. The d15N of N2 gas fluxing out of the sediments could provide additional information on the N isotope balance (see Brandes and Devol, 1997). However, large amounts of background N2 in marine waters limit the use of N2 isotope 12 M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 measurements to constrain the system. It can also be shown that the d15N of N2 would not provide a constraint on the net isotope effect of sedimentary denitrification because it does not allow for the actual DIN pathways to be distinguished. Ammonium isotope data should allow for characterization of QNH4 but not the degree of nitrification – denitrification coupling. More generally, N-isotope ratios in nitrogen species do not unequivocally allow for a deconvolution of coexisting nitrate production, consumption, and gross diffusive transport terms. Nitrate d18O provides the constraint needed to disentangle these co-occurring N cycle reactions (Lehmann et al., 2003; Sigman et al., 2003). Nitrate 15 14 N/ N and 18O/16O are similarly affected by microbial denitrification. As a result, if the isotope effect is expressed at all, both the d15N and d18O in residual nitrate increase with the degree of nitrate consumption, and, at least in marine environments, the ratio of 15 N and 18O enrichment seems to be close to unity (Casciotti et al., 2002; Sigman et al., 2003). In contrast to denitrification, nitrate production by nitrification has a very different effect on the d15N and d18O of the nitrate pool. It was traditionally assumed that two-thirds of the oxygen atoms in nitrificationderived nitrate originated from ambient H2O and onethird from dissolved O2 (Kumar et al., 1983; Durka et al., 1994; Bohlke et al., 1997). More recently, Casciotti et al. (2002) argued that less than one sixth of the oxygen atoms in marine nitrate come from dissolved O2, which was supported by their observation that the d18O of deep-sea nitrate is close to 0xversus V-SMOW. Either way, given a d18O of + 1.6 F 0.8x for bottom-water nitrate in the SMB, the d18O of nitrate within the benthic chamber should change very little due to microbial nitrification. This is in contrast to the large 15N-depleting effect of nitrification (see above). Thus, if there were a flux of remineralized (i.e., nitrified) nitrate out of the sediments, the N and O isotopes of chamber nitrate would evolve in different ways. This does indeed occur in chamber 5-Y, which shows the greatest evidence for irrigationdriven efflux of solutes from the sediments: the chamber nitrate in 5-Y decreases in d15N but not in d18O (filled circles in Fig. 4). Thus, the possible contribution of nitrification to nitrate in the chamber can be discerned by the combined measurement of nitrate d15N and d18O. We use a simple one-box model (based on the scheme depicted in Fig. 1) that simulates the change of nitrate d15N and d18O in a benthic chamber as a function of the effective isotope enrichment associated with denitrification, and as a function of the N and O isotopic composition of newly produced nitrate. Based upon recent marine studies (Casciotti et al., 2002; Sigman et al., 2003), we assume that the isotope effect for O is the same as for N. We assume steadystate conditions within the sediments and an N isotope effect for nitrification of 17x(the effective N isotope fractionation may actually be lower because ammonium is not nitrified from a homogenized ammonium pool). d18O of newly produced nitrate is assumed to be 0x, independent of the ammonium – efflux ratio. For each time step, we calculate the amount and the isotopic composition of nitrate lost due to denitrification and gained through nitrification, respectively. The isotopic composition of newly produced nitrate is controlled by the partitioning of DIN fluxes which, in turn, is determined by the degree of coupling between nitrification and denitrification (given as GCND = % of newly produced nitrate being denitrified within the sediments) and, in the case of d15N, by QNH4 (45% in Fig. 5A and 10% in Fig. 5B). Again, GCND is the ratio Flux Dn/(Flux C + Flux Dn), where Flux C represents the fraction of newly produced nitrate that escapes from the sediment to the overlying water and Flux Dn represents that fraction of newly produced nitrate that is denitrified within the sediments before it diffuses into the bottom water. It can be seen that the trend in nitrate d15N that is observed in the model benthic chamber depends on the relative contribution of denitrification and nitrification as well as on the isotope fractionation associated with denitrification. Denitrification occurring with even a small apparent isotope effect will drive the d15N of nitrate in overlying water towards higher values if ammonium oxidation does not play a significant role at all or if it is closely coupled to denitrification in the sediments. On the other hand, if nitrification contributes to a gross flux of nitrate out of the sediments and that nitrate results from incomplete nitrification, the d15N of benthic chamber nitrate can decrease. The DIN pathways control the degree of N isotope change. The simultaneous effects of the apparent N isotope enrichment due to isotope frac- M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 tionation during sedimentary denitrification and the release of 15N-depleted microbial nitrate can potentially cancel each other (circles in the upper panels of Fig. 5A and B), producing no N isotope change. Similarly, no N isotope change is observed when both e = 0xand GCND = 100% (crosses in the upper panels of Fig. 5A and B). Hence, two completely different scenarios can create the same N isotope trends in benthic chamber nitrate, indicating that the d15N values alone do not yield unambiguous information on GCND and e. The nitrate d18O that is observed in the model chamber is also dependent on the isotope effect associated with sedimentary denitrification and the coupling of nitrification and denitrification (Fig. 5A and B, lower panels). However, the opposing effect of microbial nitrate input to the chamber water on the d18O of chamber nitrate is insignificant because, in contrast to d15N, the initial nitrate d18O and that from nitrification do not differ much. Different scenarios (varying values for e and GCND) result in characteristic d15N –d18O combinations; hence, these signatures provide constraints on the apparent isotope effects and the degree of coupled nitrification– denitrification. It is noteworthy that the exact sensitivity of this dual-isotope technique depends on the boundary conditions (initial d15 N NO 3 , d18 O NO 3 , [NO 3], eNitrification , Q NH 4 , denitrification rates, d15Norg). For example, in marine environments with relatively high QNH4, the dissimilarity of coupled N and O signatures for varying values of GCND will be more pronounced because (more 15N-depleted) microbial nitrate added to the chamber nitrate will have a higher potential to drive the N isotopic composition of chamber nitrate away from its initial value. The fact that both nitrate d15N and d18O generally do not change during the benthic chamber experiments in the SMB is consistent only with a setting where the ‘‘effective’’ N and O isotope effects associated with denitrification are not greater than 1 F 1x, and most of the nitrate from microbial nitrification is completely consumed within the sediments; that is, no significant gross efflux of nitrate occurs (Fig. 6). As outlined above, the ammonium – efflux ratio QNH4 represents one control on the amount and isotopic composition of nitrate potentially contributing to the bottom-water nitrate pool. Accordingly, additional evidence in support of our conclusion that the gross 13 flux of nitrate out of the sediments is generally close to zero is provided by the fact that the SMB chamber isotope trends show no sign of correlation with variations in QNH4. As can be seen when comparing Fig. 5A and B, different QNH4 values as observed in the SMB would lead to significant changes in the spectrum of possible effective isotope effects if a significant flux of nitrate out of the sediments occurred, yet this is not observed. Although overall denitrification rates varied quite significantly among benthic chamber incubations (Table 3), the N and O isotopic trends observed during deployment were always similar, showing no significant change. We thus infer that, independent of the overall denitrification rate, diffusion-limitation led to complete nitrate removal within SMB sediments. In the SMB, solute exchange between pore water and bottom water is dominated by nondiffusive transport (Fraser and Berelson, unpublished data), so that it would seem to be an ideal environment to observe some expression of the large ( f 25x ) intrinsic isotope effect of denitrification. This suggests that the lack of expression of the denitrification isotope effect originally noted by Brandes and Devol (1997) is likely to be a widespread phenomenon. Mechanistically, it yields two possible alternative causes for this lack of expression. The impact of microsites, highlighted by Brandes and Devol as the cause for the lack of isotope effect expression, may be important in a broad range of environments. Alternatively, other geometries of nitrate diffusion and uptake may be equally as effective at driving diffusion-limitation, such that the loss of nitrate by denitrification is not associated with any net isotope fractionation. Generally, no apparent isotope fractionation occurs when denitrification completely consumes the nitrate in sediment pore waters. It needs to be stressed that even if sedimentary denitrification is not associated with any significant effective isotope effect (measured in the overlying water), it is certainly associated with an intrinsic/biologic isotope effect. Pore water nitrate isotope data (Sigman et al., 2001; Lehmann et al., unpublished) clearly show that denitrification within sediments is associated with a marked intrinsic isotope effect probably not much different from that observed in open-water environments. Yet, whether this isotope effect is communicated with nitrate in overlying water is likely to depend on the extent of 14 M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 diffusion limitation across the sediment – water interface or into microzones. We argue above that the overall effect of benthic nitrogen cycling on the isotopic composition of nitrate in a benthic chamber is sensitive to the degree of coupling between nitrification and denitrification. As a consequence, one might expect that the redox condition of the sediments, which controls the spatial coexistence of O2, NH4+, and NO3, and thus regulates nitrification and denitrification (Rysgaard et al., 1994), has a strong influence on the nitrate N and O isotope effects perceived in the water column. Based on work by Bender (1990) and Brandes and Devol (1997), one tends to expect a significantly higher apparent N and O isotope fractionation associated with denitrification in less actively denitrifying environments, where concentration gradients are weaker and diffusion-limitation is less important. However, particularly in environments where nitrate production exceeds nitrate consumption (e.g., deep-sea sediments), it is very likely that nitrate that escapes to the overlying water carries the depleted isotope signal of nitrification rather than that of denitrification because nitrification occurs in shallower parts of the sediment column than denitrification (Sigman et al., 2001). As a result, we hypothesize that, in general and probably independent of environmental conditions, marine sediments are not a significant direct source of 15N-enriched nitrate to the overlying water. 3.4. The effect of effluxing NH4+ In those cases where the isotope effect of nitrification produces low-d15N nitrate, it also results in a high-d15N for the ammonium in pore water (Brandes and Devol, 1997). In our simulation depicted in Fig. 5, 15 for example, the d15N of ammonium diffusing out of the sediments would be + 15.4xand + 21.5xfor QNH4 values of 45% and 10%, respectively. Apparently, this ammonium is not oxidized rapidly within the chamber water (as outlined above, we would expect to a see dual isotope signature typical for nitrification if reoxidation rates were high). However, it would eventually be oxidized to nitrate in the ocean water column, so that 15N depletion of nitrate caused by nitrification in the sediments may be completely compensated by the oxidation of the 15N-enriched ammonium in the water column. Accordingly, if denitrification expresses an isotope effect, the nitrate pool in an ocean region or the global ocean will become enriched in 15N. In the same sense, nitrification does not represent a real (long-term) mechanism for lowering the d15N of oceanic nitrate. Quite the contrary, if high-d15N ammonium (from partial nitrification) escapes from the sediments and is eventually oxidized in the open ocean, and nitrification is closely coupled to denitrification (preventing efflux of the complementary 15N-depleted nitrate), the marine nitrate pool will become enriched in 15N. From this perspective, if the N isotopes are to be used as a tool to study N loss from the ocean, it is crucial not only to distinguish between denitrification without an isotope effect and denitrification with an isotope effect that is countered by nitrification in the short term (as we have done here using chamber experiments) but also to constrain actual ammonium fluxes out of the sediments and their isotopic composition. In the case of the SMB, we have demonstrated that (a) the sediment denitrification expresses no isotope effect and (b) no significant gross flux of newly produced nitrate out of the sediments occurs. Hence, the SMB sediments represent a sink for fixed Fig. 5. Simulation of nitrate d15N and d18O trends during benthic chamber incubation. Boundary conditions in (A) QNH4 = 45%, with NO 3 flux = 0.8 mmol m 2 d 1, NH+4 flux = 1.0 mmol m 2 d 1, and PO34 flux = 0.14 mmol m 2 d 1 and in (B) QNH4 = 10%, with NO 3 flux = 0.8 mmol m 2 d 1, NH+4 flux = 0.2 mmol m 2 d 1, and PO34 flux = 0.14 mmol m 2 d 1. In both simulations, the initial nutrient concentrations were: 22.0 Amol/l for nitrate, 0.0 Amol/l for ammonium, and 1.4 Amol/l for phosphate (see small graphs), and the initial nitrate d15N and d18O values were + 7.8xand + 1.2, respectively. Fluxes and initial conditions in (A) are similar to those observed at the Hyperion site in March 2001. In (B), the net ammonium flux was lowered to create a lower QNH4. QNH4 has a direct effect on the d15N of newly produced nitrate, with a lower d15N value at higher QNH4 (see text). In (A), the d15N of newly produced nitrate was 1.6x ; in (B), it was 4.5x(given a sedimentary organic d15N of + 6xand assuming an N isotope effect for nitrification of 17x ). Simulated isotope trends are illustrated for different effective isotope enrichment factors (e) and varying degrees of the coupling of nitrification and denitrification ( GCND, which is inversely proportional to the efflux of microbial nitrate from the sediments). The combined nitrate d15N and d18O trends give unequivocal constraints on the actual pathways of nitrate and the apparent isotope effects associated with sedimentary denitrification (see text for discussion). 16 M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 Fig. 6. Measured versus predicted nitrate d15N and d18O at Station 1-Y (A) and Station 5-Y (B). Real nitrate isotope trends observed during incubation experiments in the SMB are only reconcilable with a setting where the apparent N and O isotope effects are not greater than 1 F 1x and GCND (the fraction of newly produced nitrate that is denitrified) is not lower than 85 F 5% (that is, nitrification is tightly coupled to denitrification and very little of the newly produced nitrate escapes from the sediments). See also Fig. 5. N from the ocean that has no short-term effect on the d15N of the overlying nitrate pool. However, benthic N-cycling reactions can still combine to produce a flux of 15N-enriched ammonium that eventually contributes to the N isotope enrichment of water-column nitrate in the SMB and, in turn, the global ocean. This remains to be addressed. 3.5. Alternate dissimilatory N transformations In coastal marine sediments, dissimilatory nitrate reduction to ammonium (DNRA) has been reported to be a nitrate-reduction pathway as important as denitrification (Rysgaard et al., 1996; An and Gardner, 2002). However, most of these high-DRNA environments (mostly lagoonal) were characterized by hypersulfidic conditions. High sulfide concentrations inhibit nitrification and denitrification but may enhance DRNA (Rysgaard et al., 1996). We do not have sulfide data for the SMB sediments but the benthic stoichiometry indicates that sulfate reduction is not the dominant respiratory mechanism in the SMB, suggesting that DRNA is not prevalent. Even if DNRA were an important nitrogen pathway within the SMB benthic N cycle, the dual-isotope approach to constrain the pathway of newly produced microbial nitrate would still be valid. To our knowledge, reports on the N isotope fractionation associated with DRNA do not exist. However, data by McCready et al. (1983) have shown that ammonium produced during DRNA is strongly depleted in 15N. Hence, DRNA would presumably make the chamber isotopic measurements more sensitive towards the efflux of microbial nitrate because the partial nitrification of DNRA-derived ammonium would lead to an even more distinct (i.e., more negative) nitrate d15N signal. If the isotope M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 fractionation pattern for DRNA and denitrification were similar, the direct effect of DRNA on the pore water nitrate pool would not be different from that of denitrification. Thus, fractionation by DRNA would be constrained by the dual isotopes just as denitrification is. Another aspect of benthic N cycling that requires further investigation are alternate modes of suboxic N2 production, e.g., the anammox process, Mn(II) oxidation by nitrate, or Mn(IV) reduction by ammonium (Luther et al., 1997; Dalsgaard et al., 2003; Kuypers et al., 2003). We have no evidence to suggest that these processes are important in the SMB, but they may be important in other environments and on a global scale (Devol, 2003). At this point, it is difficult to address the potential effect of those particular N-cycling reactions on the benthic nitrate isotope dynamics because we do not know about the possible range of isotope effects associated with these reactions. It is possible that, if they are indeed important elements of the benthic N cycle, the alternate modes of suboxic N2 production may affect the sensitivity of the dual isotope approach, as well as the overall impact of benthic N cycling on the isotopic composition of marine nitrate. We are only beginning to understand all the pathways that can lead to N2 production in marine sediments. However, using stable isotope ratios in DIN species in a similar way as has been done in this study may eventually help to gain insight into the complexity of closely coupled nitrogen conversion processes in benthic environments. 4. Summary and concluding remarks The Santa Monica Bay shelf sediments represent an effective sink for fixed nitrogen, indicating that denitrification in these sediments is an important pathway for organic matter mineralization. Independent of denitrification rates and of the partitioning of ammonium fluxes, sedimentary denitrification did not cause any significant N and O isotope enrichment in residual nitrate in the benthic incubation chambers. The ultimate N and O isotope impacts of benthic nitrogen cycling provide a constraint on the interplay between nitrification and denitrification and on actual nitrogen pathways during microbial organic matter 17 decomposition. Comparing the benthic flux N and O isotope data with a model of the relevant benthic processes, we find that, in the SMB, microbial nitrate is readily denitrified almost to completion within the sediments; thus, coupled nitrification –denitrification sensu stricto accounts for the major portion of total denitrification (generally between 50% and 100%). Moreover, both N and O isotope enrichment associated with denitrification was effectively f 0x, indicating that essentially all bottom-water nitrate entering the sediment was denitrified before having the opportunity to escape back into the overlying chamber water. Given that bio-irrigation is an extremely active process in the SMB sediments (Fraser and Berelson, unpublished data), one might have expected the SMB to be an environment where isotope enrichment in the sediment pore water nitrate would be rapidly communicated into the overlying water. Thus, the lack of such isotope enrichment in the SMB suggests that the diffusion-limitation, the explanation previously posited for negligible or very small N isotope fractionation in other environments (Brandes and Devol, 1997, 2002), can still dominate NO3 isotope behavior in sediments where solute transport across the sediment – water interface and within the sediments is facilitated by nondiffusive processes. Thus, this study extends significantly the range of sedimentary environments in which the nitrate isotope effect of microbial denitrification is essentially not expressed. On the scale of an oceanic region, total DIN fluxes (including ammonium) have to be considered when addressing the N isotopic effects of benthic N cycling on the marine nitrate pool. The low mean d15N of f + 5xfor nitrate in the global ocean interior has been explained as the result of very low degrees of isotope enrichment associated with sedimentary denitrification (Brandes and Devol, 2002). Brandes and Devol (2002) proposed a ratio of global sedimentary to pelagic denitrification of up to 4:1, based upon estimates on effective N isotope fractionation for sedimentary and water-column denitrification (e = 1.5xand 25x , respectively), yielding a rate of global sedimentary denitrification (21 Tmol N/ yr) that is significantly higher than most previously published rates (Codispoti and Christensen, 1985; Devol et al., 1997). In many actively denitrifying marine sediments, the efflux of NH4+ represents an 18 M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 important component of the benthic N cycle (e.g., Seitzinger, 1988 and references therein; Hopkinson et al., 2001; Laursen and Seitzinger, 2002). We argue that this ammonium would ultimately contribute to the marine nitrate pool. If, on average, the released ammonium is indeed enriched in 15N relative to mean oceanic nitrate (as a result of partial nitrification), the overall N isotope effect of benthic N cycling on oceanic nitrate may be greater than previously assumed. In turn, an even higher ratio of sedimentary to pelagic denitrification would be needed to meet the global N isotope balance according to Brandes and Devol (2002). Addressing this issue further will require isotope measurements of ammonium in pore waters and benthic chamber waters. Acknowledgements This study was supported by DFG grant LE 1326/ 1-1 to M.F.L., NSF grants OCE-0136449 and OCE9981479 to D.M.S., and USC Sea Grant to W.M.B. We thank R. Archer, N. Fraser, G. Cane, and R. Ho for their help and technical assistance. We thank J. Brandes and R. Jahnke for valuable comments on an earlier version of the manuscript. We also thank associate editor M. Altabet who helped to improve the manuscript significantly. Associate editor: Dr. Mark Altabet References Altabet, M.A., et al., 1999. The nitrogen isotope biogeochemistry of sinking particles from the margin of the eastern North Pacific. Deep-Sea Res., Part 1, Oceanogr. Res. Pap. 46 (4), 655 – 679. An, S.M., Gardner, W.S., 2002. Dissimilatory nitrate reduction to ammonium (DNRA) as a nitrogen link, versus denitrification as a sink in a shallow estuary (Laguna Madre/Baffin Bay Texas). Mar. Ecol., Prog. Ser. 237, 41 – 50. Barford, C.C., Montoya, J.P., Altabet, M.A., Mitchell, R., 1999. Steady-state nitrogen effects of N2 and N2O production in Paraoccus denitrificans. Appl. Environ. Microbiol. 65 (3), 989 – 994. Bender, M.L., 1990. The d18O of dissolved O2 in seawater: a unique tracer of circulation and respiration in the deep sea. J. Geophys. Res. (Part C) 95 (12), 22243 – 22252. Berelson, W.M., Hammond, D.E., 1986. The calibration of a new free-vehicle benthic flux chamber for use in the deep-sea. DeepSea Res., Part 1, Oceanogr. Res. Pap. 33 (10), 1439 – 1454. Berelson, W.M., Heggie, D., Longmore, A., Kilgore, T., Nicholson, G., Skyring, G., 1998. Benthic nutrient recycling in Port Phillip Bay, Australia. Estuar. Coast. Shelf Sci. 46 (6), 917 – 934. Blackburn, T.H., Henriksen, K., 1983. Nitrogen cycling in different types of sediments from Danish waters. Limnol. Oceanogr. 28 (3), 477 – 493. Böhlke, J.K., Ericksen, G.E., Revesz, K., 1997. Stable isotope evidence for an atmospheric origin of desert nitrate deposits in northern Chile and southern California, USA. Chem. Geol. 136 (1 – 2), 135 – 152. Böhlke, J.K., Mroczkowski, S.J., Coplen, T.B., 2003. Oxygen isotopes in nitrate: new reference materials for 0-18:0-17:0-16 measurements and observations on nitrate-water equilibration. Rapid. Commun. Mass. Sp. 17 (16), 1835 – 1846. Brandes, J.A., Devol, A.H., 1995. Simultaneous nitrate and oxygen respiration in coastal sediments—evidence for discrete diagenesis. J. Mar. Res. 53 (5), 771 – 797. Brandes, J.A., Devol, A.H., 1997. Isotopic fractionation of oxygen and nitrogen in coastal marine sediments. Geochim. Cosmochim. Acta 61 (9), 1793 – 1801. Brandes, J.A., Devol, A.H., 2002. A global marine-fixed nitrogen isotopic budget: implications for Holocene nitrogen cycling. Global Biogeochem. Cy. 16 (4) (Art. No.1120), doi:10.1029/ 2001GB001856. Brandes, J.A., Devol, A.H., Yoshinari, T., Jayakumar, D.A., Naqvi, S.W.A., 1998. Isotopic composition of nitrate in the central Arabian Sea and eastern tropical North Pacific: a tracer for mixing and nitrogen cycles. Limnol. Oceanogr. 43 (7), 1680 – 1689. Brzezinski, M.A., 1985. The Si – C – N ratio of marine diatoms— interspecific variability and the effect of some environmental variables. J. Phycol. 21 (3), 347 – 357. Casciotti, K.L., Sigman, D.M., Hastings, M.G., Bohlke, J.K., Hilkert, A., 2002. Measurement of the oxygen isotopic composition of nitrate in seawater and freshwater using the denitrifier method. Anal. Chem. 74 (19), 4905 – 4912. Casciotti, K.L., Sigman, D.M., Ward, B.B., 2003. Linking diversity and stable isotope fractionation in ammonia-oxidizing bacteria. Geomicrobiol. J. 20 (4), 335 – 353. Christensen, J.P., 1994. Carbon export from continental shelves, denitrification and atmospheric carbon-dioxide. Cont. Shelf Res. 14 (5), 547 – 576. Christensen, J.P., Murray, J.W., Devol, A.H., Codispoti, L.A., 1987. Denitrification in continental shelf sediments has major impact on the oceanic nitrogen budget. Global Biogeochem. Cy. 1 (2), 97 – 116. Cline, J.D., Kaplan, I.R., 1975. Isotopic fractionation of dissolved nitrate during denitrification in the eastern tropical North Pacific Ocean. Mar. Chem. 3 (4), 271 – 299. Codispoti, L.A., Christensen, J.P., 1985. Nitrification, denitrification and nitrous-oxide cycling in the eastern tropical South-Pacific ocean. Mar. Chem. 16 (4), 277 – 300. Codispoti, L.A., Brandes, J.A., Christensen, J.P., Devol, A.H., Naqvi, S.W.A., Paerl, H.W., Yoshinari, T., 2001. The oceanic fixed nitrogen and nitrous oxide budgets: Moving targets as we enter the anthropocene? Sci. Mar. 65, 85 – 105. Suppl. M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 Codispoti, L.A., Yoshinari, T., Devol, A.H., in press. Suboxic respiration in the oceanic water column. In: P.A. delGiorgio, W.P.B. (Eds.), Respiration in Aquatic Ecosystems. Blackwell Scientific. Dalsgaard, T., Canfield, D.E., Petersen, J., Thamdrup, B., AcunaGonzalez, J., 2003. N-2 production by the anammox reaction in the anoxic water column of Golfo Dulce, Costa Rica. Nature 422 (6932), 606 – 608. Deutsch, C., Gruber, N., Key, R.M., Sarmiento, J.L., Ganachaud, A., 2001. Denitrification and N-2 fixation in the Pacific Ocean. Global Biogeochem. Cy. 15 (2), 483 – 506. Devol, A.H., 1991. Direct measurement of nitrogen gas fluxes from continental-shelf sediments. Nature 349 (6307), 319 – 321. Devol, A.H., 2003. Nitrogen cycle—solution to a marine mystery. Nature 422 (6932), 575 – 576. Devol, A.H., Christensen, J.P., 1993. Benthic fluxes and nitrogen cycling in sediments of the continental margin of the eastern North Pacific. J. Mar. Res. 51, 345 – 372. Devol, A.H., Codispoti, L.A., Christensen, J.P., 1997. Summer and winter denitrification rates in western Arctic shelf sediments. Cont. Shelf Res. 17 (9), 1029 – 1050. Durka, W., Schulze, E.D., Gebauer, G., Voerkelius, S., 1994. Effects of forest decline on uptake and leaching of deposited nitrate determined from N-15 and O-18 measurements. Nature 372 (6508), 765 – 767. Froelich, P.N., et al., 1979. Early oxidation of organic-matter in pelagic sediments of the eastern equatorial Atlantic—suboxic diagenesis. Geochim. Cosmochim. Acta 43 (7), 1075 – 1090. Gonfiantini, R., Stichler, W., Rosanski, K., 1995. Standards and Intercomparison Materials Distributed by the IAEA for Stable Isotope Measurements. International Atomic Energy Agency, Vienna. Gruber, N., Sarmiento, J.L., 1997. Global patterns of marine nitrogen fixation and denitrification. Global Biogeochem. Cy. 11 (2), 235 – 266. Hammond, D.E., Giordani, P., Berelson, W.M., Poletti, R., 1999. Diagenesis of carbon and nutrients and benthic exchange in sediments of the northern Adriatic Sea. Mar. Chem. 66 (1 – 2), 53 – 79. Hartnett, H., 1998. Organic carbon input, degradation and preservation in continental margin sediments: an assessment of the role of a strong oxygen deficient zone. PhD thesis, University of Washington. Hopkinson, C.S., Giblin, A.E., Tucker, J., 2001. Benthic metabolism and nutrient regeneration on the continental shelf of eastern Massachusetts, USA. Mar. Ecol., Prog. Ser. 224, 1 – 19. Horrigan, S.G., Montoya, J.P., Nevins, J.L., McCarthy, J.J., 1990. Natural isotopic composition of dissolved inorganic nitrogen in the Chesapeake Bay. Estuar. Coast. Shelf Sci. 30 (4), 393 – 410. Jahnke, R.A., Jahnke, D.B., 2000. Rates of C, N, P and Si recycling and denitrification at the US Mid-Atlantic continental slope depocenter. Deep-Sea Res., Part 1, Oceanogr. Res. Pap. 47 (8), 1405 – 1428. Kendall, C., 1998. Tracing nitrogen sources and cycling in catchments. In: Kendall, C., McDonnell, J.J. (Eds.), Isotope Tracers in Catchment Hydrology. Elsevier, Amsterdam, pp. 519 – 576. Kumar, S., Nicholas, D.J.D., Williams, E.H., 1983. Definitive 15N NMR evidence that water serves as a source of O during nitrite oxidation by Nitrobacter agilis. FEBS Lett. 152, 71 – 74. 19 Kuypers, M.M.M., et al., 2003. Anaerobic ammonium oxidation by anammox bacteria in the Black Sea. Nature 422 (6932), 608 – 611. Laursen, A.E., Seitzinger, S.P., 2002. The role of denitrification in nitrogen removal and carbon mineralization in Mid-Atlantic Bight sediments. Cont. Shelf Res. 22 (9), 1397 – 1416. Lehmann, M.F., Bernasconi, S.M., Reichert, P., Barbieri, A., McKenzie, J.A., 2003. Nitrogen and oxygen isotope fractionation during denitrification in a lacustrine redox-transition zone. Geochim. Cosmochim. Acta 67 (14), 2529 – 2542. Liu, K.K., Kaplan, I.R., 1989. The eastern tropical Pacific as a source of N-15-enriched nitrate in seawater off southern-California. Limnol. Oceanogr. 34 (5), 820 – 830. Luther, G.W., Sundby, B., Lewis, B.L., Brendel, P.J., Silverberg, N., 1997. Interactions of manganese with the nitrogen cycle: Alternative pathways to dinitrogen. Geochim. Cosmochim. Acta. 61 (19), 4043 – 4052. Mariotti, A., et al., 1981. Experimental determination of nitrogen kinetic isotope fractionation: some principles; illustration for the denitrification and nitrification processes. Plant Soil 62, 413 – 430. Mayer, M.S., Schaffner, L., Kemp, W.M., 1995. Nitrification potentials of benthic macrofaunal tubes and burrow walls—effects of sediment NH+4 and animal irrigation behavior. Mar. Ecol., Prog. Ser. 121 (1 – 3), 157 – 169. McCready, R.G.L., Gould, W.D., Barendregt, R.W., 1983. Nitrogen + isotope fractionation during the reduction of NO 3 to NH4 by Desulfovibrio sp. Can. J. Microbiol. 29, 231 – 234. McNichol, A.P., Lee, C., Druffel, E.R.M., 1988. Carbon cycling in coastal sediments: 1. A quantitative estimate of the remineralization of organic-carbon in the sediments of Buzzards Bay, MA. Geochim. Cosmochim. Acta 52 (6), 1531 – 1543. Middelburg, J.J., Soetaert, K., Herman, P.M.J., Heip, C.H.R., 1996. Denitrification in marine sediments: a model study. Global Biogeochem. Cy. 10 (4), 661 – 673. Redfield, A.C., 1958. The biological control of chemical factors in the environment. Am. Sci. 46 (3), 205 – 221. Revesz, K., Böhlke, J.K., Yoshinari, T., 1997. Determination of d18O and d15N in nitrate. Anal. Chem. 69, 4375 – 4380. Rysgaard, S., Risgaard-Petersen, N., Sloth, N.P., Jensen, K., Nielsen, L.P., 1994. Oxygen regulation of nitrification and denitrification in sediments. Limnol. Oceanogr. 39 (7), 1643 – 1652. Rysgaard, S., Risgaard-Petersen, N., Sloth, N.P., 1996. Nitrification, denitrification, and nitrate ammonification in sediments of two coastal lagoons in southern France. Hydrobiologia 329 (1 – 3), 133 – 141. Sebilo, M., Billen, G., Grably, M., Mariotti, A., 2003. Isotopic composition of nitrate – nitrogen as a marker of riparian and benthic denitrification at the scale of the whole Seine River system. Biogeochemistry 63 (1), 35 – 51. Seitzinger, S.P., 1988. Denitrification in freshwater and coastal marine ecosystems: ecological and geochemical significance. Limnol. Oceanogr. 33 (4), 702 – 724. Sigman, D.M., Altabet, M.A., McCorkle, D.C., Francois, R., Fischer, G., 2000. The delta N-15 of nitrate in the Southern Ocean: nitrogen cycling and circulation in the ocean interior. J. Geophys. Res.-Oceans 105 (C8), 19599 – 19614. 20 M.F. Lehmann et al. / Marine Chemistry 88 (2004) 1–20 Sigman, D.M., et al., 2001. A bacterial method for the nitrogen isotopic analysis of nitrate in seawater and freshwater. Anal. Chem. 73 (17), 4145 – 4153. Sigman, D.M., et al., 2003. Distinguishing between water column and sedimentary denitrification in the Santa Barbara basin using the nitrogen isotopes of nitrate. Geochem. Geophys. Geosyst. 4 (5), 1040 (doi:10.1029/2002GC000384). Silva, S.R., et al., 2000. A new method for collection of nitrate from fresh water and the analysis of nitrogen and oxygen isotope ratios. J. Hydrol. 228, 22 – 36. Voss, M., Dippner, J.W., Montoya, J.P., 2001. Nitrogen isotope patterns in the oxygen-defficient waters of the eastern tropical North Pacific Ocean. Deep-Sea Res., Part 1, Oceanogr. Res. Pap. 48, 1905 – 1921. Wada, E., Goldberg, E.D., Horibe, Y., Saruhashi, K., 1980. Nitrogen isotope fractionation and its significance in biogeochemical processes occurring in marine environments. In: Goldberg, E.D., Horibe, Y., Saruhashi, K. (Eds.), Isotope Marine Chemistry. Uchida Rokakuho Publ., Tokyo, pp. 375 – 398.
© Copyright 2026 Paperzz