nitrate deficits by nitrification and denitrification …deep-sea research i 53 (2006) 94–110...

ARTICLE IN PRESS

0967-0637/$ - se

doi:10.1016/j.ds

�CorrespondiE-mail addre

Deep-Sea Research I 53 (2006) 94–110

www.elsevier.com/locate/dsr

Nitrate deficits by nitrification and denitrificationprocesses in the Indian Ocean

Yuan-Hui Lia,�, Laurie Menviela, Tsung-Hung Pengb

aDepartment of Oceanography, University of Hawaii, Honolulu, Hawaii 96822bNOAA/AOML, Ocean Chemistry Division, Miami, Florida 33149, USA

Received 3 March 2005; received in revised form 12 September 2005; accepted 19 September 2005

Available online 1 December 2005

Abstract

The three-end-member mixing model of Li and Peng [Latitudinal change of remineralization ratios in the oceans and its

implication for nutrient cycles. Global Biogeochemical Cycles 16, 1130–1145] was applied to the World Ocean Circulation

Experiment (WOCE) data from Indian Ocean to obtain additional estimates on the remineralization ratios ðP\N\Corg\-O2Þ

of organic matter in the oxygenated regions. The results show systematic changes of the remineralization ratios with

latitude and depth in the Indian Ocean. The average remineralization ratios for Indian warm water masses (potential

temperature y4�10 �C) are P\N\Corg\-O2 ¼ 1\ð15:6� 0:7Þ\ð110� 9Þ\ð159� 8Þ. These are comparable to the traditional

Redfield ratios ðP\N\Corg\-O2 ¼ 1\16\106\138Þ, and are in good agreement with Anderson’s [On the hydrogen and oxygen

content of marine phytoplankton. Deep-Sea Research I 42, 1675–1680.] values of P\N\Corg\-O2 ¼ 1\16\106\150 within the

given uncertainties. Separation of nitrate deficits resulting from aerobic partial nitrification (dN) and anaerobic

denitrification (dN00) processes using empirical equations is shown to be useful and consistent with other observations. The

dN maximum coincides with the phosphate and nitrate maximums, lies within the oxycline below the oxygen minimum

zone, and is in contact with the continental slope sediments. The dN00 maximum lies within the oxygen minimum zone with

O2o�2mmol/kg, is in contact with shelf or upper slope sediments, and is always associated with a secondary nitrite

maximum in the water column. The spatial extent of dN is much larger than that of dN00. The low N/P remineralization

ratio (o15) for deep waters (yo�10 1C) and the dN maximum in the lower oxycline can be best explained by the partial

conversion of organic nitrogen into N2, N2O, and NO by yet unidentified bacteria during oxidation of organic matter.

These bacteria may have evolved in a low oxygen and high nitrate environment to utilize both oxygen and nitrate as

terminal electron acceptors during oxidation of organic matter (i.e. the partial nitrification hypothesis). Direct proof is

urgently needed.

r 2005 Elsevier Ltd. All rights reserved.

Keywords: Remineralization ratios; Nitrification; Partial nitrification; Denitrification; Anammox; Nitrate deficit; Indian Ocean

e front matter r 2005 Elsevier Ltd. All rights reserved

r.2005.09.009

ng author. Fax: +1808 9567112.

ss: [email protected] (Y.-H. Li).

1. Introduction

Nitrate losses by the denitrification process in thenorthern Arabian Sea and in the eastern tropicalPacific Ocean at intermediate depths (100–800m)with zero oxygen (anoxic) to near zero oxygen

.

www.elsevier.com/locate/dsr

ARTICLE IN PRESSY.-H. Li et al. / Deep-Sea Research I 53 (2006) 94–110 95

(suboxic) concentrations are well documented(Naqvi et al., 1982, 1990; Naqvi and Sen Gupta,1985; Codispoti and Christensen, 1985; Ward andZafiriou, 1988; Ward et al., 1989; Mantoura et al.,1993). The denitrification process is suppressedwhen the concentration of oxygen in the easterntropical North Pacific water is higher than about2 mmol/kg (Devol, 1978). Therefore, we set theupper limit of the term ‘‘suboxic’’ at oxygenconcentration about 2 mmol/kg in this paper. Inthe oxygen minimum zone with O2o�2 mmol/kg,some nitrate is actively reduced to N2 by denitrify-ing bacteria (Codispoti and Christensen, 1985;Ward et al., 1989) through reactions such as

4NO3� þ 5Corg þ 3H2O

! 2N2 þ 5HCO3� þHþ; (1a)

3NO3� þ 5ðNH3Þorg þ 3Hþ ! 8N2 þ 9H2O; (1b)

where Corg and (NH3)org are carbon and nitrogen inorganic matter. For clarity, the term ‘‘denitrifica-tion’’ is strictly applied only to the reduction/oxidation processes occurring in an oxygen mini-mum zone with O2o�2 mmol/kg, and we refer tothis oxygen minimum zone as the denitrificationzone throughout this paper. Below and above thedenitrification zone are oxygen gradient layers(oxyclines), where the abundance and activity ofnitrifying bacteria (including ammonium oxidizersand nitrite oxidizers) are highest (Ward et al., 1989).Furthermore, nitrifying bacteria were also shown tobe present throughout the denitrification zone(Ward et al., 1989). In oxyclines, most of the organicnitrogen is oxidized into nitrite and nitrate (i.e.nitrification processes) through net reactions such as

ðNH3Þorg þ 1:5O2 ! NO2� þHþ þH2O; (1c)

NO2� þ 0:5O2 ! NO3

�: (1d)

However, some is oxidized only up to N2O and NO(Codispoti and Christensen, 1985; Ward and Zafir-iou, 1988; Ward et al., 1989; Naqvi and Noronha,1991) through net reactions such as

2ðNH3Þorg þ 2O2 ! N2Oþ 3H2O and (1e)

N2Oþ 0:5O2 ! 2NO: (1f)

Thermodynamically, the production of N2 by yetunidentified bacteria in the same oxyclines by a netreaction such as

2ðNH3Þorg þ 1:5O2 ! N2 þ 3H2O (1g)

is a real possibility, though a direct proof is stilllacking. Schmidt et al. (2004) did show that a wildtype of Nitrosomonas europaea bacteria in chemo-stat cell cultures can produce nitrogen gases(N2, NO, and N2O) during aerobic oxidation ofammonia, using genes encoding reduction enzymessuch as nitrite reductase, nitric oxide reductase etc.For example,

NH4+ (ammonia monooxygenase) - NH2OH

(hydroxylamine oxidoreductase) - NO2� (nitrite

reductase) - NO (nitric oxide reductase) - N2O(not yet identified nitrous oxide reductase) - N2.

The same species also can produce N2 withaddition of NO2 as oxidizing agent in the reactorduring anoxic oxidation of ammonia (Schmidtet al., 2004).

The N2O and NO produced in the oxyclines maydiffuse into the denitrification zone and be reducedto N2 by the following reactions

2N2Oþ Corg þH2O

! 2N2 þHCO3� þHþ and (1h)

2NOþ Corg þH2O ! N2 þHCO3� þ Hþ; (1i)

forming nitrification–denitrification coupled path-ways to produce N2 (Codispoti and Christensen,1985).

Another important reaction to produce N2 is theanaerobic ammonia oxidation (anammox) reaction

NH4þ þNO2

� ! N2 þ 2H2O; (1j)

mediated by the anammox bacteria belonging to theorder Planctomycetales (Strous et al., 1999; Thamdr-up and Dalsgaard, 2002; Dalsgaard and Thamdrup,2002; Kuypers et al., 2003, 2005). Hydroxylamine(NH2OH) and hydrazine (N2H4) have been shownto be involved in the anammox reaction asintermediate species (Van de Graaf et al., 1997).

Nielsen et al. (2005) demonstrated that aerobicand anaerobic ammonia-oxidizing bacteria actclosely to remove ammonia in an oxygen-limited(�5 mMO2) CANON (acronym of CompletelyAutotrophic Nitrogen removal Over Nitrite) reac-tor. Aerobic ammonia-oxidizing bacteria are activein the thin surface layer of organic particles in theCANON reactor and oxidize ammonia into nitrite.Nitrite is then diffused into the anoxic interior of theparticles where anammox bacteria convert nitriteand ammonia into N2. Marine snow and fecal pelletparticles suspended in a low oxygen water column(such as oxyclines) may prove to be the ideal sitesfor the coupling of aerobic ammonia oxidation and

ARTICLE IN PRESSY.-H. Li et al. / Deep-Sea Research I 53 (2006) 94–11096

anammox reactions. However, anaerobic conditionswere never observed within the marine snow andfecal pellets suspended in high oxygen watercolumn, using O2 microelectrode (Alldredge andCohen, 1987). Only one large fecal pellet (6.2mmlong) showed the complete depletion of O2. Thenitrogen isotopic data of N2O from the westernNorth Pacific (Yamagishi et al., 2005) may suggestpartial oxidation of organic nitrogen to nitrite first,then, reduction of nitrite to N2O (as well as to NOand N2; Schmidt et al., 2004).

Gruber and Sarmiento (1997) introduced a newparameter N* that is defined as

N� ¼ 0:87½N � N̄�, (2a)

where N̄ ¼ 16ðP� 0:181Þ, representing a globalmean relationship between nitrate+nitrite (N) andorthophosphate (P) concentrations based on allGEOSECS data. Later, Deutsch et al. (2001)redefined N* to

N� ¼ N � N̄. (2b)

Therefore, N* is a measure of the deviation of theobserved N from N̄ for a given P. They interpreted anegative N* as mostly caused by the loss of nitratethrough the denitrification process, and a positiveN* by the gain of nitrate through N2 fixation byN2-fixing bacteria. By this interpretation, the loss ofnitrate by the denitrification process could bewidespread even in the oxic regions of the oceans(Gruber and Sarmiento, 1997).

Codispoti et al. (2001) estimated the excess N2 inthe water column of the Arabian Sea, using the Ar/N2 ratio in the water column and in the air. Theyfound that the excess N2 is substantially greaterthan the N2 produced by the denitrification process.The implication is that there must be some otherN2-producing mechanisms in the water column.

Li and Peng (2002) introduced a new three-end-member mixing model to estimate the remineraliza-tion ratios of organic matter in the oxic watercolumn. They showed that the remineralizationratios ðP\N\Corg\-O2Þ of organic matter in the oxicdeep ocean (yo10–12 1C) change systematicallywithin one ocean and from one ocean to another.In particular, the remineralization ratios for thedeep equatorial Indian Ocean ½P\N\Corg\-O2 ¼

1\ð10� 1Þ\ð94� 5Þ\ð130� 7Þ� are quite differentfrom those for the Southern Indian Ocean½P\N\Corg\-O2 ¼ 1\ð15� 1Þ\ð83� 2Þ\ð134� 9Þ�. Theapparent low N/P remineralization ratio for theoxygenated deep equatorial Indian Ocean may

indicate that most of the organic nitrogen wasoxidized into nitrate but some into NO, N2O andN2 by yet unidentified bacteria during oxidation oforganic matter (Li and Peng, 2002). This interpreta-tion is called the partial nitrification hypothesisthroughout this paper. As will be shown later, thepartial nitrification may occur in the oxyclinesaround the denitrification zone, the steep oxygengradients within marine-snow particles and fecalpellets suspended in a low oxygen (but still oxic)water column, and within the bottom sediments ifthe pore water or overlying water is low in oxygenand enriched in nitrate.

Naqvi and Noronha (1991) confirmed the pro-duction of N2O in the oxyclines around thesecondary nitrite maximum zone (corresponding tothe denitrification zone) in the Arabian Sea.Furthermore, the observed strong correlation be-tween apparent oxygen utilization (AOU) andDN2O (the N2O level above atmospheric equili-brium) in the oxyclines supports the dominant roleof the partial nitrification process for producingN2O (probably NO and N2 as well). Nevison et al.(2003) also reached the same conclusion, based onthe global distribution pattern of N2O and thepositive correlation between AOU and DN2O.

The purposes of the present study are threefold:First, to reconfirm the variability of remineraliza-tion ratios in the Indian Ocean by applying thethree-end-member mixing model (Li and Peng,2002) to the World Ocean Circulation Experiment(WOCE) data along the E–W sections shown inFig. 1. Second, to introduce an empirical model todistinguish the nitrate deficit caused either byaerobic partial nitrification (dN) or by anaerobicdenitrification (dN00) processes and elucidate theirrelative importance. Finally, to discuss the spatialdistribution patterns of dN and dN00 in the IndianOcean and their relationship with those for O2 andnitrite (NO2

�) concentrations. Similar results fromthe Pacific Ocean will be presented in a separatepaper (Y.H. Li, L. Menviel, and T.H. Peng, Nitratedeficits by nitrification and denitrification processesin the Pacific Ocean, in preparation).

2. Water mass end members and remineralization

ratios

Many useful section plots of potential tempera-ture (y), salinity (S), oxygen (O2), nutrients (phos-phate, nitrate, and silica), dissolved inorganiccarbon (DIC), alkalinity (Alk), and chlorinated

ARTICLE IN PRESS

Fig. 1. WOCE cruise tracks in the Indian Oceans.

Y.-H. Li et al. / Deep-Sea Research I 53 (2006) 94–110 97

fluoro-carbon (CFC-11) along the WOCE cruisetracks can be found in the eWOCE Gallery as establi-shed by Schlitzer (2000; http://www.awi-bremerhaven.de/GEO/eWOCE/index.html). Additional sectionplots for the Indian Ocean based on GEOSECSdata are given by Spencer et al. (1982). Those plotsare helpful in identifying water masses and theirspatial extent, and the remineralization sites oforganic matter, etc. The emphasis here is to firstidentify the water mass end members along the E–Wsections (WOCE cruise tracks i01 to i05 and s4i inFig. 1) for the three-end member mixing model of Liand Peng (2002).

From Figs. 2 to 4, one can easily identify thefollowing end members of water masses. I ¼Antarctic Bottom Water, II ¼ Circumpolar DeepWater, III ¼ Antarctic Intermediate Water,IV ¼ Sub-Antarctic Oxygen Maximum Water,

V ¼ South Indian Central Water, III* ¼ EquatorialIndian Phosphate Maximum Water, IV*e andIV*w ¼ eastern and western Equatorial IndianCentral Water, V*

¼ Red Sea Water, and varioussurface waters. The asterisks indicate the endmembers formed north of the equator. Data pointsbetween two dotted horizontal lines (i.e. the intervalbetween two given potential temperatures, y) inFigs. 2a–d can be produced by mixing the threeappropriate end members shown in those figures.Therefore those data within a given y interval can befitted by the three-end-member mixing model of Liand Peng (2002) to obtain remineralization ratios.This model essentially involves a multiple-variablelinear regression of hydrographic data, including y,salinity, O2, and one of nutrients (nitrate, phos-phate, and DA [ ¼ DIC–Alk/2]). The advantage ofthis model is that it requires no knowledge of end

http://www.awi-bremerhaven.de/GEO/eWOCE/index.html

http://www.awi-bremerhaven.de/GEO/eWOCE/index.html

ARTICLE IN PRESS

0

5

10

15

20

25

30

0

5

10

15

20

25

30

0

5

10

15

20

25

30

θ

-1

0

1

2

θ θ

34.0 35.0 36.0 37.0S

34.0 34.5 35.0 35.5

34.5 35.0 35.5 36.0S

34.0 34.5

II

III

II

V

I

IV*w

IV

III

III

IV

6N-11Ni01Ei01W

12S-4Si02Ei02W

22S-20Si03Ei03W

66S-60SS4iES4iW

III

V*

IV*e

(a) (c)

(b) (d)

Fig. 2. Salinity (S) versus potential temperature (y) for data from (a) cruise s4i, (b) cruise i03, (c) cruise i02, and (d) cruise i01. Roman

numerals represent various water mass end members: I ¼ Antarctic Bottom Water, II ¼ Circumpolar Deep Water, III ¼ Antarctic

Intermediate Water, IV ¼ Sub-Antarctic Oxygen MaximumWater, V ¼ South Indian Central Water, III* ¼ Equatorial Indian Phosphate

Maximum Water, IV*e and IV*w ¼ eastern and western Equatorial Indian Central Water, respectively, V*¼ Red Sea Water. Sections E

and W for each cruise are roughly separated along the Mid-Indian Ridge.

Y.-H. Li et al. / Deep-Sea Research I 53 (2006) 94–11098

member values for those variables. Model resultsare summarized in Table 1. In Fig. 2d for cruise i01,the data points for y between 2 and 10 1C can beproduced by mixing end members II, IV*e, andIV*w, but there is an additional end member III* inthe same y interval as shown in Fig. 4d. Thus, thosedata cannot be fitted to our three-end-membermixing model. Furthermore, the data points abovey ¼ 11 �C for the cruise i01w (Fig. 2d) need four-end-member mixing and show some effect of thedenitrification (to be discussed later), so those arealso not suitable for the model.

The newly calculated remineralization ratios fordeep water along E–W sections are in goodagreement with the earlier estimates along N–Ssections (Table 1; Li and Peng, 2002); they confirmthat the N/P ratio decreases from 1571 in theSouthern Oceans to 1071 in the deep equatorial

Indian Ocean. As shown in Figs. 3a–c, those deepwaters are all high in dissolved O2 concentrations(450 mmol/kg); thus, the decrease of the N/P ratiois difficult to explain by the loss of nitrate into N2

through the denitrification process. In short, thevariation of dissolved O2, P, N and DA ( ¼ DIC—Alk/2; Li and Peng, 2002) in any properly chosenoxic region can be best explained by the water massmixing and the in situ partial nitrification process.

The interesting results are the remineralizationratios obtained for the shallower warm watermasses ðy4�10 �CÞ above the Sub-Antarctic Oxy-gen Maximum Water (IV), including South IndianCentral Water (V), eastern Equatorial IndianCentral Water (IV*e), and various surface waters.The average remineralization ratios for theIndian warm water masses are P\N\Corg\-O2 ¼

1\ð15:6� 0:7Þ\ð110� 9Þ\ð159� 8Þ (from bold values

ARTICLE IN PRESS

0

5

10

15

20

25

30

0

5

10

15

20

25

30

0

5

10

15

20

25

30

θ

-1

0

1

2

θ

0 50 100 150 200 250

O2 (umol/kg)

0 50 100 150 200 250

0 50 100 150 200 250

O2 (umol/kg)

150 200 250 300 350

6N-11Ni01Ei01W

12S-4Si02Ei02W

22S-20Si03Ei03W

66S-60SS4iES4iW

I

II

III

III

IV

V

III

III

IV

III

IV*

V*

(a)

(b)

(c)

(d)

Fig. 3. Oxygen concentration (O2; mmol/kg) versus y for the same transects as in Fig. 2.


in Table 1), which are distinctly different from thosefor the deep Indian Ocean (Table 1). Using differentapproaches, Shaffer (1996) and Shaffer et al. (1999)also reached the same conclusion that remineraliza-tion ratios change with depth. The remineralizationratios for the warm waters are comparable to thetraditional Redfield ratios ðP\N\Corg\-O2 ¼ 1\16\106\138Þ, and in good agreement with that sug-gested by Anderson (1995; P\N\Corg\-O2 ¼ 1\16\106\150) within the given uncertainties. The low-O2/P ratio of 138 for the traditional Redfield ratiosresults from Redfield’s (1958) unrealistic assump-tion that all organic carbon in plankton existspurely as carbohydrates, represented in the idealizedchemical formula (CH2O)106(NH3)16(H3PO4) formarine plankton. It is time to discontinue the usageof this formula for marine plankton but notRedfield’s concept.

One can rewrite Anderson’s formula for marinephytoplankton (C106H175O42N16P) into C106H48

(H2O)38(NH3)16(H3PO4) according to the proceduregiven by Li and Peng (2002). Its oxidation by

oxygen (nitrification) through nitrifying bacteriacan be written as

150O2 þ C106H48ðH2OÞ38ðNH3Þ16ðH3PO4Þ

! H2PO4� þ 16NO3

� þ 106HCO3�

þ 123Hþ þ 62H2O (3a)

and oxidation by nitrate (denitrification) throughdenitrifying bacteria under anoxic or suboxiccondition as

104NO3� þ C106H48ðH2OÞ38ðNH3Þ16ðH3PO4Þ

! H2PO4� þ 60N2 þ 106HCO3

�

þ 3Hþ þ 32H2O: (3b)

In short, nitrifying bacteria use oxygen anddenitrifying bacteria use nitrate as terminal electronacceptors during oxidation of organic matter. Incontrast, some yet unidentified bacteria (probablysimilar to Nitrosomonas europaea; Schmidt et al.,2004) may have evolved in oxyclines, where theconcentration of oxygen is relatively low and nitratehigh, and are able to utilize both oxygen and nitrate

ARTICLE IN PRESS

0

5

10

15

20

25

30

0

5

10

15

20

25

30

0

5

10

15

20

25

30

θ

-1

0

1

2

θ

0 1 2 3

P (umol/kg)

0 1 2 3

0 1 2 3

P (umol/kg)

0 1 2 3

6N-11Ni01Ei01W

12S-4Si02Ei02W

22S-20Si03Ei03W

66S-60SS4iES4iW

I

II

I+II

III

IV

V

I+II

III

IVeIVw

I+II

V*

IV*

III*

(a)

(b)

(c)

(d)

Fig. 4. Orthophosphate concentration (P; mmol/kg) versus y for the same transects as in Fig. 2, with end member III* representing the

Equatorial Indian Phosphate maximum water.


as terminal electron acceptors. The coupling ofaerobic and anaerobic ammonia oxidizing bacteria(Nielsen et al., 2005) in marine snow and fecalpellets suspended in low oxygen water column (suchas in oxyclines) may provide additional mechanismsfor the partial nitrification.

As schematically shown in Fig. 5, the cycling ofnitrogen in the oceans also involves other nitrogenspecies with different oxidation states duringnitrification, denitrification, and anammox pro-cesses; all are facilitated by different bacteria specieswith genes encoding specific enzymes (Amend andShock, 2001). One important difference betweenreactions (3a) and (3b) is that nitrification producesmany hydrogen ions but denitrification does not.Also reaction (3a) reduces Alk by 16 units andreaction (3b) adds Alk by 104 units for everyincrement of one P unit in the water column.

In summary, our present and earlier studies (Liand Peng, 2002) have amply demonstrated thevariability of remineralization ratios from region

to region and from the surface to deep waters. Oneshould be cautious about the earlier assumptionthat remineralization ratios are constant throughoutthe whole ocean (Broecker et al., 1985; Andersonand Sarmiento, 1994).

3. Nitrate deficits by partial nitrification (dN) and

denitrification (dN00)

As shown in Figs. 6a–d (P versus N), the datapoints from the shallow water of cruise s4i and thewarm water masses (y4�10 1C) of cruises i03, i02,and i01 all fall on straight lines with a slope of about16.570.5. Therefore, both P and N are controlledby the complete nitrification of organic nitrogenwithout net production of NO, N2O and N2 duringoxidation of organic matter in the oxygenatedshallow water column. One may write the relation-ship between P and N as

Na ¼ mðP� PoÞ; (4a)

ARTICLE IN PRESS

Table 1

Remineralization ratios based on the WOCE data from the Indian Ocean

Transect Latitude Long. (E) Stations y (1C) rp rn rp/rn rc rp/rc

Redfield 138 8.6 16 1.30 106

E–W (present study)

i01E 10N-6N 80-98 962-1014 11-29 15674 10.670.2 14.770.5 1.3570.04 11574

i02E 8S-9S 71-106 1077-1154 2-10 11572 10.370.1 11.270.2 1.4370.04 8072

i02W 4S-12S 40-71 1155-1244 2-10 11971 9.670.1 12.470.2 1.5870.03 7571

i02EW 4S-12S 40-106 1077-1244 10-24 14672 9.070.1 16.270.3 1.4970.02 9972

i03E 20S-22S 66-114 444-522 2-10 13171 10.770.1 12.270.1 1.5970.03 8271

i03W 20S-22S 53-66 523-573 2-10 12672 10.070.1 12.670.2 1.5570.03 8172

i03EW 20S-22S 53-114 444-573 10-24 16877 10.170.6 16.671.2 1.5470.05 10978

i04W 25S 35-44 585-610 2-12 11271 9.170.1 12.370.2 1.4970.02 7571

12-20 15872 10.470.1 15.270.2 1.5470.03 10372

i05W 31S-34S 30-50 611-669 2-10 11771 9.470.1 12.470.2 1.4770.02 8071

10-20 16973 11.070.2 15.470.4 1.3570.06 12573

i05E 30S-33S 88-115 395-442 2-8 12772 10.470.1 12.270.2 1.5970.10 8071

S4iE 60S-62S 84-120 65-108 0.2-2.1 11877 8.370.3 14.271.0 1.3870.13 8676

S4iW 61S-66S 34-83 20-64 0.2-2.1 12275 8.670.3 14.270.8 1.3970.12 8875

N–S (Li and Peng, 2002)

i09N 19.6N-2.7N 213-277 0.8-11 13171 13.070.2 10.170.2 1.3770.04 9572

i09N 2N-18.5S 172-212 0.8-8 14271 13.470.2 10.670.2 1.3970.04 10272

i09N 19S-31S 148-171 2-10 13471 10.570.1 12.870.2 1.5770.03 8671

i08S 30.3S-51S 4-49 2-11 13373 9.570.3 14.070.5 1.6670.08 8073

i08S 51.4S-63.3S 50-84 0.2-2 13073 8.570.2 15.370.5 1.5270.07 8573

Note: rp ¼ -O2=P; rn ¼ -O2=N; rp=rn ¼ N=P; rc ¼ -O2=Corg; rp=rc ¼ Corg=P.

Bold numbers are for warm water masses with y410–12 1C.

+5 ← NO3- NO3

- ←Anammox

+3 ← ↔ NO2- ↔ NO2

- ↔

+2 ↔ NO NO ↔ ↔

+1 ↔ N2O N2O ↔ ↔

0 → → N2 N2 N2 ← ↔fixation

-1 ← ↔ NH2OH NH2OH ↔

-3 ← ↔ NH4+ ↔ PON ↔ DON → NH4

+ →Uptake &

ammonificationOxidation state

[Anoxic]Dissimilative

reduction(Denitrification)

[Oxic]Dissimilative

oxidation(Nitrification)

Gasexchange

?

Assimilativereduction

Air Air

Fig. 5. Schematic representation of the nitrogen cycle in the ocean. Dissimilative processes do not involve cellular incorporation, and are

mediated by bacteria at cell–water interface. Anammox is the anaerobic ammonia oxidation.


where m ¼ slope, Po ¼ intercept of P at N ¼ 0, andNa represents the expected concentration of N fromthe right hand side of Eq. (4a) for a completenitrification. The m and Po can be obtained by thelinear least-square fitting of the shallow water data

with NX�1 mmol/kg as well as Pp�1.6 mmol/kg ina chosen region. In case of PoPo for some shallowwater (Figs. 6b–d), Na becomes negative accordingto Eq. (4a). A negative Na only means zeroconcentrations of nitrate and nitrite, and a potential

ARTICLE IN PRESS

0

10

20

30

40

0

10

20

30

40

0

10

20

30

40

10

20

30

40

N (

umol

/kg)

N (

umol

/kg)

0 1 2 3

P (umol/kg)

0 1 2 3

0 1 2 3

P (umol/kg)

0 1 2 3

6N-11Ni01Ei01W

12S-4Si02Ei02W

22S-20Si03Ei03W

66S-60SS4iES4iW

(a)

(b) (d)

(c)

N=

16.8

(P- 0

.27)

N=

16.9

(P- 0

.23)

N=

16.4

(P -

0.18

)

N=

16(P

- 0.2

7)

I

II

I+II

III

IVV

I+II

III

IVe

IVw

I+II

IV*w

III*N

N

N

a

a

b

a

b

a

b

Fig. 6. P versus N ( ¼ nitrate+nitrite; mmol/kg) versus y for the same transects as in Fig. 2. See the text and Eq. (3a)–(3b) for the

explanation of the lines Na and Nb.


for microbial N2 fixation, but does not indicate anyactual N2 fixation and its intensity.

In contrast, the data points for the AntarcticBottom Water (I), Circumpolar Deep Water (II),and Antarctic Intermediate Water (III) from cruisesi03 and i02, which have the highest P and N

concentrations (Figs. 4b and c), all fall below thestraight line Na in Figs. 6b and c. The AntarcticIntermediate Water (III) mixes smoothly with thedeeper waters (end members I and II) and with theshallower Sub-Antarctic Oxygen Maximum Water(IV). The N/P slope of much less than 16.570.5 forthe Antarctic Intermediate Water (III) again can bebest explained by the partial nitrification in oxyge-nated water columns as discussed earlier. The P–N

relationship resulted from in situ oxidation oforganic matter and the mixing of end members IIIand IV can be represented by an empirical cubicequation such as

Nb ¼ c0 þ c1Pþ c2P2 þ c3P

3; (4b)

where ci are constants, and Nb represents theexpected concentration of N from the right handside of Eq. (4b) during a partial nitrification (withnet N2, N2O and NO production). The ci can beobtained by the least-square fitting of the datapoints with depths equal to and shallower thanthose of the Antarctic Intermediate Water (III)(y4�5–6 1C) and NX�1 mmol/kg, using the Statis-tical Package for the Social Sciences (SPSS) curvefitting program. Nb is always equal to or lessthan Na.

For a given P, when Na4N4Nb, we define thenitrate deficit by the partial nitrification (dN) as

dN ¼ Na �N: (4c)

The Equatorial Indian Phosphate Maximum Water(III*) data also fall below the Na line (Fig. 6d) andform a smooth mixing line Nb with the end membersI+II and the eastern Equatorial Indian CentralWater (IV*e). However, the samples shallower thanthe end member III* in the western Indian Ocean,

ARTICLE IN PRESS

10

20

30

40

N (

umol

/kg)

5

10

15

20

25

30

5

10

15

20

25

30

θ

5

10

15

20

25

30

0 1 2 3P (umol/kg)

0 50 100 150 200 250O

2 (umol/kg)

0 1 2 3P (umol/kg)

34.5 35.0 35.5 36.0 36.5 37.0S

i07n 20S<lat.<4N 26N>lat.>4N

(a) (c)

(b) (d)

θθ

N = 1

6(P

- 0.1

6)

N

a

b

Fig. 7. (a) P vs. N, (b) P vs. y, (c) O2 vs. y, and (d) S vs. y for the cruise i07n data. Na and Nb are Eqs. (5a) and (5b) in the text.


including the western Equatorial Indian CentralWater (IV*w), all fall below the mixing line Nb

(crosses in Fig. 6d), and have low O2 concentrationbut not below 2 mmol/kg (Fig. 3d). Those samplesrepresent the nitrate deficit by denitrification, mixedin from the main denitrification zone of the ArabianSea (see the discussion later).

For a given P, when NoNb, the nitrate deficitcan be caused by both partial nitrification anddenitrification. We define here the nitrate deficit bypartial nitrification (dN) operationally as

dN ¼ Na2Nb (4d)

and the nitrate deficit by denitrification (dN00) as

dN 00 ¼ Nb2N: (4e)

In practice, Na is almost equal to Nb for sampleswith P less than 1.6–1.8 mmol/kg (Figs. 6b–d).Therefore, dN are set to zero for all shallow watersamples with Pp�1:6mmol=kg to eliminate unne-cessary data noise.

The internal consistency and usefulness of thenew parameters dN and dN00 are demonstrated by

the following example for the cruise i07n, whichtraverses the Arabian Sea from 271N to 181S(Fig. 1). As shown in Fig. 7a, we first obtained thefollowing Na and Nb equations for the cruise i07n bycurve fittings:

Na ¼ 16ðP� 0:16Þ (5a)

and

Nb ¼ �3:223þ 16:772Pþ 0:374P2 þ 0:465P3: (5b)

The data points between 41N and 261N (crossesin Fig. 7a) and with depth shallower than thephosphate maximum (Fig. 7b) all fall below themixing line Nb in Fig. 7a. Those data points alsoencompass the oxygen minimum zone witho�2 mmol/kg (Fig. 7c), confirming the dominanceof the denitrification process. In contrast, the datapoints at and deeper than the phosphate maximumall fall between the Na and Nb lines and encompassthe oxycline below the oxygen minimum zone(Fig. 7c), confirming the dominance of the partialnitrification process.


4. Spatial distributions of dN and dN00

Substituting Eqs. (5a)–(5b) into Eqs. (4d)–(4e),the dN and dN00 values for each given pair of P andN in the cruise i07n were calculated. From thespatial distribution patterns of dN and O2 concen-trations (Fig. 8; all in units of mmol/kg), it is evidentthat the dN maximum zone (X7 mmol/kg) lieswithin the oxycline just below the oxygen minimumzone from 261N down to the equator. South of theequator, the dN maximum tends to coincide withthe oxygen minimum. The dN maximum in thelower oxycline is on the order of 8 mmol/kg (Fig. 8),while the N2O maximum in the oxyclines of theIndian and the equatorial eastern Pacific Oceans isonly on the order of 0.08 mmol/kg (Nevison et al.,2003). Allowing some diffusive or mixing loss ofN2O, most of dN must be caused by conversion of(NH3)org into N2 to keep the nitrogen mass balance.The dN maximum zone (X7 mmol/kg) is in contactwith the continental slope sediments at depthsbetween 1000 and 1700m (Fig. 8), indicating somenitrate deficit sources from sediments probably by

Fig. 8. dN (nitrate deficit by partial nitrification; mmol/kg) and O

all N2-producing processes including partial nitrifi-cation, denitrification and anammox. It is unlikelythat nitrate deficit sources from sediments canexplain the whole integrated dN throughout cruisei07n (Fig. 8). Since dN maximum zone coincideswith the P or N maximum zone (compare Figs. 7band 10a) and lies within the lower oxycline, one canenvision that some yet unidentified bacteria haveevolved there to utilize both oxygen and nitrate tooxidize organic matter. Identification of thesebacteria and elucidation of the detailed pathwaysof the partial nitrification in the oxyclines areurgently needed.

The spatial distributions of dN00 and nitrite asshown in Fig. 9 are confined within the area definedby the oxygen contour line of 10 mmol/kg (Fig. 8).The dN00 maximum (10 mmol/kg) is in contact withthe upper continental slope sediments at depthsbetween 300 and 500m (Fig. 9), indicating somedenitrification inputs from the sediments as well.There is no easy way to distinguish the relativemagnitude of the dN00 inputs from sediments andfrom in situ water column production. Obviously,

2 contours in a latitude-depth space for the cruise i07n data.

ARTICLE IN PRESS

Fig. 9. dN00 (nitrate deficit by denitrification; mmol/kg) and nitrite (mmol/kg) contours in a latitude-depth space for the cruise i07n data.


the spatial extent of dN00 is much smaller than thatof dN. The three dN00 maxima (X10 mmol/kg) areaccompanied by three secondary nitrite maximaand lie within the oxygen contour line of 2 mmol/kg(Fig. 8), which is the upper limit of suboxiccondition as defined earlier. The surface dN00

maximum around 181N in Fig. 9 is related to theupwelling along the Omani coast of the East Africaduring the summer SW monsoon. The almostcontinuous primary nitrite maximum from 271Nto 181S at depth around 100m is probably causedby the release of nitrite by bacteria and phytoplank-ton that reside just below the chlorophyll maximumand reduce nitrate to nitrite internally (assimilativereduction; Karl and Michaels, 2001). This processwill be discussed later.

In the plot of dN or dN00 versus potentialtemperature for the cruise i07n data (north of 41Nonly; Fig. 10a), the numbers without double primeindicate possible end members of dN (large changein slope). Peaks 2 and 3 represent, respectively, thedN maximum in the lower and the upper oxyclines.Variation of dN can be explained by the mixingof dN end members in the simple order of 1–5

(Figs. 10a–b). Fig. 10b also shows the inverserelationship between O2 and dN (or positive correla-tion between AOU and dN) for samples between endmembers 1 and 2, and shows a slope dN\-O2 of aboutð6� 1Þ\130. Therefore, one may rewrite the reminer-alization ratios for the deep equatorial Indian Ocean(Li and Peng, 2002) into P\N\dN\Corg\-O2 ¼ 1\ð10�1Þ\ð6� 1Þ\ ð94� 5Þ\ð130� 7Þ. The (N+dN)/P ratioof 16 is as expected from the model. The numberswith double prime indicate the possible end membersof dN00. Peak 200 represents the dN00 maximum; 100 and300 the lower and upper boundaries of the denitrifica-tion zone; 400 the subsurface dN00 maximum; and 500

the upwelled surface dN00 maximum (Fig. 9). Mixingof peak 200 with 100 and 300 forms the broad dN00

maximum within the denitrification zone (Figs. 10aand c). Samples with finite dN00 and O2 values all canbe produced by mixing end members 000 and 400 with100, 200 and 300 (Fig. 10c). The sharp nitrite peak at O2

less than 2mmol/kg (Fig. 10d) corresponds nicely tothe sharp dN00 peak in Fig. 10c.

Our dN and dN00 and the parameter N* byDeutsch et al. (2001) are conceptually quite differ-ent. However, they can be related in magnitude by

ARTICLE IN PRESS

0

5

10

15

20

25

30

θ

2

4

6

8

10

12

14

2

4

6

8

10

dN

1

2

3

4

5

0 5 10 15dN or dN" (umol/kg)

0 50 100 150 200O2 (umol/kg)

0 50 100 150 200O2 (umol/kg)

0 50 100 150 200O2 (umol/kg)

dN"dN

dN"

Nitr

ite

(a)

(b)

(c)

(d)

Fig. 10. (a) dN or dN00 vs. y, (b) O2 vs. dN, (c) O2 vs. dN00, and (d) O2 vs. NO2� for the cruise i07n data north of 41N. Numbers without and

with double prime represent, respectively, the possible end members of dN and dN00. See the text for details.


the following relationship, if N̄ in Eq. (2b) (Deutschet al., 2001) is made equal to our Na

�N� ¼ dN þ dN 00: (6)

Therefore, the �N* values will give a rough estimateof the total nitrate deficit by partial nitrification anddenitrification. For a steady ocean, nitrate convertedinto N2 through partial nitrification and denitrifica-tion should be balanced by the gain through N2

fixation as suggested by Gruber and Sarmiento(1997), Deutsch et al. (2001) and Karl et al. (2002).Nitrogen-fixing bacteria are ubiquitous in subtropi-cal and tropical open-ocean habitats (Karl et al.,2002; Church et al., 2005), and may be able to fixenough N2 to balance the nitrate lose. The recentestimate of the total N2 fixation by Trichodesmiumcyanobacterium in the ocean (Capone et al., 2005)indeed suggests this balance. Negative Na values inshallow water samples with PoPo in our modelroughly correspond to some of the positive N* valuesgiven by Gruber and Sarmiento (1997). A negative

Na (or positive N*) only indicates zero N concentra-tion and may suggest the potential site for microbialN2 fixation, but does not translate into the degree orintensity of N2 fixation as postulated by Gruber andSarmiento (1997).

Similar studies of dN and dN00 in the easternIndian Ocean (including the Bay of Bengal) aresummarized in Figs. 11 and 12, using the WOCEdata for cruises i08s and i09n (Fig. 1; hereafter asi8s9n). Na and Nb obtained by curve fitting for thecruise i8s9n are

Na ¼ 16:5ðP� 0:20Þ (7a)

and

Nb ¼ �2:7573þ 14:6025Pþ 2:046P2 � 0:7488P3:

(7b)

The dN maximum (X5 mmol/kg) again lies withinthe lower oxycline below the oxygen minimumzone and intersects continental slope sediments at

ARTICLE IN PRESS

Fig. 11. dN (nitrate deficit by partial nitrification; mmol/kg) and O2 contours in a latitude-depth space for the cruise i8s9n data.


depth between 800 and 1400m (Fig. 11). Otherwise,distribution patterns of dN and �O2 are verysimilar.

The distribution of dN00 maximum (X2 mmol/kg;Fig. 12a) is confined within a small depth range(50–150m) and a narrow latitudinal interval(18–211N). However, it does not associate with theoxygen minimum (Fig. 11b). Therefore, this dN00

maximum must be from sediment inputs. As noticedby Rao et al. (1994), the nitrite maximum in the Bayof Bengal is the primary nitrite maximum and is notrelated to any denitrification process. The subsur-face and surface nitrite maxima (Fig. 12b) are againrelated to nitrite production/release by bacteria andphytoplankton below or within the chlorophyllmaxima. Some surface samples south of 401S havehigh P concentration (1.4–2.0 mmol/kg) and highN/P slope (416.5), indicating additional oxidationof nitrite into nitrate.

5. Summary and conclusions

1. The newly calculated remineralization ratiosalong E–W sections of the WOCE data in the deep

Indian Ocean confirm the earlier conclusion that theremineralization ratios change systematically withlatitude and from one ocean to another. In addition,the remineralization ratios are also shown to changewith water depth. Therefore, any assumption ofconstancy of remineralization ratios throughout thewhole ocean is not warranted.

2. The low remineralization ratio of N/P (con-siderably less than 16) for deep oxic waters can bebest explained by the partial conversion of organicnitrogen into N2, N2O, and NO during oxidation oforganic matter in oxyclines by yet unidentifiedbacteria (i.e. the partial nitrification hypothesis).The unidentified bacteria may resemble Nitrosomo-

nas europaea in chemostat cell cultures (Schmidtet al., 2004) and/or the aerobic and anaerobicammonia oxidizing bacteria coupled within suspendedorganic particles in a low oxygen CANON reactor(Nielsen et al., 2005). Identification of those uni-dentified bacteria and the actual microbial reactionpathways (both oxidation and reduction processes)for the partial nitrification process is urgently needed.

3. The average remineralization ratios for theIndian warm water masses (y4�10–12 1C) are

ARTICLE IN PRESS

Fig. 12. dN00 (nitrate deficit by denitrification; mmol/kg) and nitrite (mmol/kg) contours in a latitude-depth space for the cruise i8s9n data.


P\N\Corg\-O2 ¼ 1\ð15:6� 0:7Þ \ð110� 9Þ\ð159� 8Þ.These are comparable to the traditional Redfieldratios ðP\N\Corg\-O2 ¼ 1\16\106\138Þ, and are ingood agreement with the average suggested byAnderson (1995; P\N\Corg\-O2 ¼ 1\16\106\150)within the given uncertainties.

4. It is time to discontinue the usage of Redfield’sformula for marine plankton (CH2O)106(NH3)16(H3PO4), but not his concept. The preferred formulahere is C106H48(H2O)38(NH3)16(H3PO4) by Ander-son (1995).

5. Separation of nitrate deficits resulting fromaerobic partial nitrification (dN) and anaerobicdenitrification (dN00) using empirical equations isshown to be useful and consistent with otherobservations. The dN maximum coincides with theP and N maximums, lies within the oxycline belowthe oxygen minimum zone, and intersects thecontinental slope sediments. In contrast, the dN00

maximum lies within the oxygen minimum zonewith O2o�2 mmol/kg (i.e. denitrification zone), isalways associated with a secondary nitrite max-imum in the water column, and is in contact withthe continental shelf or upper slope sediments.

6. Remineralization ratios for the deep equatorialIndian Ocean can be written as P\N\dN\ Corg\-O2 ¼

1\ð10� 1Þ\ð6� 1Þ\ð94� 5Þ\ð130� 7Þ. The (N+dN)/P ratio of 16 is consistent with the partialnitrification hypothesis.

7. The spatial extent and magnitude of dN00 aremuch larger in the Arabian Sea than in the Bay ofBengal. The spatial extent of dN in the IndianOcean is much wider than that of dN00, indicatingthe prevalence of the partial nitrification over thedenitrification processes.

8. Conceptually, our dN and dN00 are quitedifferent from the parameter N* defined by Deutschet al. (2001). However, they can be roughly relatedin magnitude by �N*

¼ dN+dN00.9. Negative Na values in shallow water samples

with PoPo roughly correspond to some of thepositive N* values given by Gruber and Sarmiento(1997). Negative Na (or positive N*) means onlyzero concentrations of nitrate and nitrite, and maypoint to potential sites for microbial N2 fixation, butdoes not translate into the degree or intensity of N2

fixation as postulated by Gruber and Sarmiento(1997).

ARTICLE IN PRESSY.-H. Li et al. / Deep-Sea Research I 53 (2006) 94–110 109

10. The almost continuous primary nitrite max-imum from the northern coastal margin down to401S at depths around 100m is probably caused bythe release of nitrite by bacteria and phytoplanktonthat reside just below the chlorophyll.

Acknowledgment

Lauren Kaupp patiently showed one of us (YHL)how to use the Ocean Data View PC program,which was kindly provided by Dr. Rainer Schlitzer.Discussions with Drs. James Cowen, David Karl,Marcel Kuypers, Fred Mackenzie, Wajih Naqvi andHiroaki Yamagishi were most fruitful. Two anon-ymous reviewers provided useful comments andsuggestions. The final draft was finished at theHanse Institute for Advanced Study, Delmenhorst,Germany, when YHL was a Fellow (6/1/05–11/30/05). Diane Henderson provided helpful editorialassistance. This work was supported by the NOAA/OAR/OGP Grant GC02-386. SOEST contribution6649.

References

Alldredge, A.L., Cohen, Y., 1987. Can microscale chemical

patches persist in the sea? Science 235, 689–691.

Amend, J.P., Shock, E.L., 2001. Energetics of overall metabolic

reactions of thermaphilic and hyperthermophilic archaea and

bacteria. FEMS Microbiology Review 25, 175–243.

Anderson, L.A., 1995. On the hydrogen and oxygen content of

marine phytoplankton. Deep-Sea Research I 42, 1675–1680.

Anderson, L.A., Sarmiento, J.L., 1994. Redfield ratios of

remineralization determined by nutrient data analysis. Global

Biogeochemical Cycles 8, 65–80.

Broecker, W.S., Takahashi, T., Takahashi, T., 1985. Sources and

flow patterns of deep-ocean water as deduced from potential

temperature, salinity, and initial phosphate concentration.

Journal of Geophysical Research 90, 6925–6939.

Capone, D.G., Burns, J.A., Montoya, J.P., Subramaniam, A.,

Mahaffey, C., Gunderson, T., Michaels, A.F., Carpenter,

E.J., 2005. Nitrogen fixation by Trichodesmium spp.: an

important source of new nitrogen to the tropical and

subtropical North Atlantic Ocean. Global Biogeochemical

Cycles 19, GB2024.

Church, M.J., Jenkins, B.D., Karl, D.M., Zehr, J.P., 2005.

Vertical distributions of nitrogen-fixing phylotypes at Stn

Aloha in the oligotrophic North Pacific Ocean. Aquatic

Microbial Ecology 38, 3–14.

Codispoti, L.A., Christensen, J.P., 1985. Nitrification, denitrifica-

tion and nitrous oxide cycling in the eastern tropical South

Pacific Ocean. Marine Chemistry 16, 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? Scientia Marina 65 (2), 85–105.

Dalsgaard, T., Thamdrup, B., 2002. Factor controlling anaerobic

ammonium oxidation with nitrite in marine sediments.

Applied and Environmental Microbiology 68, 3802–3808.

Deutsch, C., Gruber, N., Key, R., Sarmiento, J.L., Ganachaud,

A., 2001. Denitrification and N2 fixation in the Pacific Ocean.

Global Biogeochemical Cycles 15 (2), 483–506.

Devol, A.H., 1978. Bacterial oxygen uptake kinetics as related to

biological processes in oxygen deficient zones of the oceans.

Deep-Sea Research 25, 137–146.

Gruber, N., Sarmiento, J.L., 1997. Global patterns of marine

nitrogen fixation and denitrification. Global Biogeochemical

Cycles 11, 235–266.

Karl, D.M., Michaels, A.F., 2001. Nitrogen cycle. In: Steele, J.,

Thorpe, S., Turekian, K. (Eds.), Encyclopedia of Ocean

Sciences. Academic Press, London, pp. 1876–1884.

Karl, D.M., Michaels, A., Bergman, B., Capone, D., Carpenter,

E., Letelier, R., Lipschultz, F., Paerl, H., Sigman, D., Stal, L.,

2002. Dinitrogen fixation in the world’s oceans. Biogeochem-

istry 57/58, 47–98.

Kuypers, M.M.M., Sliekers, A.O., Lavik, G., Schmid, M.,

Jorgensen, B.B., Kuenen, J.G., Damste, J.S.S., Strous, M.,

Jetten, M.S.M., 2003. Anaerobic ammonium oxidation by

anammox bacteria in the Black Sea. Nature 422, 608–611.

Kuypers, M.M.M., Lavik, G., Woebken, D., Schmid, M., Fuchs,

B.M., Amann, R., Jorgensen, B.B., Jetten, M.S.M., 2005.

Massive nitrogen loss from the Benguela upwelling system

through anaerobic ammonium oxidation. Proceedings of

National Academy of Science, 6478–6483.

Li, Y.H., Peng, T.H., 2002. Latitudinal change of remineraliza-

tion ratios in the oceans and its implication for nutrient

cycles. Global Biogeochemical Cycles 16, 1130–1145.

Mantoura, R.F.C., Law, C.S., Owens, N.J.P., Burkill, P.H.,

Woodward, E.M.S., Howland, R.J.M., Llewillyn, C.A., 1993.

Nitrogen biogeochemical cycling in the northwestern Indian

Ocean. Deep-Sea Research 40, 651–671.

Naqvi, S.W.A., Noronha, R.J., 1991. Nitrous oxide in the

Arabian Sea. Deep-Sea Research 38, 871–890.

Naqvi, S.W.A., Sen Gupta, R., 1985. ‘NO’, a useful tool for the

estimation of nitrate deficits in the Arabian Sea. Deep-Sea

Research 32, 665–674.

Naqvi, S.W.A., Noronha, R.J., Reddy, C.V.G., 1982. Denitrifi-

cation in the Arabian Sea. Deep-Sea Research 29, 459–469.

Naqvi, S.W.A., Noronha, R.J., Somasundar, K., Sen Gupta, R.,

1990. Seasonal changes in the denitrification regime of the

Arabian Sea. Deep-Sea Research 37, 593–611.

Nevison, C., Butler, J.H., Elkins, J.W., 2003. Global distribution

of N2O and the (N2O-AOU) yield in the subsurface ocean.

Global Biogeochemical Cycles 17, 1119–1136.

Nielsen, M., Bollmann, A., Sliekers, O., Jetten, M., Schmid, M.,

Strous, M., Schmid, I., Larsen, L.H., Nielsen, L.P., Revsbech,

N.P., 2005. Kinetics, diffusional limitation and microscale

distribution of chemistry and organisms in a CANON

reactor. FEMS Microbiology Ecology 51, 247–256.

Rao, C.K., Naqvi, S.W.A., Kumar, M.D., Varaprasad, S.J.D.,

Jayakumar, D.A., George, M.D., Singbal, S.Y.S., 1994.

Hydrography of the Bay of Bengal: possible reasons for a

different water-column cycling of carbon and nitrogen from

the Arabian Sea. Marine Chemistry 47, 279–290.

Redfield, A.C., 1958. The biological control of chemical factors in

the environment. American Science 46, 205–221.

Schlitzer, R., 2000. Electronic atlas of WOCE hydrographic and

tracer data now available. Eos Transaction AGU 81 (5), 45.


Schmidt, I., Rob, van Spanning, J.M., Jetten, M.S.M., 2004.

Denitrification and ammonia oxidation by Nitrosomonas

europaea wild-type, and NirK- and NorB-deficient mutants.

Microbiology 150, 4107–4114.

Shaffer, G., 1996. Biogeochemical cycling in the global ocean 2:

new production, Redfield ratios, and remineralization in the

organic pump. Journal of Geophysical Research 101 (C2),

3723–3745.

Shaffer, G., Bendtsen, J., Ulloa, O., 1999. Fractionation during

remineralization of organic matter in the ocean. Deep-Sea

Research I 46, 185–204.

Spencer, D., Broecker, W.S., Craig, H., Weiss, R.F., 1982.

GEOSECS Indian Ocean Expedition, vol. 6. Sections and

Profiles, US Government Printing Office.

Strous, M., Fuerst, J.A., Kramer, E.H.M., Logemannn, S.,

Muyzer, G., Van De-Pas-Schoon, K.T., Webb, R., Kuenen,

J.G., Jetten, M.S.M., 1999. Missing lithotroph identified as

new planctomycete. Nature 400, 446–449.

Thamdrup, B., Dalsgaard, T., 2002. Production of N2 through

anaerobic ammonium oxidation coupled to nitrate reduction

in marine sediments. Applied and Environmental Microbiol-

ogy 68, 1312–1318.

Van de Graaf, A.A., de Bruijin, P., Robertson, L.A., Jetten,

M.S.M., Kuenen, J.G., 1997. Metabolic pathway of anaerobic

ammonium oxidation on the basis of 15N studies in a

fluidized bed reactor. Microbiology 143, 2415–2421.

Ward, B.B., Zafiriou, O.C., 1988. Nitrification and nitric oxide in

the oxygen minimum of the eastern tropical North Pacific.

Deep-Sea Research 35, 1127–1142.

Ward, B.B., Glover, H.E., Lipschultz, F., 1989. Chemoauto-

trophic activity and nitrification in the oxygen minimum zone

off Peru. Deep-Sea Research 36, 1031–1051.

Yamagishi, H., Yoshida, N., Toyoda, S., Popp, B.N., Westley,

M.B., Watanabe, S., 2005. Contributions of denitrification

and mixing on the distribution of nitrous oxide in the North

Pacific. Geophysical Research Letters 32, L04603(pp4).

nitrate deficits by nitrification and denitrification …deep-sea research i 53 (2006) 94–110...

Documents