Sulfate reduction in shelf sediments in the upwelling region off Central Peru

ContinentalShelfResearch,Vol. 10, No. 4, pp. 355-367, 1990. Printed in Great Britain.

0278--4343/90$3.00 + 0.00 ~) 1990PergamonPressplc

Sulfate reduction in shelf sediments in the upwelling region off Central Peru HENRIK FOSSING*

(Received 12 June 1989; accepted 8 February 1990) Abstract---Concentration profiles of sulfate, hydrogen sulfide, ferrous sulfide, elemental sulfur, and pyrite are presented from five stations from the upwelling region off Peru near 15°S together with profiles of water content, density and organic content. Four of the stations were located within the oxygen minimum zone (80-500 m depth) and one was located at 2650 m depth. Sulfate reduction rates were determined from radiotracer measurements on all stations by acid distillation followed by chromium reduction. Reduced 35S recovered as acid volatile sulfide (H2S and FeS) was compared to 35S recovered as chromium reducible sulfur (S ° and FeS2). In the top 10 cm of the sediment 15-40% of 35Sred was recovered as acid volatile sulfide, deeper into the sediment it was less than 10%. Maximum reduction rates of 68-312 nmol cm -3 day- 1 were observed within the top 4 cm of sediment from stations in the oxygen minimum zone. Integrating sulfate reduction rates over whole sediment cores under the oxygen depleted water column gave an average rate of 20 + l l m m o l SO 2- m -2 day - l (n = 3). This corresponds to the mineralization of 9-29% of the planktonic primary production in the overlying water column,

INTRODUCTION

THE upwelling region off Peru has very high primary productivities of 1-10 g C m -2 day-1 (RrmER et al., 1971; LEE and CsoNI1q, 1982; HErqFaCrlSand FAsmN~rOrq, 1984). This is 10 times higher than the average primary productivity of the ocean, 0.1-0.2 g C m -2 day -1 (MENzEL and RrraEt, 1960). Up to 90% of the primary production is remineralized within the water column (LEE and CRol~Irq, 1982; RowE, 1985), resulting in oxygen depletion from 80 to 500 m depth (<0.1 ml 021-1) in the upweUing zone. When oxygen is depleted the microbial degradation of organic material takes place with NO3 or NO2 as terminal electron acceptors (CoDISPOTI, 1983; CODISPOTIand CnPaSTENSEN, 1985; CoI)ISPOrIet al., 1986). Total depletion of NO3 in the water column has been observed only once in the Peru upwelling area when H2S was detected (DUGDALEet al., 1977). In the sediment the mineralization processes proceed with NO3, NO2 and SO2- as electron acceptors. Despite the importance of sulfate as the terminal electron acceptor in marine sediments in general (FRoELICHet al., 1979; JORGENSEN,1983; LEIN, 1984; Sr~'lUr~G, 1987) only one study of sulfate reduction has been reported in sediments influenced by upwelling (Rowe and HOWARTIa,1985). In the study of Rowe and Howarth, maximum sulfate reduction rates of approx. 30 nmol cm -3 day- 1 were measured between 9 and 12 cm depth and total sulfate reduction rates were 5--6 mmol m -2 day -i, calculated by *Institute of Ecology and Genetics, University of Aarhus, Ny Munkegade, DK-8000 Aarhus C, Denmark. 355

356

H. FOSSING i

14°50 S'

~~-~:":!:.

,t/ R/V Moana Wave :~'aca , • July 1987

~

5.:..:'

~

15° 10

'~

i~.....;,,.i:...:. "

00,

2~...

-

! ~ : ~ ~i!.-.-:..... \

~

x

~

ni:a San Nicolas .w.::......

• 10 c~0tm ZN~TB~ 0. 5 10 t,s 2.0 KILOMETERS

].5o30 76°00 'W Fig. 1.

f

I

i 75°30 '

75°00 '

Sampling locations in the upwelling region off Central Peru.

integration over the depth of the cores. These rates were 10 to 100-fold higher than observed ocean-wide at comparable water depth (SrYPdNG, 1987). Presented here are sulfate reduction rate measurements from five stations visited during PUBS I expedition with R.V. Moana Wave (cruise 87-08), July 1987. The purpose is to report how fractions of reduced sulfate are incorporated into acid volatile sulfide (AVS: H2S and FeS) and chromium reducible sulfur (CRS: S° and FeS2). In addition, the sediment chemistry of SO42-, H2S, S°, FeS and FeS2 are also reported and discussed. MATERIAL AND METHODS The depths ranged from 130 to 2500 m and ranged in bottom water oxygen content from <2 to 100% of air saturation. Stations 4, 6, 8 and 10 were located 15-450 km from the coast near 15°S, whereas Sta. 2 was positioned further north at 11°S (Fig. 1, Table 1).

Sampling and sediment handling Sediment was obtained from all stations by a box corer (BC) (Sandia-Hessler MK-III Ocean Instruments, San Diego; 0.25 m 2, depth 30 cm). Sediment cores of varying lengths were also sampled with gravity corer (GC) (200 cm length, 15 cm diameter) at Stas 2, 4 and 8 and at Sta. 6 with a piston corer (PC) (500 cm length, 10 cm diameter). Sediment was also sampled by a Soutar corer (SC) (0.1 m 2, depth 1 m) at Sta. 4 (Table 1). Sediments from BC and SC were visually inspected for an undisturbed surface. The GC and PC were always recovered with the surface layer lost. The loss of sediment was estimated from the chemical data when these observations were aligned with the data obtained from the BC. Subsampling was done with Plexiglas coring tubes pressed into the sediment. To avoid compression during subsampling, a piston mounted inside the coring tube was pulled simultaneously as the tube was pressed into the mud. Subcores of 15-20 cm length were sampled in triplicate from BC, GC, PC and SC with 3 cm diameter coring

Sulfate reduction in shelf sediments

357

Table 1. Sampling stations and sediment cores sampled

Water depth (m)

Bottom temp. (°C)

Oxygen content (ml 021 -l)

Core type*

Top lostt (cm)

Station

Location

2

11°04.2'S 78°31.2'W

255

14

0.07

BC GC

0 11

4

15°06.2'S 75°42.1 'W

268

14

0.04

BC SC GC

6 0 19

6

15°11.5'S 75°34.5'W

502

15

0.12

BC PC

0 8

8

15°00.0'S 75°39.2'W

135

16

0.08

BC GC

0 16

2650

6

Saturation

BC

0

10

15°20'S 75°50'W

*BC: MK-III box corer; GC: large diameter gravity corer; SC: Soutar type box corer; PC: piston corer. tThe calculated thickness of sediment that was blown off from the surface when sampled.

tubes. The gravity and piston cores were pushed from the coreliners as consecutive subcores were taken. The subcores were all sealed with butyl rubber stoppers and stored by in situ t e m p e r a t u r e <1.5 h before radiolabeling. Analyses Sulfate reduction measurement. Injections of 2 ktl carrier-free 35SO2- (200 MBq m1-1) ( A m e r s h a m Corp.) were done horizontally, at 1 cm intervals, into the sediment through silicone-stoppered ports (JORGENSEN, 1978). The radiolabeled sediment was incubated for 8-16 h at in situ t e m p e r a t u r e (cf. Table 1) in a thermostated van. To prevent AVS oxidation, the cores were extruded in 1 or 2 cm steps from the coreliner, immediately cut into 10 ml 5% zinc acetate fixative in a 20 ml vol. plastic vial, thoroughly mixed, and weighed. The sediment was then centrifuged within the plastic vial and radioactive sulfate determined from 5 ml of a diluted subsample of the supernatant mixed with 5 ml of scintillation liquid (Dynagel, B a k e r Chemicals). The fraction of 35SO2- reduced during incubation was determined from an acid distillation followed by a chromium reduction (ZHABINAand Volkov, 1978; WESTRICH, 1983; HOWARTH and JORGENSEN, 1984; CANFIELD et al., 1986, FOSSING and JORGENSEN, 1989). The sediment pellet was washed twice and homogenized. A 2-3 g subsample was transferred to a reaction flask with condenser and 12 N HCI was added to a final concentration of 2 N HCI. Distillation c o m m e n c e d for 30 miD and liberated all A V S as H2S, which was collected as ZnS in two sequential traps containing 10 ml of 5% zinc acetate (ZnAc) plus a drop of antifoam. The traps were then renewed and chromium reduction proceeded on the same sediment with Cr 2+ and HC1 added. All CRS was dissolved as H2S and carried to the traps within 40 miD. For both digestions, the pairs of Zn35S traps were pooled, a 5 ml subsample was mixed with 5 ml of scintillation fluid, and 35S was counted on a scintillation counter (Packard, Tri-Carb 2200 CA). The sulfate

358

H. FOSSlN~

reduction rates were calculated as described by JORGENSEN (1978) and FOSSlNG and JOR~ENSEN (1989). Radiolabeled elemental sulfur was extracted with 5 ml of CS2 from 2 g of the washed and homogenized sediment. A 2 ml subsample of the CS2 was vaporized, the 35S° was redissolved into 5 ml of scintillation liquid, and counted after addition of 5 ml distilled H20. All radioactivities were quench corrected. Chemical analyses. Sulfate and free sulfide were obtained from porewater by pressure filtering through a 0.45/~m membrane filter (Millipore) under a gas-impermeable latex membrane. The first 5 drops of porewater were discharged. To minimize exposure to the atmosphere up to 2 ml of porewater (determined by weight) was transferred directly into 1.0 ml of 2% ZnCI 2 fixative in a plastic vial. The concentration of sulfate was determined on diluted porewater samples by non-suppressed ion chromatography (Waters) using an anion exchange column and conductivity detector. Free sulfide was measured spectrophotometrically from precipitated ZnS as described by CI~INE (1969), as were concentrations of AVS and CRS on subsamples of the ZnS from the H235Sdistillations. Elemental sulfur concentration was measured on a 1 ml subsample of CS2 from the S° extraction as described by TROELSENand JORGENSEN(1982). Porosity and organic content. Known volumes of sediment were transferred into 20 ml plastic vials and weighed. The densities were determined and used to recalculate rates and concentrations from weight to volume. The sediment was then dried overnight at 105°C to estimate the water content. Porosity was calculated from density and % water content (w/w). Organic content was measured as loss on ignition from the dried sediment (4 h at 450°C). RESULTS The sediment cores were all sampled from the oxygen minimum zone (<0.1 ml 021-1) except from the deep Sta. 10 at 2650 m, where the oxygen concentration was estimated to be at air saturation. From each station, except Sta. 10, the profiles of water content, density, and AVS in the sediments obtained by box, gravity, piston, or soutar coring were aligned in order to estimate how much of the surface was lost when the sediment was sampled (Table 1). All graphic presentations are based on these depth estimates. In the oxygen minimum zone the upper 5-10 cm of the sediment was brown to blackish green, with a gelatin-like consistency and contained fish scales, bones, and other organic debris (in the top 1 cm of fresh sediment, 150-250 g organic material m-2). The porosity ranged from 95 to >99% in the top 10 cm. At Stas 2, 4 and 8 (Figs 2, 3 and 5), filamentous colorless sulfur bacteria of the genus Thioploca were also observed within the upper 15 cm of sediment (GALLARDO,1977), with the highest density at Sta. 8. Deeper into the core the sediment changed to an olive green diatomaceous ooze with yellow laminations and the water content decreased. At Sta. 6 a compact layer of mud was located from 10 to 20 cm on top of sediment with a higher porosity (Fig. 4).

Sulfate At all stations the sulfate concentration was approx. 28 mM in the top 10-20 cm of the cores (Figs 2-6). Below this depth the concentration decreased. Only at Sta. 4 did SOZ4become depleted within the length of the core, at 140 cm depth.

Sulfate reduction in shelf sediments WATER {ml "9-11 0.6 0 .B 1.0 i

,

i ~

i

°~w

DENSITY (g .cm-al x ORGANIC

g-

1.5

20

2 0

(w/w)

40

359 504~- (mM) 10 20

t -~

(mM)

HzS

5

30

10

%°

50. eo

loo

qje

41

ee

t5C

S° (~mol .cm-3) 0.1 •

0.2 0

FeS

{#tool"cm-3) 0.5

!

FeS~ (~mol S .cm-3) 50

I00

150 0

SRRAvs nm01-cm-~.day -t 5 I0

SRRAvs÷cRS rim01 "cm- 3 -(lay - i 50 100

~O0

50.

i / /

1oo

150.

Fig. 2. Station2. Porewaterand solid phase datawith depth in core. Organiccontent determined as percent loss on ignition from dried sediment. SRRAvs: sulfate reduction rate determined from acid distillation of H2S + FeS. SRRAvs+cRS: sulfate reduction rate determined from acid distillation followed by chromiumreduction of S° + FeS2. Increase in sulfate reduction rate below 40 cm (broken line) may be an artifact.

Free sulfide N o free sulfide was detected within the upper 10-15 cm at any of the stations investigated, but below that depth free sulfide concentration increased. This increase was reversely related to the sulfate concentration and reached the highest concentration, 15 mM, at Sta. 4.

Ferrous sulfide The concentration of FeS was calculated from the difference between A V S and free sulfide concentrations. In the oxygen minimum zone, FeS was detected between the sediment surface and the boundary of the free sulfides, which gave the sediment a black appearance. The maximum concentrations of FeS varied from 0.5/zmol cm -3 at Sta. 2 to 2 . 0 / z m o l cm -3 at Sta. 8, with maxima at a depth of 11 and 0.5 cm, respectively. As expected from the grey-colored (oxidized) top sediment at Sta. 10, FeS was first detected deeper in the sediment where the concentration increased up to 5/zmol c m - 3 at 10 cm.

Elemental sulfur Elemental sulfur was not detected at Sta. 4 and only at a very low concentration ( < 0 . 1 /zmol cm -a) within the upper 5 cm at Stas 2 and 8. A maximum concentration of 1.57/zmol

360

H. FOSSING

S ° c m - 3 was measured at Sta. 6 at a depth of 3.5 cm, but S ° was depleted at 10 cm. The more oxidized sediment of Sta. 10 showed an increasing concentration of elemental sulfur below 5 cm.

Pyrite Pyrite normally exceeds elemental sulfur by orders of magnitudes in marine sediments (GOLDHABER and KAPLAN, 1974) and therefore the bulk concentration of CRS should be FeS2. Pyrite was the largest reduced sulfur pool in the sediment at all stations but low in concentration near the surface (Figs 2-6).

Sulfate reduction rates Sulfate reduction rates (SRR) based on A V S alone (SRRAvs) were significantly lower than rates determined from the combined distillation of A V S and CRS (Figs 2--6). When most extreme, SRRAvs+cR s exceeded SRRAvs by more than 10-fold as in the top sediment of Sta. 2. Sulfate reduction was most intense a few cm below the sediment surface on all stations in the oxygen minimum zone. The reduction rate decreased rapidly with depth except at Sta. 2 where sulfate reduction increased again below 40 cm (broken line in Fig. 2). I cannot explain this SRR increase as it is not correlated to any decrease in sulfate concentration. At present, thus, I consider this observation to be an artifact. Station 10 (Fig. 6) revealed no sulfate reduction in the upper 5 cm based on A V 3 5 S alone, but CR35S indicated that sulfate reduction was taking place in the oxidized part of the sediment.

WATER (ml-g-tl

o6 o8 Ip

OENSITY

50

C3

ioo

2 a

i

,l k -~

(g.cm-3)

1.5

ORGANIC (w/w) 20 40 ,

S04z10

i

(mM) 20 30 i , i

HzS (mM) 10 20

30

oO •e

15C

SRRAvs÷CRS

SRRAvs S = (#tool-cm-3l 0.1 i

FeS (amol cm-31 0.2 i

_ ,

•

I i

J~

2 i

FeSz (t/tool S .cm-3) 0 ~I~

50 i

100 i

1500

i

nmol .cm - ~ .day - ~ 10

20

30 0

nmQ1 .cm - 3 - d a y - ~ 50 100

J%

i -~

50~

u

~oo

i50

Fig. 3.

Station 4. Porewater and solid phase data with depth in core. See Fig. 2 legend.

Sulfate reduction in shelf sediments WATER (ml .g-q 0.6 0.8 1.0

.

i

,

q~

a

...

DENSITY (g "cm-3] t.5

% ORGANIC [w/w) 2

......

n

raP"-

0

20

..

40 O

,

361 SO4a1o

HaS (mM)

(raM) 2o

30 o

10

20

30

i

i

i

e

100 E

•

200.

•

•

• -

• o•

NA

°o

••

300.

.2

% o°

,100.

S ° (umol-cm-3) •

Oq

•

1 w J~

'.

(#tool .cm-3) FeSa (#mol S .cm-3) 2.5 5 200 ,u,0

AVS

2 i

,L.v' ,~,.

t

SRRAvs n m o l .cm - a . d a y - ~ 50 tOO ~ '

SRRAvs+cRs lmol,cm-3.day -~ 400 200

tO0 •

50

50

200.

300.

•

•

o

400.

•

Fig. 4.

tO0.

3001

400

400

Station 6. Porewater and solid phase data with depth in core. See Fig. 2 legend. NA: not analysed.

WATER 0.6 i

[mlg -a} 0.8 i ~

,

DENSITY [g 'cm-3}

1.0 u

$ 20

t~

tD0 3DD-

| e°

;,':.

l.fl . . . . . . . . . .

% ORGANIC (w/wl

2 0

20 n

,

40 i

SO4z- (raM) 10 20 a

i

30

HzS lmMl 5 . . . . . . . .

1o ,

,

u

't

It

40

60-

•

B0100 (umol 'cm-3}

S°

o.o Ok •

20

E

0.1 ,f

n

FeS (#mol "CFF3)

0.2

2 '

'e~

I,

FeSa [#tool S .cm-3) 0

g ,

P

50 i

@ ,,,o

t

100 i

•

150 g

SRRAvs nmol .cm-3 -day-~ 10

20

30 0

SRRAvs+cRs nmol -cm-3 .clay-~ |00

•

6o~ 80 1(10

Fig. 5.

Station 8. Porewater and solid phase data with depth in core. See Fig. 2 legend.

200

362

H. Foss]No WATER [ml .g-~}

0,6 . 0.r. , * t.,0

DENSITY[g .cm-al .

.

.

.

t.5 ,

.

.

.

x ORGANIC {w/w) 2 0

20\

,

",

SO4z tO

40 1 i

:"

•

,

HaS (mM}

[mM) 20 .

,

tp

30 0 ,

2p

3p

oo

5•.

E

•

t0,.=, o 15.

20

S= (/Jmol .cm-3) u.,1

8

FeS (umol "cm-31 0.2 0 I

I

2.5 I

I

I

I

FeSa [/Jmol S .cm-3) 5I 0

e•

-

50 0

'p . . . . . . . . . . . . . . . . . .

•

5.

25

SRRAvs

nmol .cm-3 .aay-=

.

E

50

tOOl

SRRAvs+CRS nmol.cm-=.day-~

, , i

50

tO0

.,=

t0, t5

2C

Fig. 6.

Station 10. Porewater and solid phase data with depth in core. See Fig. 2 legend.

DISCUSSION

More than 50% of the total sulfate reduction took place within the upper 20 cm of the sediment at Stas 4, 6 and 8 (Table 2). At these stations, the SRR reached a maximum 1-4 cm below the sediment surface (Figs 3-5). Oxygen concentrations were measured with microelectrodes and showed that oxygen decreased from <0.1 ml 02 1-1 (<5 #M 02) at Table 2. Sulfate reduction rate per m 2 o f the stations investigated. Sulfate reduction rates integrated over the full core length and the top 20 cm o f the sediment are compared to reduced 35S (35Sred) recovered in AVS in % o f total 35Sred (AV35S + CR35S) Sulfate reduction rate (mmol SO 2- m -2 day -n) Full core length

Station

Core length (cm)

2 4 6 8 10

100 160 400 80 10

* Overestimated: see text.

0-20 cm

355re d

355re d

SRR

in AVS (%)

SRR

in A V S (%)

32.5* 8.8 34.3 16.5 5.2

7.1 26.1 20.4 14.1 32.0

6.0 4.4 25.5 9.3 --

11.3 32.3 19.6 16.8 --

Sulfate reduction in shelfsediments

363

the surface to zero within <2 mm (NELSON and FOSSING, unpublished results). The oxygen consumption was <5 mmol 02 m -2 day-1 as calculated from this oxygen gradient. Oxygen was therefore only of minor importance for the mineralization of organic matter in the sediment compared to sulfate. As nitrate and nitrite were not measured in the sediment on this cruise it is not possible to estimate the importance of denitrification on these stations. Sulfate reduction in sediments under the oxygen minimum zone Sulfate which was reduced in the top 15 cm was immediately re-oxidized or precipitated with ferrous iron, as no free sulfide was detected. In the sediment tops at Stas 2, 4 and 8, colorless sulfur bacteria of the genus Thioploca are probably associated with the oxidation of H2Sto elemental sulfur (LARKINand STROHL,1983; MAIERand GALLARDO,1984). So far, the mechanism of this oxidation has not been elucidated, as Thioploca has never been successfully cultured. The relative recovery of reduced 35S (355red) in AVS showed an exponential decrease with depth (Fig. 7). Between 15 and 40% of the sulfate reduced within the upper 10 cm of the sediment was recovered as AVS. Less than 10% of 35Sred was recovered as AV35S below 20 cm where a considerable sulfate reduction was still taking place (Table 2). Significantly larger fractions of 355red have been recovered in AVS in shallower marine sediments and salt marshes, typically 50-90% (WESTmCH, 1983; HOWARTHand JORGENSEN, 1984; HOWES et al., 1984; KING et al., 1985; FOSSlNGand JORGENSEN, 1989). LEIN et al. (1982) reported as little as 1% of reduced 35S in AvaSs during studies of sulfate reduction in the Baltic Sea. The mechanism of tracer incorporation into the sulfur compounds that comprise the CRS pool, mainly FeS2 is still not clear. The CRS pool comprises pyrite, elemental sulfur, and other reduced sulfur compounds that are not dissolved during cold acid distillation, e.g. greigite (BERNER, personal communication in HOWARTHand JORGENSEN, 1984). Also elemental sulfur that is formed from oxidation of H2S by Fe 3÷ during acid distillation (BERNER,1964, 1974) is recovered by the chromium reduction. However, in nearshore sediments and in estuaries this fraction has been found only to vary from I to 5% of the reduced 355 (HOWARTHand JORGENSEN, 1984; THODE-ANDERSEN and JORGENSEN, 1989). 100 X o10 ta ot •

1 t

0.1

1

i

, % **

i

10 100 DEPTH (cm)

1000

Fig. 7. Recoveryof reduced 35S (35S,~d)in AVS in percent of total reduced 3SS (=AV3SS + CR35S) vs depth (cm) into the sediment. A double logarithmicplot of 115 data points from the oxygenminimumzone (r = -0.67).

Sulfatereductionin shelfsediments

365

they compared the measured SRR to rates estimated from SO~4- fluxes according to Fick's first law of diffusion (cf. ROWEand HOWARTH,1985). However, they argued that the higher SRR most likely were explained by an overestimation of the SO~4- flux at the interface because they "did not know the distribution of SO42- in the bottom few centimeters of the water column or within the top 3 cm of the sediment." (cited). To my concern, the difference could also be explained from the low recovery of radiolabeled CRS. Higher SRR than measured by Rowe and Howarth can be estimated from SO42- concentrations at three stations in the same area, reported by HENRICHSand FARRINGTON(1984). Based on these observations, I have estimated the SRR from the SO42- flux using Fick's first law of diffusion and the same parameters as did Rowe and Howarth. For comparison these SRR are presented in Table 3. Based on SRR in this study, I find an average rate of 20 _+ 11 mmol SO ]- m -2 day -1 (n = 3; Table 3) to be a realistic estimate for sediments in the oxygen minimum zone. This SRR is 50 to 300-fold higher than SRR measured on outer continental shelf sediments (150500 m water) outside upwelling areas (IVANOVet al., 1980; BATTERSBYand BROWN, 1982). This difference is probably explained by the high flux of organic matter to the sediment due to intense primary production in the upwelling waters. It is therefore not surprising that sulfate reduction rates are as high as observed in estuaries and coastal marine sediments (SKYRING, 1987; CANFIELD,1989). To evaluate the importance of sulfate reduction for the carbon cycling in the Peruvian upwelling system the sulfate reduction rates were compared to rates of water-column primary production. Using an average SRR of 20 + 11 mmol SO42- m -2 day -1, and assuming that 2 moles of CO2 are formed for each mole SO42- respired, the sulfate reducers are able to remineralize (20-12-2) 480 + 260 mg C m -2 day -1. Primary production ranges from 1 to 4 g C m -2 day -1 in most studies (LEE and CRONIN, 1982; HENRICHS and FARRINGTON,1984), although RYrHERet al. (1971) have reported values up to 11.7 g C m -2 day -1, the highest ever reported for this area. The sulfate reducers thus remineralize 9-29% (0.48 + 0.26/2.5) of the total primary production of which 10-15%, based on sedimentation trap measurements, was estimated to sink to the sea bed (LEE and CRONIN, 1982; HENRICHSand FARRINGTON,1984; ROWE,1985). However, this does not correspond with the observation that about half the organic carbon is permanently buried in the sediments (Figs 2-5). The sedimentation, though, may change from year to year. The sedimentation trap measurements are also sparse and these measurements might not correspond with the sedimentation on the stations investigated during this study. Organicrich sediments could also be transported downslope, as the bottom slope in this area is about 1:10. A higher amount of organic carbon will be available in this way for mineralization than can be calculated from the sedimentation of the local primary production. A nonsteady-state deposition of this kind was proposed by HENRICHS and FARRINGTON (1984) from 21°pb observations. The downslope transportation may also explain the almost uniform SO42- gradients in the top 10-20 cm of the sediment in the oxygen minimum zone (Figs 2-5). No bioturbation was observed on these stations and diffusion alone cannot explain the SO 2- concentrations measured. It is, however, evident from this study that sulfate-reducing bacteria must play a most important role in this sediment ecosystem. Acknowledgements--I thankJohnW. Farringtonfor the invitiationto participatein the PUBSI, 1987expedition

supported by Grant OCE-85-09859 U.S. NationalScienceFoundation. I also thank the officers,crew, and the

366

H, FOSSING

scientific party of R.V. Moana Wave for their contributions to the success of the sampling program. Thanks are especially due to Eben Franks and Douglas C. Nelson for their help during subsampling. Else Bundgaard Frentz carefully assisted with the analytical work and Bo Barker Jcrgensen reviewed the manuscript and provided helpful suggestions. Thanks to two anonymous reviewers for comments on an earlier version of this manuscript. Funding was provided from the Danish Natural Science Research Council, Grant nos 11-5736 and 81-5671.

REFERENCES BATrERSBY N. S. and C. M. BROWN (1982) Microbial activity in inorganically enriched marine sediments. In: Sediment microbiology, D. B. NEDWELL and C. M. BROWN, editors, Academic Press, London, published for the Society for General Microbiology, pp. 147-170. BERNER R. A. (1964) Distribution and diagenesis of sulfur in some sediments from the Gulf of California. Marine Geology, 1,117-140. BERNER R. A. (1974) Iron sulfides in Pleistocene deep Black Sea sediments and their paleo-oceanographic significance. In: The Black Sea--geology, chemistry, and biology, E. T. DEGENS and D. A. Ross, editors, The American Association of Petroleum Geologists, Tulsa, pp. 524-531. CANFIELD D. E. (1989) Sulfate reduction and oxic respiration in marine sediments: implications for organic carbon preservation in euxinic environments. Deep-Sea Research, 36, 121-138. CANFIELD D. E., R. RAISWELL, J. T. WESTRICH, C. M. REAVESand R. A. BERNER (1986) The use of chromium reduction in the analysis of reduced inorganic sulfur in sediments and shales. Chemical Geology, 54,149155. CL1NE J. D. (1969) Spectrophotometric determination of hydrogen sulfide in natural waters. Limnology and Oceanography, 14,454--458. CODISPOTI L. A. (1983) Nitrogen in upwelling systems. In: Nitrogen in the environment, E. J. CARPENTERand D. G. CAPONE, editors, Academic Press, New York, pp. 513-564. CODISPOTI L. A. and J. P. CHRISTENSEN(1985) Nitrification, denitrification and nitrous oxide cycling in eastern tropical South Pacific Ocean. Marine Chemistry, 16,277-300. CODISPOTIL. A., G. E. FRIEDERICH,T. T. PACKARD,H. E. GLOVER,P. J. KELLY,R. W. SPINRAD,R. T, BARBER,J. W. ELR1NS, B. B. WARD, F. LIVSCHULTZand N. LOSXAUNAU(1986) High nitrite levels off Northern Peru: a signal of instability in the marine denitrification rate. Science, 233, 1200-1202. DUGDALE R. C., J. J. GOEVaNG, R. T. BARBER, R. L. SMITH and T. T. PACKARD (1977) Denitrification and hydrogen sulfide in the Peru upweUing region during 1976. Deep-Sea Research, 24, 601-608. FOSSlNGH. and B. B. JORGENSEN(1989) Measurement of bacterial sulfate reduction in sediments: evaluation of a single-step chromium reduction method. Biogeochemistry, 8,205-222. FROELICH P. N., G. P. KLINKHAMMER,M. L. BENDER et al. (1979) Early diagenesis of organic matter in pelagic sediments of the eastern equatorial Atlantic: suboxic diagenesis. Geochimica et Cosmochimica Acta, 43, 1075-1090. GALLARDO V. A. (1977) Large benthic microbial communities in sulphide biota under Peru-Chile Subsurface Countercurrent. Nature, 268, 331-332. GOLDHABER M. B. and I. R. KAPLAN (1974) The sulfur cycle. In: The sea, Vol. 5, Marine chemistry, E. D. GOLDBERG,editor, John Wiley, New York, pp. 569-655. HENRICHS S. M. and J. W. FARRINGTON(1984) Peru upwelling region sediments near 15°S. 1. Remineralization and accumulation of organic matter. Limnology and Oceanography, 29, 1-19. HOWARTH R. W. and J. M. TEAL (1979) Sulfate reduction in a New England salt marsh. Limnology and

Oceanography, 24,598--608. HOWARTH R. W. and A. GIBLIN (1983) Sulfate reduction in the salt marshes at Sapelo Island, Georgia. Limnology and Oceanography, 28, 70-82. HOWARTH R. W. and B. B. JI3RGENSEN(1984) Formation of 35Sqabelled elemental sulfur and pyrite in coastal marine sediments (Limfjorden and Kysing Fjord, Denmark) during short term 35SO~4--reduction measurements. Geochimica et Cosmochimica Acta, 48,180%1818. HOWLS B. L., J. W. DACEYand G. M. KING (1984) Carbon flow through oxygen and sulfate reduction pathways in salt marsh sediments. Limnology and Oceanography, 29,103%1051. IVANOV M. V., A. Yu. LE1N, S. S. BEYLAEV, A. I. NESTEROV, V. A. BONDER and N. N. ZHABINA (1980) Geochemical activity of sulfate-reducing bacteria in benthal sediments in North-West Indian Ocean. Geokimiya, No. 8, pp. 1238-1249.

Sulfate reduction in shelf sediments

367

JORGENSEN B. B. (1978) A comparison of methods for the quantification of bacterial sulfate reduction in coastal marine sediments. I. Measurement with radiotracer techniques. Geomicrobiology Journal, 1, 11-27. JORGENSENB. B. (1983) Processes at the sediment-water interface. In: The major biogeochemical cycles and their i,,teractions, B. BOLIN and R. C. COOK, editors, SCOPE, Stockholm, pp. 477-509. KIN6 G. M. (1983) Sulfate reduction in Georgia salt marsh soils: An evaluation of pyrite formation by use of 3SS and SSFe tracers. Limnology and Oceanography, 28,987-995. KInG G. M., B. L. HowEs and J. W. H. DACEY (1985) Short-term end products of sulfate reduction in a salt marsh: formation of acid volatile sulfides, elemental sulfur, and pyrite. Geochimica et Cosmochimica Acta, 49, 1561-1566. LARKINJ. M. and W. R. STROHL(1983) Beggiatoa, Thiotrix, and Thioploca. Annual Review of Microbiology, 37, 341-367. LEE C. and C. CRONlr~(1982) The vertical flux of particulate organic nitrogen in the sea: decomposition of amino acids in the Peru upwelling area and equatorial Atlantic. Journal of Marine Research, 40, 227-251. LEIN A. Y. (1984) Anaerobic consumption of organic matter in modern marine sediments. Nature, 312, 148-150. LEIN A. Y., M. B. VAYNSHTEYN,B. B. NAMSARAYEV,Y. E. KASHPAROVA,A. G. MATROSOV,V. A. BONDARand M. V. IVANOV (1982) Biogeochemistry of anaerobic diagenesis of recent Baltic sediments. Geochemistry International, 19, 90-103. MAIER S. and V. A. GALLARDO (1984) Nutritional characteristics of two marine thioplocas determined by autoradiography. Archives of Microbiology, 139, 218-220. MENZEL D. W. and J. H. RYTHER (1960) The annual cycle of primary production in the Sargasso Sea off Bermuda. Deep-Sea Research, 6, 351-367. Rowe G. T. (1985) Benthic production and processes off Baja California, Northwest Africa and Peru: A classification of benthic subsystems in upwelling ecosystems. Simposio Internacional Sobre las Areas de Afloramiento mas Importantes del Oeste Africano, Instituto de Investigaciones Pesqueras, Barcelona, pp. 589-612. Rowe G. T. and R. HOWARTH(1985) Early diagenesis of organic matter in sediments off the coast of Peru. DeepSea Research, 32, 43-55. RYTHER J. H., D. W. MENZEL, E. M. HULBURT, C. J. LORENZEN and N. CORWIN (1971) The production and utilization of organic matter in the Peru coastal current. Investigacion Pesquera, 35, 43-59. SKYR1N~ G. W. (1987) Sulfate reduction in coastal ecosystems. Geomicrobiology Journal, 5,295-374. THODE-ANDERSON S. and B. B. J¢RGENSEN(1989) Sulfate reduction and the formation of 35S-labelled FeS, FeS2, and S° in coastal marine sediments. Limnology and Oceanography, 793-806. TROELSEN H. and B. B. JglRGENSEN (1982) Seasonal dynamics of elemental sulfur in two coastal sediments. Estuarine, Coastal and Shelf Science, 15,255-266. WESTRICHJ. T. (1983) The consequences and controls of bacterial sulfate reduction in marine sediments. Ph.D. thesis, Yale University, New Haven, Conn. Zhabina N. N. and I. I. VOLKOV(1978) A method of determination of various sulfur compounds in sea sediments and rocks. In: Environmental biogeochemistry and geomicrobiology, Vol. 3, W. E. KRUraBI~IN,editor, Ann Arbor Science Publishers, Michigan, pp. 735-746.