Vol. 55, No. 5APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 1989, p. 1234-12410099-2240/89/051234-08$02.00/0Copyright © 1989, American Society for Microbiology

Microzonation of Denitrification Activity in Stream Sediments asStudied with a Combined Oxygen and Nitrous Oxide MicrosensorPETER BONDO CHRISTENSEN,* LARS PETER NIELSEN, NIELS PETER REVSBECH, AND JAN S0RENSEN

Department of Ecology and Genetics, University of Aarhus, Ny Munkegade, DK-8000 Arhus C, Denmark

Received 19 September 1988/Accepted 9 February 1989

Microzonation of denitrification was studied in stream sediments by a combined 02 and N20 microsensortechnique. 02 and N20 concentration profiles were recorded simultaneously in intact sediment cores in whichC2H2 was added to inhibit N20 reduction in denitrification. The N20 profiles were used to obtainhigh-resolution profiles of denitrification activity and N03- distribution in the sediments. 02 penetrated about1 mm into the dark-incubated sediments, and denitrification was largely restricted to a thin anoxic layerimmediately below that. With 115 ,uM N03- in the water phase, denitrification was limited to a narrow zonefrom 0.7 to 1.4 mm in depth, and total activity was 34 nmol of N cm2 h'. With 1,250 ,uM N03 in the water,

the denitrification zone was extended to a layer from 0.9 to 4.8 mm in depth, and total activity increased to 124nmol of N cm-2 h-1. Within most of the activity zone, denitrification was not dependent on the N03-concentration and the apparent Km for N03- was less than 10 ,uM. Denitrification was the only N03--consuming process in the dark-incubated stream sediment. Even in the presence of C2H2, a significant N20reduction (up to 30% of the total N20 production) occurred in the reduced, N03--free layers below thedenitrification zone. This effect must be corrected for during use of the conventional C2H2 inhibition technique.

Recent development of in situ techniques has demon-strated the importance of denitrification in nitrogen cyclingin both freshwater and marine sediments (8, 10, 11, 13, 14,21, 33). Among the assays introduced for measuring in situdenitrification in undisturbed sediment cores, the acetyleneinhibition technique appears to be the most convenientmethod (6, 24). By this method the C2H2, which inhibits thereduction of N20 to N2 in denitrifying bacteria (4, 41), isinjected directly into the sediment and accumulation of N20is taken as a measure of denitrification activity (34).For a better understanding of the distribution and regula-

tion of sediment denitrification, it is essential to develop newtechniques by which the process can be studied at a higherspatial resolution. Recent attempts have been made todetermine both N03 concentration profiles (12, 15) anddenitrification activity (19) by sectioning sediment cores in 1-to 5-mm-thick segments. The depth penetration of NO3 issometimes limited to a very narrow zone, however, andsectioning in the soft and flocculent surface layer may beimpossible.

In the study of 02 transformations in sediments, the 02microelectrodes have proven to be extremely useful (28).From extension of this work, a new microsensor has re-cently been developed by which 02 and N20 concentrationsin sediment, biofilm, etc., can be detected simultaneously ona microscale (30). After C2H2 is introduced, the developmentof N20 concentration profiles can be followed and denitrifi-cation can be quantified at a very high spatial resolution. Inthe present study, the microsensor was applied to studydenitrification activity in lowland stream sediments. Thedistribution of activity was measured together with 02 andN03 profiles on a microscale, and the results were com-pared with the activities that were determined by the con-ventional C2H2 inhibition technique.

* Corresponding author.

MATERIALS AND METHODS

Study sites. The study was carried out in three Danishlowland streams, Arhus A, Gudena, and D0de A, all ofwhich are located in the eastern part of Jutland, Denmark.Sediment cores were sampled in 10-cm-long and 2.4-cm-wide Plexiglas tubes during summer (June 1987), whenNO- concentrations in the stream water were about 100,uM (Arhus A), 600 p.M (Gudena), and 1,000 p.M (D0de A).The sediments were fine sand and silt, with an organiccontent of typically 3 to 5% (ignition loss) in the upper 2 cm.After sampling, the cores were stored overnight at roomtemperature under aerated stream water from the samplinglocality.02 and N20 concentration profiles measured by use of a

microsensor. To determine the microzonation of denitrifica-tion, a recently developed polarographic microsensor forsimultaneous 02 and N20 measurements was used (30). Inshort, the sensor has both a gold-plated and a silver-platedcathode behind a gas-permeable membrane of silicone rub-ber at the sensor tip (diameter, 20 Rm). By charging thecathodes against an Ag/AgCl anode (-1.04 and -1.18 V,respectively), the 02 diffusing into the sensor is reduced atthe gold cathode, while N20 is reduced at the silver cathode.The silver cathode is situated slightly behind the gold cath-ode, and 02 therefore does not interfere with N20 measure-ments. The 90% response time was 1.5 s for 02 and 12 s forN20, and the detection limit for both gases was about 1 ,uM.

Acetylene was added to the undisturbed sediment core byinjecting C2H2-saturated distilled water through small sili-cone-filled holes in the Plexiglas tube (10). Several injectionsof small portions ensured a homogeneous distribution ofC2H2 in the upper 2 cm of the sediment, and a finalconcentration of approximately 10 kPa in the pore water wasobtained. A total of 10% of the water phase in the cores wasreplaced by C2H2-saturated stream water, and the coreswere incubated in the dark at 22°C. The overlying water wasstirred by a small rotating magnet during the experiment.

1234

on May 27, 2021 by guest

http://aem.asm

.org/D

ownloaded from

MICROZONATION OF DENITRIFICATION 1235

The microsensor was lowered into the sediment with a

micromanipulator (Mertzheuser, Steindorf/Wetzlar, FederalRepublic of Germany), and changes in the 02 and N20profiles could be followed over time at exactly the same

point. After each determination of 02 and N20 profiles, thesensor was withdrawn from the sediment and calibrated inair-saturated water and in N20 standard solutions.

Denitrification and N03- concentration profiles determinedby modeling. Apparent diffusion coefficients for N20 andN03 in the sediment were estimated from recorded coeffi-cients of 02, assuming a similar ratio between the values as

between those in pure water (7, 26). The apparent diffusioncoefficient Of 02 was determined in HgCl2-inactivated sedi-ment from Arhus A by using an 02 microelectrode technique(27a). In the upper 1 mm of the sediment, the coefficient was80% of the value in pure water, and the porosity (vol/vol)was 0.8. In layers below 1 mm, these values were 70% and0.7, respectively.The measured N20 profiles were simulated by a diffusion-

reaction model (29) to determine the depth distribution ofN20 production, i.e., denitrifying activity. The model was

based on an extended version of Fick's second law ofdiffusion (5):

dC(x,t) a2C(X,t)=Ds(x)

aDs(x)ax

+ DS(x) )(X)2 .dC(x, )+ [P(x)-R(x)].

For a given depth x, C(x,t) is the concentration at time t;

D,(x) is the apparent diffusion coefficient; +~(x) is the poros-ity, and P(x) and R(x) are production and reduction rates,respectively. R(x) is 0 if N20 reduction is completely inhib-ited by C2H2. Rates of N20 production and reduction wereinserted until the modeled concentration profiles closelysimulated the measured ones.

Profiles ofNO3- concentration could be obtained from thedepth distribution of N20 production rates under the as-

sumptions that denitrification was the only N03--consumingprocess and that N03- production was 0 because of inhibi-tion of nitrification by C2H2 (16, 39). The N03- reductionrates were then twice the N20 production rates on a molarbasis, and the downward diffusion flux of N03- at eachdepth [F(x)] could be calculated as the sum of N03-consumption in all underlying layers. In accordance withFick's first law of diffusion:

F(x) = +(x) D,(x) (dC/ax); (2)

the change in N03- concentration with depth (dClax) is thenobtained from the flux divided by the apparent diffusioncoefficient for N03 and the porosity. All calculations were

relatively simple since the depth variation of N03- con-

sumption, apparent diffusion coefficient, and porosity was

described as constants for discrete depth intervals in thesediment. Assuming an N03 concentration of 0 at thebottom of the N20 production zone, the entire N03 con-

centration profile was eventually constructed by a cumula-tion of the dClax values from the bottom of the denitrificationzone to the water column. The NO3- profile thus gave a

simulated N03- concentration in the water phase whichcould be compared with the measured one as an independenttest of the validity of the profile.

Denitrification determined by the conventional C2H2 inhibi-tion technique. A separate experiment compared the deter-mination of denitrification by the microsensor technique and

by the conventional C2H2 inhibition technique with head-space extraction and gas chromatographic determination ofN20 (2, 10, 34). A set of five sediment cores with anoverlying water column of approximately 4 cm in height wassampled in Arhus A. Acetylene was added to both thesediment and the water phase as described above, andgas-tight Plexiglas caps were mounted on top of the tubes.Small magnetic bars placed underneath the caps were movedby using a large external magnet. After 10 min, watersamples (2 ml) were taken by syringe and frozen for lateranalysis of NO3- (initial concentration). The volumes re-moved were replaced by equivalent amounts of air, and theincubation was continued in darkness at 22°C. After 1.5 h ofincubation, the heights (volumes) of the water and sedimentphases were recorded. While the caps were kept on thetubes, water samples (2 ml) were taken for determination ofN03- (final concentration).The N20 accumulation in the water phase was determined

in a 21-ml water sample which was injected into a closed,preevacuated, 40-ml serum flask. The flask was shakenvigorously for 2 min to obtain equilibrium, and a 1-ml gassample was transferred to a preevacuated, 3-ml tube(Venoject; Terumo Europe N.V., Leuven, Belgium) for lateranalysis of N20. In two of the five cores, N20 concentrationprofiles were measured by use of the microsensor before thecores were sacrificed for the headspace extraction. The capswere removed from these cores immediately after the 21-mlwater sample was taken, and the sensor was carefullylowered into the sediment by the micromanipulator. Forboth cores, a set of three concentration profiles was re-corded before the tubes were capped again. The amount ofN20 that accumulated in the sediment plus the remainingfraction of the water phase was then determined in all fivecores by the headspace extraction technique; the entire tubewas shaken vigorously for 2 min before a 1-ml gas samplefrom the headspace (25 ml) was transferred to a Venojecttube.

All NO3- concentrations were determined colorimetri-cally in an autoanalyzer (Chemlab Instruments Ltd., Essex,England) by the method of Armstrong et al. (3). Gas sampleswere analyzed for N20 on a gas chromatograph (model 427;Packard Instrument Co., Inc., Rockville, Md.) equippedwith a 63Ni electron capture detector (3200C). The compo-nents were separated on a 80/100 mesh Porapak Q column(length, 2 m; width, 3.2 mm) at a temperature of 60°C withpure N2 used as the carrier gas (flow rate, 15 ml min-1). Inthe headspace analysis, a correction was made for dissolvedN20 by using a Bunsen solubility coefficient of 0.65 (40).Total denitrification activity, expressed in nanomoles ofnitrogen per square centimeter per hour, was estimated fromthe N20 accumulation in the cores, i.e., the sum of N20 inthe water sample and in the suspension of sediment andwater.

RESULTS

Depth distribution of denitrification activity. (i) Gudenawith 550 ,uM N03- in the water. In an experiment withsediment from Gudena and an N03 concentration in thewater phase of 550 FLM, modeling of the N20 concentrationprofiles showed that denitrification was distributed in a zoneabout 3.6 mm deep and starting immediately below thesediment surface (Fig. 1B). A good fit of the recorded N20concentration profiles was obtained throughout the incuba-tion period (Fig. 1A). Within the denitrification zone, the

VOL. 55, 1989

on May 27, 2021 by guest

http://aem.asm

.org/D

ownloaded from

1236 CHRISTENSEN ET AL.

N20 (pM) 02 and NO (uM)0 400 800 1200

0 200 400 600Denitrification (nmol N cm-3 h-1)

FIG. 1. (A) Measured (datum points) and simulated (solid lines) N20 profiles after various incubation periods following C2H2 addition toa sediment core from Gudenfa with 550 ,uM N03 in the water phase. (B) Depth distribution of denitrification activity as obtained by modelingof the time course of N20 accumulation. Measured 02 and simulated N03- profiles are indicated.

activity was almost constant and decreased only at thebottom. By accumulating all activities, a total denitrificationrate of 99 nmol of N cm-2 h-1 was calculated.The simulated N03- profile gave an NO3- concentration

of 556 ,uM in the water phase, which was in excellentagreement with the measured concentration of 550 piM (Fig.1B).

(ii) Arhus A with 115 ,uM N03 in the water. In C2H2-

treated Arhus A sediment, 02 and N20 concentration pro-files were measured at the in situ N03 concentration in thewater phase (Fig. 2A). After 110 min, there was no furtherchange in the N20 profile, indicating a steady-state conditionwhere N20 production equaled N20 diffusion loss plusconsumption. This steady-state profile was stable for morethan 2 h. Denitrification activity was estimated by modelingthe steady state (equation 1), and the N03 concentrationprofile was simulated; both are shown together with themeasured 02 profile in Fig. 2B. Activity was limited to anarrow zone from 0.7 to 1.4 mm in depth in the sediment.Oxygen penetrated to a depth of about 1 mm, and denitrifi-cation activity was thus only found when 02 was eitherpresent at a very low concentration (<10 ,uM) or totallyabsent (Fig. 2B). A vertical integration of the activity gavean overall denitrification rate of 34 nmol of N cm-2 h-'.The steady-state profile could be used to predict the time

course of N20 accumulation after C2H2 addition. It wasevident that shortly after C2H2 addition (7 and 24 min ofincubation; Fig. 2A), the measured N20 concentrations inthe sediment were considerably lower than those predictedby the model by using steady-state conditions.The steady-state N20 profile further revealed that even in

the presence of C2H2, a significant N20 reduction took placein the deeper layers below the denitrification zone. Thelinearity of the N20 profile below the production zone (from

1.4 to 3.5 mm in depth) indicated that the zone of N20reduction was separated in depth from the zone of produc-tion (Fig. 2A) and by simulation of the steady-state profile: areduction rate of 16 nmol of N20-N cm-3 h-1 [R(x) value ofequation 1] was obtained for sediments below 3.5 mm. Anoverall reduction rate of 10 nmol of N20-N cm-2 h-',corresponding to 29% of the production rate, was found.The simulated N03 profile gave an N03 concentration

of 117 puM in the water phase, which was identical to themeasured concentration of 115 ,uM. In the denitrificationzone, the availability of NO3 was lower because of thediffusion barrier of the oxic surface layer; the N03 concen-tration at the upper edge of the denitrifying zone was about30 p,M (Fig. 2B).

(iii) Arhus A with 1,250 ,uM N03- in the water. In aseparate core from Arhus A, the water column was amendedwith N03 to a final concentration of 1,250 ,uM, which wasa 10-fold increase compared with in situ conditions. The corewas preincubated for 5 h at room temperature before theC2H2 was added and the 02 and N20 profiles were recorded.Compared with the sediment with in situ N03- concen-

trations (Fig. 2), the zone of denitrification activity wasmarkedly deeper in the N03(-amended sediment; denitrifi-cation was thus observed from 0.9 to 4.8 mm below thesediment surface (Fig. 3B). Also in this sediment, theactivity was relatively constant throughout most of thedenitrification zone, and the rates (per unit volume ofsediment) were not significantly different from those ob-tained at the in situ N03 concentration. Accumulation ofthe whole activity profile gave an overall denitrification rateof 124 nmol of N cm-2 h-1, which was 3.5-fold higher thanthat at the in situ N03- concentration.No discrepancy between measured and simulated N20

concentration profiles was observed in the early phase of the

-C

CD

° °2 ) N03O o.0 .

...p.I,

B

APPL. ENVIRON. MICROBIOL.

on May 27, 2021 by guest

http://aem.asm

.org/D

ownloaded from

MICROZONATION OF DENITRIFICATION 1237

N20 (p1M) 02 and N0O (M)0 400 800 12

NO3| 6 OLz

I-m0 : 02 .

!00

0 200 400 600Denitrification (nmol N cm-3 h-I)

FIG. 2. (A) Measured (datum points) and simulated (solid lines) N,O profiles after various incubation periods following C2H2 addition to

a sediment core from Arhus A with 115 ,uM NO3- in the water phase. The measured values after 7 and 24 min of incubation are connectedby a dotted line to indicate the discrepancy from simulated N,O profiles. (B) Depth distribution of denitrification activity obtained by modelingof the steady-state N,O concentration profile (110 min of incubation). Measured 02 and simulated N03 profiles are indicated.

incubation with C2H2 (Fig. 3A), but the simulations revealeda significant N20 reduction in the deeper part of the sedi-ment. Thus, for sediment below 5 mm in depth, a reductionrate of 22 nmol of N cm-3 h-' was used in the model [R(x)value of equation 1] to give a perfect simulation of the N2Oconcentration profiles. Steady-state conditions were notobtained during the incubation period of 55 min; until a

steady state was reached, the overall N2O reduction in-creased along with the extension in depth of the N20 profile.An overall reduction rate of 8 nmol of N20-N cm-2 h-V,corresponding to 7% of the total N20 production, was

calculated for 55 min of incubation.The simulated N03- profile gave a concentration of 845

F.M at the upper edge of the denitrification zone and a

concentration of 1,170 F.M in the overlying water phase (Fig.3B); the latter was very close to the measured concentrationof 1,250 ,uM.Comparison of denitrification activity measured by the

microsensor and by the conventional C2H2 inhibition tech-niques. To compare determination of N2O accumulation bythe microsensor and by the conventional C,H, inhibitiontechniques (headspace extraction followed by gas chromato-graphic detection of N20), both techniques were applied tosediment cores from Arhus A with the in situ N03 concen-tration of 110 ,uM in the overlying water. By the conven-

tional technique, N,O accumulation was determined byheadspace extraction in five parallel cores after 1.5 h ofincubation with C,H2. Since nitrification was inhibited byC2H2, total N03- consumption could be estimated from thedecreasing N03- concentration in the water phase duringthe denitrification assay. Mean activity (+standard error ofthe mean) for total N03 consumption and denitrification inthe five cores were 62 + 2 and 53 + 3 nmol of N cm 2 h-1,

respectively, and N20 accumulation thus accounted for 85%of the total NO3 consumption.

Nitrous oxide concentration profiles were measured bythe microsensor technique in two of the five cores just beforethe headspace extraction (Fig. 4A and B). The profiles gave

an indication of the heterogeneity in spatial distribution ofthe denitrification. While one of the cores from Arhus A onlyshowed little variation in depth distribution and total activity(Fig. 4A), the other one had N20 maxima at different depthsand showed a significant variation in the recorded accumu-lation of N20 (Fig. 4B). On the basis of six N20 concentra-tion profiles measured in two cores, the accumulated N20content made up only 64% of the N20 content measured bythe conventional C2H2 inhibition technique. For compari-son, N2O concentration profiles were also measured in a

C2H2-treated sediment core from D0de A (after 1.5 h ofincubation) with an N03- concentration of 1,000 ,uM in thewater phase (Fig. 4C). At this locality, the sediment surfacewas visually heterogeneous because of faunal activity, andthe recorded N,O profiles were highly variable within thesame sediment core. Thus, one of the profiles showed a veryirregular shape, with no distinct maximum of N20 accumu-

lation, and for the three recorded profiles, a threefold or

greater variation was found for the total amount of N20accumulated.

DISCUSSION

Depth distribution of denitrification activity. The combined02 and N,O microelectrode has made it possible for the firsttime to demonstrate and study the microzonation of denitri-fication in sediments. In the lowland streams investigatedhere, we localized denitrification activity in a narrow zone

2 Denit.

3

4

5-

6 -

7 - BQ

VOL. 55, 1989

on May 27, 2021 by guest

http://aem.asm

.org/D

ownloaded from

1238 CHRISTENSEN ET AL.

N20 (IM)0

02 and N0O (AM)400 800 1200

0 200 400 600Denitrification (nmol N cm-3 h-1)

FIG. 3. (A) Measured (datum points) and simulated (solid lines) N20 profiles after various incubation periods following C2H2 addition toa sediment core from Arhus A with 1,250 ,uM N03 in the water phase. (B) Depth distribution of denitrification activity obtained by modelingof the time course of N20 accumulation. Measured 02 and simulated N03- profiles are indicated.

within the upper few millimeters of the sediment and dem-onstrated that the activity zone was primarily determined bythe depth penetrations of 02 and N03 .

(i) Denitrification activity in relation to 02. Denitrificationat low 02 tension has previously been described for bothmarine sediments and soils (18, 38), but truly aerobic deni-trification such as that described for a denitrifying mixotroph

N20 (pM)n0 10 20 30

.4-iC.0L)03

40 50

(32) was never observed in the investigated sediments. Theresults indicated that there was a small amount of activityjust above the oxic/anoxic interface (02 concentrations from0 to 20 ,uM), but the major fraction of the denitrification wasfound in completely anoxic layers. The oxic surface layertherefore constitutes a barrier for N03 diffusion from thewater phase to the denitrification zone, and the thickness of

FIG. 4. N20 concentration profiles in sediment from Arhus A (A and B) and D0de A (C). The three lines in each panel represent different

N,0 profiles obtained in the same core. The N03- concentrations of the water phase were 115 ,uM (A and B) and 1,000 ,uM (C). All profiles

were recorded after 1.5 h of incubation with C,H,. Note the change of scale in panel C.

E

m-i0.0L)0:

°2 N03O o /-

i /0

'''L... . ,

4. Dei. .... ._

5... ..

7. B

8s.:

X-C

8IC /4

10

o ,x

B

APPL. ENVIRON. MICROBIOL.

on May 27, 2021 by guest

http://aem.asm

.org/D

ownloaded from

MICROZONATION OF DENITRIFICATION 1239

the barrier may be a major controlling factor for denitrifica-tion in stream sediments (L. P. Nielsen, P. B. Christensen,N. P. Revsbech, and J. S0rensen, manuscript in prepara-

tion).(ii) Denitrification activity in relation to N03 concentra-

tion. All simulations of N03 profiles gave N03 concen-

trations in the water phase that were almost identical tothose actually measured, indicating that denitrification in-deed was the only N03-consuming process in the sediment.The close agreement between measured and simulatedN03 concentrations also indicated that the apparent diffu-sion coefficient for N03- in the sediment was correctlyestimated, assuming a ratio between diffusion coefficients of02, N20, and N03 in the sediment similar to that in purewater. A specific reduction of N03- diffusion because ofelectrical interactions between the NO3- ion and chargedsurfaces, as observed in trickling filter matrices (31), was

apparently unimportant in the sediments since the estimatedN03 profile otherwise would have underestimated theN03 concentration of the water phase. We therefore con-

cluded that the estimated N03 profiles described the realN03 concentration profiles in the sediments.

Denitrification was limited to a narrow zone of 0.7 mm indepth in the Arhus A sediment when N03- was present at a

relatively low concentration (110 ,uM) in the water phase(Fig. 2B). At a higher N03 concentration (1,250 ,uM), thedenitrification zone was extended to a 4-mm-deep zone (Fig.3A). The activities recorded in the denitrification zone were

not stimulated by higher N03 availability per se. Thus,even if the N03 concentration was increased from 30 to 850,uM at the upper edge at the activity zone in the Arhus Asediment, the specific activities within the zone were stillcomparable (Fig. 2B and 3B). Furthermore, the activityseemed to be constant throughout the denitrification zone inthe sediments from both Gudena (Fig. iB) and Arhus A (Fig.3B). Only at the lower edge of the activity zone, whereNO3 dropped to a concentration below 10p,M, the activitydecreased significantly. This indicated a low apparent Kmvalue (<10 pM) for the reduction of N03 by denitrificationin these sediments.The depth distribution of denitrification activity has pre-

viously been illustrated by vertical sectioning of C2H2-incubated sediment cores (2, 11, 19). From data obtainedwith the microsensor, it was possible to define the spatialmicrozonation of denitrification accurately and in muchgreater detail than before. Denitrification was never re-

corded at depths below 5 mm in the sediment, and severalearlier observations of denitrification activity below a 1-cmdepth in such sediments may therefore be erroneous becauseof the downward diffusion of N20 during incubation. Excep-tionally deep zones of denitrification can only be foundwhere bioturbation by sediment fauna, macrophyte rootactivity, or upwelling of NO3-rich groundwater provides an

additional N03 source in the deeper layers (1, 10, 11).Combination of the microsensor and conventional C2H2

inhibition techniques. Both the conventional C2H2 inhibitiontechnique and the microsensor technique have advantagesand disadvantages, when used alone. The former techniquemay be preferable when determining an overall activity insediments, while the latter should be used when detailedstudies of the denitrification process are required. However,by combining the two techniques, even more information onthe denitrification can be obtained and a better estimate ofthe activity can be given.

(i) Denitrification in relation to sediment heterogeneity. Byusing the microsensor, an in situ activity of 34 nmol of N

cm-' h-' was obtained in the Arhus A sediment (Fig. 1).This activity was significantly lower than both the N03consumption rates (62 nmol of N cm-2 h-1) and the N20accumulation rates (53 nmol ofN cm-2 h-1) obtained by theconventional C2H2 technique. When the microsensor wasapplied, however, activity was recorded within a very smallvolume of sediment, and so several microprofiles must bemeasured to compensate for heterogeneity (Fig. 4A and B).The variability of N20 profiles can probably be explained bythe presence of faunal burrows that create microsites with ahigher activity than that in the bulk sediment (1, 9, 25). Sucha faunal activity was confirmed from visual inspection of thecores from D0de A and seemed to be responsible for thevery heterogeneous distribution of denitrification in thissediment (Fig. 4C).

(ii) Disturbance of in situ N03 proffle by C2H2 injection. Adistortion of activity caused by the injection of C2H2 wasobserved in the Arhus A sediment (110 ,uM N03- in thewater phase). Marked disagreement between measured andsimulated N20 profiles was demonstrated in the early phaseof incubation; the two first N20 profiles recorded at 7 and 24min after C2H2 addition were located distinctly higher in thecore than the following ones, and the accumulated amount ofN2O was significantly lower than expected (Fig. 2A). WhenC2H2-saturated, distilled water was injected into this sedi-ment, the pore water NO3- profile, which only extendedabout 1.5 mm into the sediment (Fig. 2B), was apparentlydisplaced upwards and denitrification activity was impededbecause of lower N03 availability until a steady-stateNO3 profile was reestablished. The original NO3 profilewas eventually recovered after 50 min of incubation (datanot shown). This distortion is most pronounced in sedimentsin which denitrification activity is located immediately belowthe sediment-water interface. In such cases, the addition ofC2H2 to the water phase alone may be the preferred choice,since rapid diffusion of C2H2 to the denitrification zoneshould allow measurements of N20 accumulation from un-disturbed concentration gradients. By comparison, the injec-tion of C2H2-saturated water into Gudena sediment andNO3--enriched Arhus A sediment showed no significantdisplacement of the N2O profiles since NO3- here extendedto a depth of about 5 mm (Fig. 1B and 3B).

(iii) Incomplete inhibition of N20 reduction by C2H2. By themicrosensor technique, it was also possible to demonstrate asignificant N20 reduction in the C2H2-treated sediment. Inthe Arhus A sediment with an in situ N03- concentration inthe water phase, N20 reduction made up about 30% of thetotal N20 production under steady-state conditions; a simi-lar value has been observed in other stream sediments(Nielsen et al., in preparation). There have been severalreports on low efficacy of the C2H2 inhibition at low N03concentrations (20, 22, 27). In the sediments used in thisstudy, however, the close agreement between the predictedand the measured N03 concentrations in the water phaseexcluded the possibility that significant N20 reduction couldtake place within the production (denitrification) zone. Also,the N20 profiles demonstrated that the production andreduction zones of N20 were separated in depth (Fig. 2A),and we concluded that N20 reduction occurred only in thedeeper, NO3 -free layers. The incomplete blockage of N20reduction was therefore most likely caused by the presenceof sulfide, which alleviates the C2H2 inhibition even at lowconcentrations (36, 37).The N20 reduction in the presence of C2H2 represents a

source of error in the estimate of total denitrification by theconventional C2H2 inhibition technique. However, the N20

VOL. 55, 1989

on May 27, 2021 by guest

http://aem.asm

.org/D

ownloaded from

1240 CHRISTENSEN ET AL.

consumption can be quantified by the microsensor techniqueand a correction factor can be determined. Because of theN20 reduction in the presence of C2H2, a difference betweenNO3- consumption and N20 accumulation may not directlyindicate or qDuantify an alternative pathway of NO3- reduc-tion. In the Arhus A sediment, for instance, the lower ratesofN20 accumulation relative to the total N03 consumptionrate must thus be explained by the occurrence of N20reduction in the presence of C2H2; rates of NO3- consump-tion in this study were a better estimate of total denitrifica-tion activity than the N20 accumulation rates were. Asdiscussed above, the simulations of N03- profiles alsoindicated that denitrification is the only NO03-consumingprocess in the sediments studied here. The nitrate ammoni-fication (dissimilatory NO3 reduction to NH4+) describedin marine sediments (17, 23, 24, 35) therefore did not seem tobe important in the streams that we investigated.

In conclusion, the combined 02-N20 microsensor wasvaluable in the study of microzonation of the denitrificationprocess, and its use in combination with the conventionalC2H2 inhibition technique and flux measurements maygreatly improve the study of denitrification and other NO3 --consuming processes in sediments.

ACKNOWLEDGMENTSThe skillful technical assistance of Dorthe Olsson is gratefully

acknowledged.This work was supported by the National Agency for Environ-

mental Protection under the Danish Ministry of the Environment.

LITERATURE CITED

1. Aller, R. C. 1988. Benthic fauna and biochemical processes inmarine sediments: the role of burrow structures, p. 300-388. InT. H. Blackburn and J. S0rensen (ed.), Nitrogen cycling incoastal marine environments. John Wiley & Sons Ltd., Chich-ester, England.

2. Andersen, T. K., M. H. Jensen, and J. S0rensen. 1984. Diurnalvariation of nitrogen cycling in coastal, marine sediments, I.Denitrification. Mar. Biol. 83:171-176.

3. Armstrong, F. A., C. R. Stearns, and J. D. Strickland. 1967. Themeasurement of upwelling and subsequent biological processesby means of the Technicon autoanalyzer and associated equip-ment. Deep-Sea Res. 14:381-389.

4. Balderston, W. L., B. Sherr, and W. J. Payne. 1976. Blockage byacetylene of nitrous oxide reduction in Pseudomonas perfecto-marinus. Appl. Environ. Microbiol. 31:504-508.

5. Berner, R. A. 1980. Early diagenesis. A theoretical approach.Princeton University Press, Princeton, N.J.

6. Blackburn, T. H. 1986. Nitrogen cycle in marine sediments.Ophelia 26:65-76.

7. Broecker, W. S., and T.-H. Peng. 1974. Gas exchange ratesbetween sea and air. Tellus 26:21-35.

8. Chan, Y.-K., and N. E. R. Campbell. 1980. Denitrification inlake 227 during summer stratification. Can. J. Fish. Aquat. Sci.37:506-512.

9. Chatarpaul, L., J. B. Robinson, and N. K. Kaushik. 1980.Effects of tubificid worms on denitrification and nitrification instream sediment. Can. J. Fish. Aquat. Sci. 37:656-663.

10. Christensen, P. B., and J. S0rensen. 1986. Temporal variation ofdenitrification activity in plant-covered littoral sediment fromLake Hampen, Denmark. Appl. Environ. Microbiol. 51:1174-1179.

11. Christensen, P. B., and J. S0rensen. 1988. Denitrification insediment of lowland streams: regional and seasonal variation inGelbaek and Rabis Baek, Denmark. FEMS Microbiol. Ecol.53:335-344.

12. Cooke, J. G., and R. E. White. 1988. Nitrate enhancement ofnitrification depth in sediment/water microcosms. Environ.Geol. Water Sci. 11:85-94.

13. Cooper, A. B., and J. G. Cooke. 1984. Nitrate loss and trans-formation in 2 vegetated headwater streams. N.Z. J. Mar.Freshwater Res. 18:441-450.

14. Gardner, W. S., T. F. Nalepa, and J. M. Malczyk. 1987.Nitrogen mineralization and denitrification in Lake Michigansediments. Limnol. Oceanogr. 32:1226-1238.

15. Hordijk, C. A., M. Snieder, J. J. M. van Engelen, and T. E.Cappenberg. 1987. Estimation of bacterial nitrate reduction at insitu concentration in freshwater sediments. Appl. Environ.Microbiol. 53:217-223.

16. Hynes, R. K., and R. Knowles. 1982. Effect of acetylene onautotrophic and heterotrophic nitrification. Can. J. Microbiol.28:334-340.

17. J0rgensen, B. B., and J. S0rensen. 1985. Seasonal cycles Of 02,NO3- and s042- reduction in estuarine sediments: the signifi-cance of an N03- reduction maximum in spring. Mar. Ecol.Prog. Ser. 24:65-74.

18. J0rgensen, K. S., H. B. Jensen, and J. S0rensen. 1984. Nitrousoxide production from nitrification and denitrification in marinesediment at low oxygen concentrations. Can. J. Microbiol.30:1073-1078.

19. J0rgensen, K. S., and J. S0rensen. 1988. Two annual maxima ofnitrate reduction and denitrification in estuarine sediment(Norsminde Fjord, Denmark). Mar. Ecol. Prog. Ser. 48:147-154.

20. Kaspar, H. F. 1982. Denitrification in marine sediment: mea-surement of capacity and estimate of in situ rate. Appl. Environ.Microbiol. 43:522-527.

21. Kaspar, H. F., R. A. Asher, and I. C. Boyer. 1985. Microbialnitrogen transformations in sediments and inorganic nitrogenfluxes across the sediment/water interface on the South Islandwest coast, New Zealand. Estuar. Coast. Shelf Sci. 21:245-255.

22. Kaspar, H. F., J. M. Tiedje, and R. B. Firestone. 1981. Denitri-fication and dissimilatory nitrate reduction to ammonium bydigested sludge. Can. J. Microbiol. 38:486-493.

23. Koike, I., and A. Hattori. 1978. Denitrification and ammoniaformation in anaerobic coastal sediments. Appl. Environ. Mi-crobiol. 35:278-282.

24. Koike, I., and J. S0rensen. 1988. Nitrate reduction and denitri-fication in marine sediments, p. 251-273. In T. H. Blackburnand J. S0rensen (ed.), Nitrogen cycling in coastal marineenvironments. John Wiley & Sons Ltd., Chichester, England.

25. Kristensen, E., M. H. Jensen, and T. K. Andersen. 1985. Theimpact of polychate (Nereis virens Sars) borrows on nitrificationand nitrate reduction in estuarine sediments. J. Exp. Mar. Biol.Ecol. 85:75-91.

26. Li, Y.-H., and G. Gregory. 1974. Diffusion of ions in sea waterand in deep-sea sediments. Geochim. Cosmochim. Acta 38:703-714.

27. Oremland, R. S., C. Umberger, C. W. Culbertson, and R. L.Smith. 1984. Denitrification in San Francisco Bay intertidalsediments. Appl. Environ. Microbiol. 47:1106-1112.

27a.Revsbech, N. P. 1989. Diffusion characteristics of microbialcommunities determined by use of oxygen microsensors. J.Microbiol. Methods 9:111-122.

28. Revsbech, N. P., and B. B. J0rgensen. 1986. Microelectrodes:their use in microbial ecology. Adv. Microb. Ecol. 9:293-352.

29. Revsbech, N. P., B. Madsen, and B. B. J0rgensen. 1986. Oxygenproduction and consumption in sediments determined at highspatial resolution by computer simulation of oxygen microelec-trode data. Limnol. Oceanogr. 31:293-304.

30. Revsbech, N. P., L. P. Nielsen, P. B. Christensen, and J.S0rensen. 1988. A combined oxygen and nitrous oxide mi-crosensor for studies of denitrification. Appl. Environ. Micro-biol. 54:2245-2249.

31. Riemer, M., and P. Harremoes. 1978. Multi-component diffusionin denitrifying biofilms. Prog. Water Technol. 10:149-165.

32. Robertson, L. A., and J. G. Kuenen. 1984. Aerobic denitrifica-tion: a controversy revived. Arch. Microbiol. 139:351-354.

33. Seitzinger, S. P. 1987. Nitrogen biogeochemistry in an unpol-luted estuary: the importance on benthic denitrification. Mar.Ecol. Prog. Ser. 41:177-186.

APPL. ENVIRON. MICROBIOL.

on May 27, 2021 by guest

http://aem.asm

.org/D

ownloaded from

MICROZONATION OF DENITRIFICATION

34. S0rensen, J. 1978. Denitrification rates in marine sediments asmeasured by the acetylene inhibition technique. Appl. Environ.Microbiol. 36:139-143.

35. S0rensen, J. 1987. Nitrate reduction in marine sediment: path-ways and interactions with iron and sulfur cycling. Geomicro-biol. J. 5:401-421.

36. S0rensen, J., L. K. Rasmussen, and I. Koike. 1987. Micromolarsulfide concentrations alleviate acetylene blockage of nitrousoxide reduction by denitrifying Pseudomonasfluorescens. Can.J. Microbiol. 33:1001-1005.

37. Tam, T.-Y., and R. Knowles. 1979. Effects of sulfide and

acetylene on nitrous oxide reduction by soil and by Pseudomo-nas aeruginosa. Can. J. Microbiol. 25:1133-1138.

38. Trevors, J. T. 1985. The influence of oxygen concentrations ondenitrification in soil. Appl. Microbiol. Biotechnol. 23:152-155.

39. Walter, K. M., D. R. Keeney, and I. R. Fillery. 1979. Inhibitionof nitrification by acetylene. Soil. Sci. Am. J. 43:195-196.

40. Weiss, R. F., and B. A. Price. 1980. Nitrous oxide solubility inwater and seawater. Mar. Chem. 8:347-359.

41. Yoshinari, Y., and R. Knowles. 1976. Acetylene inhibition ofnitrous oxide reduction by denitrifying bacteria. Biochem.Biophys. Res. Commun. 69:705-710.

VOL. 55, 1989 1241

on May 27, 2021 by guest

http://aem.asm

.org/D

ownloaded from