Ammonium Concentrations in Produced Waters from a Mesothermic Oil Field Subjected to Nitrate Injection Decrease through Formation of Denitrifying Biomass and Anammox Activity

Materials.Reagent-grade chemicals were obtained from either BDH, Fisher, or Sigma. Primers were purchased from the University of Calgary Core DNA Services. Deoxynucleoside triphosphates (dNTPs) and SYBR green I were purchased from Invitrogen. Other PCR reagents were obtained from Qiagen. Compressed gases used in the anaerobic hood were obtained from Praxair.

Source of samples.A total of 23 field sites, including 3 WPs (1-WP, 17-WP, and 20-WP), 3 MWs (21-MW, 22-MW, and 23-MW), 2 IWs (8-IW and 14-IW), and 15 PWs (those listed below, as well as 6-PW, 16-PW, 18-PW, and 19-PW) were sampled. A map showing their locations has been presented elsewhere (39). Samples were collected in 1-liter Nalgene bottles filled to the brim to exclude air. PW samples were collected at the wellheads where the produced oil-water mixture came up from the subsurface, whereas IW samples were collected where injection water was injected into the subsurface (Fig. 1). WP and MW samples were collected where the make-up water and injection water stream left the MW plant or the WP, respectively (Fig. 1). The samples were transported to the laboratory and placed in an anaerobic hood with an atmosphere of 5% (vol/vol) H₂, 10% CO₂, and balance N₂ within 5 h of collection. One set of samples was collected just prior to the start of field-wide nitrate injection. Samples for analysis by DGGE were obtained only in the first year of nitrate injection.

Enrichment cultures.Produced or injection water samples (5 ml) were inoculated within 24 h into 120-ml serum bottles containing 95 ml CSB-K medium (20) with either 10 mM nitrate and 5 mM sulfide for enrichment of soNRB or 3 mM volatile fatty acids (VFA) (3 mM acetate, 3 mM propionate, and 3 mM butyrate) and 10 mM nitrate for enrichment of hNRB. Samples were incubated at 30°C. Metabolic activity of soNRB was monitored by measuring the sulfate and/or sulfide concentration, and that of hNRB was determined by measuring the nitrate and/or nitrite concentration as a function of time. Typical time courses have been presented elsewhere (11). Microbial activity in dimensionless units/day was calculated as 100/t_1/2, where t_1/2 is the time in days needed to reduce half of the added electron acceptor when using a 5% (vol/vol) inoculum (39). Average activities for a given sampling date were calculated for 11 PWs (2-PW, 3-PW, 4-PW, 5-PW, 7-PW, 9-PW, 10-PW, 11-PW, 12-PW, 13-PW, and 15-PW), for which data were collected continuously for 125 weeks from 7 May 2007 onward. Once maximal growth was attained, the enrichment cultures were also used for biomass and DNA isolation.

Isolation of DNA.DNA was isolated from a subset of samples (8-IW, 3-PW, 4-PW, 7-PW, 12-PW, 13-PW, 15-PW, and 16-PW). Enrichment cultures were centrifuged in 30-ml aliquots for 20 min at 4°C and at 12,100 × g in 50-ml Nalgene centrifuge tubes. The combined cell pellets for each enrichment were resuspended in 1 ml of 0.15 M NaCl-0.1 M EDTA, pH 8 (NaCl-EDTA), and stored at −20°C until used for DNA extraction. Produced water from field samples was separated from produced oil by placing the 1-liter Nalgene sample bottles in a ring stand, piercing the bottom three times with an 18-gauge needle, and allowing the water phase to drain into 250-ml centrifuge bottles cleaned with Alkanox laboratory detergent and rinsed with deionized water and ethanol. Collected water samples were then centrifuged for 20 min at 4°C at 12,100 × g. Pelleted solids were resuspended in 1 ml of supernatant, transferred to a 15-ml Corex centrifuge tube, and centrifuged for an additional 10 min at 4°C, 12,100 × g. Pellets were then resuspended in 0.5 to 1 ml of NaCl-EDTA and stored at −20°C. DNA was extracted using a protocol described elsewhere (40) but modified to include one or more freeze-thaw steps. DNA was dissolved in 50 to 200 μl of 10 mM Tris-0.1 mM EDTA, pH 8 (TE).

PCR of rRNA genes.Extracted DNA (2 μl of 2.5 to 25 ng/μl) was added to an amplification mixture, which included 20 pmol (each) of primers 27f-GC and 534r, 25 μl of 2× master mix (Fermentas) containing Taq DNA polymerase and dNTPs, and water to 50 μl. A description of all of the primers used is provided in Table 1. PCR incubation was carried out for 5 min at 94°C, followed by 25 cycles of 1 min at 94°C, 1.5 min at 60°C, and 1 min at 72°C. PCR products were run on 0.7% (wt/vol) agarose gels and visualized using ethidium bromide staining and UV light to ensure that only one band of the correct size was present. Archaeal rRNA genes were amplified similarly with use of the primers U515f-GC and A1041r (Table 1).

DGGE.DGGE was carried out with a modified Protean II xi cell (Bio-Rad) at 60°C, as described elsewhere (11, 24). Approximately 300 ng of PCR-amplified DNA was loaded onto a 6.5% (wt/vol) acrylamide gel with a gradient of 20 to 60% denaturant, with 100% denaturant being 40% (vol/vol) formamide and 7 M urea. The gels were run at 75 V, 60°C, for 16 h. Denaturing gels were then stained with 1× SYBR green I (Invitrogen), and the bands were visualized by exposure to UV light.

DNA sequencing.All visible bands were cut and transferred into sterile microcentrifuge tubes. TE (50 μl) was added to extract the DNA. Samples were then centrifuged, and 2 μl of the supernatant was added to PCR mixtures containing the primer pair 27f and 534r or U515f and A1041r. PCR products were purified using the QIAquick PCR purification kit (Qiagen laboratories) and sequenced at the University of Calgary DNA Laboratory using a 48-capillary ABI 3730S automated DNA sequencer. The determined sequences were edited using the software program Sequence Scanner (Applied Biosystems). Sequences obtained with both the forward and reverse primers were assembled using the Staden software program GAP4. Edited and assembled sequences were compared to those in the GenBank database using BLASTN.

Detection of anammox bacteria.Extracted DNAs were further analyzed by amplification with the primers Pla46F and 1492r and then with the nested primers amx368F and amx820r. The latter pair has been shown to amplify freshwater anammox bacteria of the genera “Candidatus Brocadia,” “Candidatus Kuenenia,” and “Candidatus Jettenia” (18, 30, 31). The PCR products were cloned into the Promega pGEM-t-Easy vector, and the clones obtained were analyzed by sequencing. The sequences obtained were compared with those of known anammox bacteria and were assembled into phylogenetic trees using the software program MEGA (35).

Analytical chemistry.Ammonium concentrations were determined spectrophotometrically with the indophenol method (1) for water samples obtained from all 23 sampling sites. Volume-weighted average ammonium concentrations for 11 PWs were calculated from the determined concentrations c_i and produced water volumes V_i for wells i (i = 1 to 11) as Σc_iV_i/ΣV_i. Concentrations of nitrate were determined by high-pressure liquid chromatography, as described elsewhere (39). Headspace methane was determined by gas chromatography. Protein concentrations were determined with the Bio-Rad DC protein assay kit, using procedures outlined by the manufacturer.

Nucleotide sequence accession numbers.Sequences for rRNA genes isolated from oil field waters and enrichment cultures have been deposited in GenBank under accession numbers FN548170 to FN548363 and FN550417 to FN550802. Sequences obtained for anammox bacteria have been deposited in GenBank under accession numbers GU011974 to GU011984.

Effect of nitrate injection on NRB activity.Nitrate injection in the Enermark field involved the following: (i) continuous, field-wide injection of 2.4 mM from week 1 to week 125, (ii) batchwise application of high concentrations (760 mM; 1 h/week) at 14-IW from week 33 to week 101, and (iii) field-wide nitrate pulses of 2.4 or 14 mM, alternating weekly, from week 64 to week 96. Field-wide applications i and iii involved injection at the WPs (Fig. 1). Prior to the start of nitrate injection, the average activities of hNRB and soNRB in incubations of 11 PWs were 50.6 (±22.7) and 25.8 (±16.5) units/day, respectively. These averages tended to increase somewhat from weeks 1 to 74 of nitrate injection (Fig. 2). However, it is clear from the data presented in Fig. 2 that increases in activity were modest and that significant activities of both hNRB and soNRB were already present prior to the start of the nitrate injection program.

• Open in new tab
• Download powerpoint

FIG. 2.

Average activity of hNRB (A) and soNRB (B) in produced waters from the Enermark field as a function of time. The averages are for 11 PWs. The standard deviation indicates the distribution of values observed at each time point. The line drawn is a moving average for two time points.

DGGE analysis of bacteria.DNA was extracted from 90 different field samples, collected from the Enermark field from week 0 to 52, and from 140 incubations of these samples, which were done to determine hNRB and soNRB activity. We will refer to these as hNRB and soNRB enrichments. The DNAs obtained were PCR amplified with bacterial primers (all samples) and archaeal primers (field samples only, since no PCR product was obtained from NRB enrichments with archaeal primers). A DGGE gel for amplicons, obtained with bacterial primers, is shown in Fig. 3. All discernible bands were excised, reamplified with the primers 27f and 534r, and sequenced. Of a total of 570 processed bands, 396 (70%) gave good-quality sequences. The closest GenBank matches of all of these are provided in Table S1 in the supplemental material. A survey of their phylogenetic group and genus classification is provided in Table 2 .

• Open in new tab
• Download powerpoint

FIG. 3.

DGGE of bacterial 16S amplicons obtained from microbial community DNAs isolated from the Enermark field. (b to i) DNAs from produced waters collected on 15 May 2007. The number of the production well is indicated at the bottom. (j to o) DNAs from hNRB enrichments grown from produced waters collected on 19 December 2007. (p to x) DNAs from soNRB enrichments grown from produced waters collected on 19 December 2007 (p to r) or 10 July 2007 (s to x). Lane a is a mixture of amplicons from markers. Only the band for Thauera sp. strain N2 is visible. All bands tagged with a number were isolated from the gel and sequenced.

A total of 51 different bacterial genera were identified in 7 higher-level phylogenetic groups (class, phylum, family, or otherwise), namely, the alpha-, beta-, gamma-, delta-, and epsilonproteobacteria, the firmicutes, and the others (Table 2). The latter included the Cytophaga, Flavobacterium, Bacteroides (CFB) group, the spirochetes, the planctomycetes, the thermotogales, the green non-sulfur (GNS) bacteria, the actinobacteria, and bacteria from the genera Geovibrio, Denitrovibrio, and Thermovirga. All phylogenetic groups, except the gammaproteobacteria, were found in the field samples. Relative to field samples, enrichment of hNRB on medium containing VFA and nitrate decreased the frequency of observations of deltaproteobacteria, which includes the SRB (Table 2, 18.7% versus 1.2%), while increasing that for the beta- and gammaproteobacteria (Table 2, 4% versus 19.9% and 0% versus 11.4%, respectively). The latter groups included Thauera and Pseudomonas spp., which are known nitrate-reducing heterotrophs. Relative to field samples, enrichment for soNRB boosted the frequency of observation of epsilonproteobacteria (Table 2, 35.3% versus 76.1%) at the expense of all other groups, except the betaproteobacteria (Table 2, 4% versus 14.1%). This is in agreement with the highly specific nature of this medium, which initially allows only the growth of sulfide-oxidizing, nitrate-reducing autotrophs. The low fraction of observed deltaproteobacteria in hNRB and soNRB enrichments (1.2 and 4.2%, respectively) compared to that for the field samples (18.7%) indicates that SRB do not do well in enrichment media for nitrate reducers.

Effect of nitrate on bacterial community composition at 13-PW.No nitrate breakthrough was observed in any of the PWs during field-wide application of a constant dose of 2.4 mM nitrate (39). Application of high nitrate batches (760 mM; 1 h/week) at 14-IW gave nitrate, nitrite, and sulfate breakthrough at neighboring 13-PW while lowering sulfide concentrations to zero from week 41 onwards (39). DGGE analysis results for field samples collected from 13-PW prior to (18 bands) and following (32 bands) nitrate breakthrough indicated an increased frequency of observation of epsilonproteobacteria (representing mainly soNRB) from 7 to 33%, respectively, and a decreased frequency of observation of deltaproteobacteria from 33 to 17%, respectively; alpha- and betaproteobacteria (11 and 3%, respectively) were detected only following nitrate breakthrough. This reflects the differences in community composition between field samples and NRB enrichments (Table 2) and indicates that nitrate breakthrough at PWs in the field causes significant changes in produced water community composition.

DGGE analysis of archaea.The archaeal component of the microbial community of the Enermark field was analyzed by amplification of DNAs from 90 field samples with the primers U515f-GC and A1041r. The resulting DGGE gels showed up to 4 distinct bands per lane (not shown). A total of 271 bands were excised, reamplified, and analyzed by sequencing. Of these, 195 (72%) gave good-quality sequences. The closest GenBank matches of all these are provided in Table S2 in the supplemental material. Identified genera included Methanolobus species, which are obligate methylotrophs known to metabolize methanol and methylamines, including trimethylamine, to methane and ammonium (17). Inoculation of 5% (vol/vol) of PW inoculum into 160-ml serum bottles containing 100 ml of methanogenic medium with 5 mM trimethylamine as the sole methanogenic substrate gave up to 10 mM methane in the headspace within 15 days for 2-PW, 4-PW, 13-PW, 15-PW, and 16-PW, whereas no methane was detected for incubations with 3-PW, 5-PW, 10-PW, 11-PW, 18-PW, and 19-PW for 30 days of incubation (see Fig. S1 in the supplemental material).

Detection of anammox bacteria.Although some DGGE-derived sequences were affiliated with the planctomycetes (Table 2), none of these were closely related to groups with known anammox function. However, since ammonium is present in oil fields throughout and nitrite is generated from injected nitrate, conditions are met for the proliferation of anammox bacteria. Efforts were therefore made to detect these by PCR with specific primers.

DNAs extracted from produced water obtained from 3-PW, 7-PW, 13-PW, and 15-PW, as well as from an hNRB enrichment from 7-PW, were amplified with the primers Pla46F and 1492r and then with the nested primers amx368F and amx820r. A PCR product was obtained with the DNA extracted from the 15-PW field sample. Cloning and sequencing gave 19 sequences affiliated with known species of freshwater anammox bacteria, “Candidatus Brocadia anammoxidans,” “Candidatus Brocadia fulgida,” and “Candidatus Kuenenia stuttgartiensis” (Fig. 4: 8 sequences that were nearly identical to those of “Ca. Kuenenia stuttgartiensis” are not shown).

• Open in new tab
• Download powerpoint

FIG. 4.

Phylogenetic tree of anammox sequences retrieved from the Enermark field. DNA was isolated from 15-PW produced water and amplified with anammox-specific primers. The PCR product was cloned, and 16 clones were sequenced. Of these, 11 were closely associated with known anammox species, as shown. “Brocadia anammoxidans,” “Brocadia fulgida,” “Kuenenia stuttgartiensis,” “Jettenia asiatica,” “Anammoxoglobus propionicus,” and “Scalindua brodae” are “Candidatus” species.

Effect of nitrate injection on ammonium concentrations in the field.The water cycle in a PWRI operation, as at the Enermark field, can be summarized as follows (Fig. 1). Produced water from multiple PWs is separated from oil and collected in a WP, where it is mixed with MW (average ammonium concentration, 0.27 ± 0.11 mM; n = 3) and nitrate concentrate; the resulting injection water is transported by pipeline to multiple IWs. The nitrate concentration in injection water leaving 1-WP, the main WP serving the field, is shown in Fig. 5 D as a function of time. It averaged 2.4 mM from week 1 to week 64 and from week 96 to week 125. From week 64 to week 96, the concentration alternated weekly between 2.4 and 14 mM, increasing the average injected concentration to 8.2 mM. The ammonium concentrations of the injection water were monitored at 1-WP and at injection well 14-IW (Fig. 5A and B). The transit time between these two sites is 1 day. Fluctuations in the ammonium concentration at 1-WP were matched at 14-IW (Fig. 5C) (the slope of the line of best fit is 1.04). This indicates that the observed changes are significant and that nitrate injected at 1-WP was not reduced to ammonium during transport to 14-IW. It is interesting to note that the lowest ammonium concentrations at 1-WP and 14-IW were observed following the period of increased nitrate injection (Fig. 5A and B).

• Open in new tab
• Download powerpoint

FIG. 5.

(A and B) Ammonium concentrations in injection water as a function of time for water leaving 1-WP (A) or water injected at 14-IW (B). (C) Relation between ammonium concentrations at 1-WP and 14-IW, which are connected by a pipeline (transit time, 1 day). (D) Nitrate concentration at 1-WP. A period of increased nitrate injection is marked (↔). The line in this panel has been drawn based on data presented previously (39).

At 14-IW, batches of nitrate (760 mM for 1 h per week) were injected from week 33 to week 101. This gave a breakthrough of nitrite (up to 0.8 mM), increased concentrations of sulfate (up to 0.25 mM), and zero sulfide at the neighboring producing well 13-PW from week 41 onwards (39), indicating a breakthrough time of 8 weeks for this well pair. Overall, the batchwise injection increased the nitrate concentration injected at 14-IW to 6.9 mM from week 33 to week 64 and to 12.7 mM from week 64 to week 96 during field-wide nitrate pulsing. Breakthrough of up to 0.4 mM nitrate was also seen from week 41 onwards (Fig. 6A). However, ammonium concentrations in samples from 13-PW were within a narrow range from 0.2 to 0.3 mM for the entire duration of increased nitrate injection from week 33 to week 101 (Fig. 6B), suggesting that little of the injected nitrate was reduced to ammonium.

• Open in new tab
• Download powerpoint

FIG. 6.

Nitrate (A) and ammonium (B) concentrations at producing well 13-PW as a function of time. Nitrate broke through in week 41 due to injection of batches of high nitrate concentration at neighboring injector 14-IW from week 33 to week 101 (↔). Increased concentrations of nitrite and sulfate and zero sulfide were also observed at 13-PW during this period, as shown elsewhere (39).

The ammonium concentrations in samples from 11 PWs are shown in Fig. 7A to K for a 2-year period of nitrate injection. A volume weighted average ammonium concentration for all 11 PWs is given in Fig. 7L. In addition to a moving average, lines of best fit have been drawn through the data. These trend downwards, except for 13-PW. The lowest recorded ammonium concentrations were often found during or following the high nitrate application (Fig. 7). Overall, the data support the notion that little of the injected nitrate is converted into ammonium by DNRA.

• Open in new tab
• Download powerpoint

FIG. 7.

(A to K) Ammonium concentration at individual PWs as a function of time. (L) Production water volume-averaged ammonium concentration for 11 PWs, calculated as described in the text. Lines representing a moving average for two adjacent data points and a linear best fit are shown. A period of increased field-wide nitrate injection is indicated (↔).

Fate of nitrate in NRB enrichment cultures.The concentrations of nitrate, nitrite, and ammonium were monitored in hNRB enrichments containing 4 mM nitrate and 3 mM VFA and in soNRB enrichments containing 4 mM nitrate and 4 mM sulfide. The ammonium concentration in the medium was lowered to 1 mM to allow more-accurate monitoring of changes in this concentration. The medium bottles were inoculated with 8% (vol/vol) of produced water. Nitrate was completely reduced within 30 to 50 h with transient formation of nitrite, which was in turn reduced within 120 h (Fig. 8). The ammonium concentration decreased on average by 0.44 ± 0.09 mM (n = 5). Biomass formation, monitored as the increase in the protein concentration with time, correlated with the decrease in the ammonium concentration (see Fig. S2 in the supplemental material). The increase in the protein concentration was on average 40 ± 5 mg/liter (n = 5). Assuming 55% of the biomass dry weight to be protein (15), we calculate an hNRB biomass concentration of 73 mg/liter. Using CH_1.2O_0.5N_0.2 as the elemental composition of biomass (M_r = 24.6), this corresponds to a concentration of 3 mM assimilated carbon and of 0.6 mM assimilated ammonium. Hence, the complete reduction of 4 mM nitrate shown in Fig. 8 yielded at most 0.16 mM (4%) ammonium through DNRA, indicating that 96% was reduced through denitrification. No decline in ammonium concentrations was seen in soNRB incubations, because at 22 ± 2 mg/liter (n = 5), the biomass concentration formed in these incubations was much smaller (results not shown).

• Open in new tab
• Download powerpoint

FIG. 8.

Change in the concentrations of nitrate (▪), nitrite (▴), and ammonium (⧫) with time in serum bottles containing 100 ml of medium with 3 mM VFA and 4 mM nitrate and inoculated with 8 ml of produced water as indicated.

Effect of nitrate on community composition in low- and high-temperature reservoirs.Continuous injection of water containing 1 mM sulfate and 2.4 mM nitrate in the low-temperature Enermark field has been proposed to cause emergence of zones of predominantly nitrate reduction, sulfate reduction, or methanogenesis (39), as indicated in Fig. 1, zones A, B, and C, respectively. Souring is not lowered under these conditions, since the excess oil electron donor will cause all injected nitrate and sulfate to be sequentially reduced. In contrast, in high-temperature reservoirs, as found in the North Sea, long-lasting elimination of souring has been demonstrated as the result of nitrate injection. A plausible reason for the difference is that microbial activity in high-temperature reservoirs is limited to the near-injection wellbore region (equivalent to Fig. 1, zone A), where cold seawater is injected, whereas the rest of the reservoir (equivalent to Fig. 1, zones B and C) is too hot for significant microbial activity. If true, then only a limited region of the reservoir needs to be treated, and this can be done effectively by continuous nitrate injection. This is in contrast to low-temperature reservoirs, where growth can occur throughout (Fig. 1).

The effect of nitrate injection on microbial community composition in high-temperature North Sea reservoirs has been studied (2, 3, 8). With respect to epsilonproteobacteria, Bødtker et al. (2, 3) reported Sulfurimonas and Arcobacter spp. as major community components at injection wellheads, whereas Gittel et al. (8) identified Sulfurospirillum and Arcobacter as being increased during nitrate injection. These genera were also identified at Enermark (Table 2). Distinct hNRB or fermentative bacteria, some thermophilic, identified in the North Sea fields but not at Enermark, included Phaeobacter, Terasakiella, and Deferribacter spp. (3, 8).

A problem in determining the effect of nitrate on oil field microbial communities is that one is often limited to sampling planktonic bacteria in PWs instead of the rock-bound biofilm in the near-injection wellbore region, which one ideally would like to sample. The modest rise in NRB activity during nitrate injection (Fig. 2) may reflect what goes on in zone C but does not reflect the potentially strong increase in rock-attached NRB in zone A (Fig. 1). It is important to keep this perspective when discussing the effect of nitrate on microbial communities in oil fields.

Microbial community at Enermark.DGGE analysis of all samples indicated the presence of 51 bacterial genera and 8 archaeal genera (Tables 2 and 3). Nitrate breakthrough at 13-PW (Fig. 6A) increased the frequency of observation of alpha-, beta- and epsilonproteobacteria (NRB) at the expense of deltaproteobacteria (SRB). Identified potential hNRB included the alphaproteobacteria Dechlorospirillum (5) and Pannonibacter and the betaproteobacteria Azoarcus, Diaphorobacter, Simplicispira and Thauera, as well as the Comamonadaceae. Thauera, also isolated from the nearby Coleville field (13), and Azoarcus and Diaphorobacter couple reduction of nitrate to N₂ to oxidation of aromatics (19, 32). Gammaproteobacterial hNRB included the genera Moritella, Pseudomonas, Pseudoxanthomonas, and Thermomonas (22, 36, 42). Denitrovibrio acetiphilus is one of few identified ammonium-generating hNRB (25). Epsilonbacterial soNRB identified at Enermark included Arcobacter, Sulfurimonas, and Sulfurospirillum spp. These were also previously isolated from Coleville produced waters (7, 10, 13). The genus Sulfurovum, which couples reduction of nitrate to oxidation of either sulfur or thiosulfate to sulfate (14), has not been found previously in Western Canadian oil reservoirs. In addition to the sulfate-reducing Desulfovibrio and Desulfobulbus, the sulfur-reducing genus Desulfuromonas was also frequently observed (Table 2). All of these belong to the Deltaproteobacteria. The nonproteobacterial Geovibrio is also a sulfur reducer (4). Sulfur may arise by partial nitrate-mediated oxidation of sulfide along the water flow path.

In the absence of sulfate or nitrate, oil components are converted to methane and CO₂ by consortia of syntrophs and methanogens (43). Syntrophs, catalyzing the initial hydrocarbon attack, include members of the genera Syntrophus and Clostridium, as well as SRB in the absence of sulfate (41). hNRB may similarly act as fermenters when nitrate is absent, and this could explain the high activity of these bacteria even prior to nitrate injection (Fig. 2, t = 0). Methanogens of the genus Methanocalculus were frequently detected at 13-PW and at 8-IW (45 of 65 and 5 of 5 DGGE bands analyzed, respectively), suggesting that these grow better than other methanogens under environments containing nitrate. Methanolobus species are of special interest for the N cycle in oil fields because these metabolize methylamines, including trimethylamine (TMA), to methane and ammonium (17). This activity was demonstrated at Enermark (see Fig. S1 in the supplemental material) and may contribute to ammonium production.

Fate of injected nitrate and role of anammox bacteria.In microcosms containing nitrate, VFA, and PW inoculum, nitrate was often quantitatively converted to nitrite before being reduced further (Fig. 8). This was also observed when oil was the only electron donor (20). Injected nitrate is thus likely converted to nitrite in the near-injection wellbore region (Fig. 1, zone A) before being reduced to either ammonium or N₂. The ammonium demand for biomass formation can cause the ammonium concentration to drop significantly when most nitrate is reduced through denitrification as opposed to DNRA (Fig. 8). Under these conditions, decreases in the ammonium concentration may be expected in the field as a result of sudden increases in the injected nitrate concentration. One would not expect these during long-term, continuous injection of the same concentration, where a steady-state biomass concentration is likely to be maintained. The data obtained (Fig. 5A and B and Fig. 7) provide evidence for sudden decreases in the ammonium concentration following periods of increased nitrate injection.

Hence, the near-injection wellbore region is expected to have millimolar concentrations of nitrite and submillimolar concentrations of ammonium, conditions suitable for the growth of anammox bacteria (33). Planctomycetes are ubiquitous in soils and aquatic environments but have been found in oil fields only as a minor component of clone libraries (26). Anammox bacteria are not easily detected using standard bacterial 16S rRNA primers (16) but were identified in community DNA obtained from producing well 15-PW but not in 3-PW, 7-PW, and 13-PW with the specific primers used here. This could mean (i) that these organisms are not widespread in the field, (ii) that they are not easily dislodged from the reservoir rock in the near-injection wellbore region, or (iii) that anammox species present in the field are not being detected efficiently with the primers used. The overall effect of the anammox reaction in an oil field subjected to nitrate injection is that a larger fraction of the injected nitrate is reduced to N₂ at the expense of ammonium as the end product. This could give rise to the slow decrease in the ammonium concentration from 0.3 to 0.2 mM observed in produced waters over a 2-year period (Fig. 7).

Hence, the results obtained here indicate that nitrate injected into the Enermark field to oxidize sulfide is reduced mostly through denitrification, not through DNRA. Anammox may cause the slow decrease in ammonium concentrations observed during continued nitrate injection (Fig. 7), with biomass formation causing transient decreases during periods of increased nitrate injection.