Article pubs.acs.org/est
Stable Isotopes Reveal Widespread Anaerobic Methane Oxidation Across Latitude and Peatland Type Varun Gupta,†,# Kurt A. Smemo,*,‡,§,# Joseph B. Yavitt,⊥,# David Fowle,¶ Brian Branfireun,∥ and Nathan Basiliko†,# †
Department of Geography, University of Toronto Mississauga, Mississauga, ON L5L 1C6, Canada The Holden Arboretum, 9500 Sperry Road, Kirtland, Ohio 44094, United States § Department of Biological Sciences, Kent State University, Kent, Ohio 44242, United States ⊥ Department of Natural Resources, Cornell University, Ithaca, New York 14853, United States ¶ Department of Geology, University of Kansas, Lawrence, Kansas 66045, United States ∥ Department of Biology and Centre for Environment and Sustainability, Western University, London, ON, N6A 5B7, Canada ‡
S Supporting Information *
ABSTRACT: Peatlands are an important source of the atmospheric greenhouse gas methane (CH4). Although CH4 cycling and fluxes have been quantified for many northern peatlands, imprecision in processbased approaches to predicting CH4 emissions suggests that our understanding of underlying processes is incomplete. Microbial anaerobic oxidation of CH4 (AOM) is an important CH4 sink in marine sediments, but AOM has only recently been identified in a few nonmarine systems. We used 13C isotope tracers and followed the fate of 13 C into CO2 and peat in order to study the geographic extent, relative importance, and biogeochemistry of AOM in 15 North American peatlands spanning a ∼1500 km latitudinal transect that varied in hydrology, vegetation, and soil chemistry. For the first time, we demonstrate that AOM is a widespread and quantitatively important process across many peatland types and that anabolic microbial assimilation of CH4−C occurs. However, AOM rate is not predicted by CH4 production rates and the primary mechanism of C assimilation remains uncertain. AOM rates are higher in fen than bog sites, suggesting electron acceptor constraints on AOM. Nevertheless, AOM rates were not correlated with porewater ion concentrations or stimulated following additions of nitrate, sulfate, or ferric iron, suggesting that an unidentified electron acceptor(s) must drive AOM in peatlands. Globally, we estimate that AOM could consume a large proportion of CH4 produced annually (1.6−49 Tg) and thereby constrain emissions and greenhouse gas forcing.
■
INTRODUCTION
Microbial anaerobic oxidation of CH4 (AOM) has received little attention in peatlands, even though AOM in sulfate (SO42−)-rich marine sediments is a globally important CH4 sink7 and is estimated to consume 70−300 Tg yr−1.7 In freshwater aquatic ecosystems, evidence suggests that AOM is coupled to SO42−,8,9 ferric iron (Fe3+),10,11 and nitrate (NO3−) reduction,12,13 and recent work has demonstrated that nitritedependent AOM is carried out by CH4-oxidizing bacteria in a polluted minerotrophic peatland receiving high groundwater NO3− inputs.14 Further evidence suggests that AOM can be carried out by methanogens via “reverse methanogenesis” in peat soils15 and that this process is quantitatively more important in systems with high CH4 production,16 such as
Emissions of the potent atmospheric greenhouse gas methane (CH4) from northern peatlands are approximately 12−23 Tg yr−1.1,2 In contrast, peat soils produce CH4 in amounts that far exceed these emissions, but aerobic CH4-oxidizing bacteria living in surface peat, pools of water, or symbiotically in Sphagnum mosses consume a portion of the CH4 produced.3 Microbial processes, CH4 transport mechanisms through peat soils, and environmental controls on CH4 emissions have been studied extensively; nevertheless, CH4 emissions remain notoriously difficult to predict, with more than 3 orders of magnitude variability among sites with similar hydrology, soil properties, and plant communities,4 and emissions are rarely explained by the balance between estimated anaerobic CH4 production and aerobic CH4 oxidation.5 It is likely that emissions are constrained by additional biogeochemical controls other than those considered in most conceptual models of peatland CH4 dynamics (Figure 1).6 © 2013 American Chemical Society
Received: Revised: Accepted: Published: 8273
January 30, 2013 June 28, 2013 July 3, 2013 July 3, 2013 dx.doi.org/10.1021/es400484t | Environ. Sci. Technol. 2013, 47, 8273−8279
Environmental Science & Technology
Article
S1). A 30 cm deep peat core (10 cm diameter) was extracted from a depth starting at 15 cm below the ambient water table (see SI Table S1) and stored in sterilized canning jars that were overfilled with pore water and transported on ice to the laboratory within 1−2 days. Within-site samples were combined, homogenized, and frozen at −20 °C until incubations commenced (∼ 2 months). Incubation Setup. Incubations were carried out in 2 batches. Seven sites were incubated in batch 1 and 8 sites in batch two, both at 19 °C. For each site and measurement time, 3 (batch 1) or 2 (batch 2) sets of triplicate 500-mL serum vials sealed with thick rubber stoppers (Geo-Microbial Technologies, Ocheleta OK, USA; Cat. 1313l) were filled with 30 g of wet peat and 70 mL of degassed, deionized (DDI) H2O. Vials were evacuated with a vacuum pump for 5 min, back-flushed with high purity N2 5 times, and subsequently filled with N2 to 4 kPa above ambient pressure after the fifth cycle. Jars were stored in the dark for 4 days to allow any trace amounts of O2 that might have remained to be consumed. After the preincubation period, each jar received 10 mL of N2, 12C− CH4, or 13C−CH4 via syringes that were fitted with sealed stopcocks and flushed 4 times with ultrahigh purity N2. Only the 7 sites incubated in batch 1 received the 12C−CH4 additions. All isotopic gases were manufactured by Cambridge Isotopes (Cambridge MA, USA, 99.99% isotope purity, where less than 10 ppm chemical impurity was confirmed not to be O2 by the manufacturer). Headspace gas was mixed using the same syringe, and 10 mL of headspace gas was removed and the stopcock closed prior to the removal of the syringe. This process did not introduce detectable O2 (with methylene blue indicator), and drawing samples from a slightly over-pressurized headspace and analyzing at ambient pressure provided flux underestimates. This measurement represented “time zero”. Headspace CH4 and CO2 concentrations were determined using an 8610C gas chromatograph (GC) equipped with a flame ionization detector and in-line methanizer (SRI Instruments, Torrance, CA, USA). The GC injector, column, and methanizer temperatures were 70, 45, and 350 °C, respectively. Volumetric concentrations of CH4 and CO2 were determined relative to commercial standards and then converted to mass. Instrument error was less than 2% based on 10 replicate samples of commercial standards. Determination of AOM to CO2. At days 3, 20, and 40, a set of three incubations was destructively sampled for each site and headspace treatment (i.e., N2, 12CH4, 13CH4) to analyze the headspace CH4 and CO2 concentration, the isotopic signature of the headspace CO2, and the isotopic signature of organic C in peat. First, the headspace gas was gently mixed and 10 mL of gas was removed for GC analysis. Second, remaining headspace gas was evacuated using a 60 mL syringe and bubbled through 5 mL of 0.2 M sodium hydroxide solution at a rate of 1 mL sec−1 (this rate was empirically optimized, data not shown) in a 15-mL falcon tube to base-capture the CO2. Once all the headspace CO2 was captured, 5 mL of 0.2 M barium chloride solution was added to each tube. Tubes were centrifuged at 4000g for 30 min at room temperature and the supernatant was gently removed using a pipet without disturbing the pellets. Then, 7 mL of DDI water was added to each tube, the pellets were resuspended, and the tubes were centrifuged again at the same setting. This process was repeated three times, after which tubes were oven-dried at 45 °C and shipped to the Keck Paleoenvironmental & Environmental Stable-Isotope Laboratory at the University of Kansas. Samples were analyzed using a
Figure 1. Conceptual diagram of CH4 fluxes between peatlands and the atmosphere where the question mark indicates a largely unknown role of AOM.
many peatlands. Identifying AOM as a CH4 sink in a broader range of peatlands might therefore be unsurprising. However, due to chronic reducing conditions and little geochemical and atmospheric input, northern peatlands typically lack high concentrations of oxidized inorganic electron acceptors often linked to AOM,17 suggesting thermodynamic constraints in otherwise CH4-rich peat soils.6 Smemo and Yavitt18 were the first to conclusively demonstrate AOM in peat soil against a large background of CH4 production, although evidence relating to global importance and metabolic pathway(s) was lacking and much about the biogeochemistry of the process remains unknown. Here, we used 13C tracers to study the geographic extent, relative importance, and biogeochemistry of AOM in 15 North American peatlands spanning a ∼1500 km latitudinal transect and varying in hydrologic, vegetation, and soil chemical properties. Our overall aim was to determine the magnitude of rates and controls on AOM in peatlands across a wide range of northern peatland types: this allowed us to determine if peatland type had a significant influence on the relative importance of AOM and the metabolic fate of CH4−C as a function of peatland type. To this end we performed an electron acceptor addition experiment to determine if AOM rates were influenced by the availability of putative inorganic electron acceptors.
■
MATERIALS AND METHODS Field Sampling. We sampled 15 peatlands (SI Figure S1) across a ∼1500 km latitudinal transect stretching from subarctic Canada (James Bay Lowlands) to the most southerly true ombrotrophic bogs in the cool-temperate region of the United States (central Appalachians). Detailed site information can be found in SI Table S1. In each of the study sites, peat was collected from 5 random locations in summer and early fall 2009 and mid-summer 2010, with more southerly sites sampled earlier in the growing season and northerly sites (James Bay lowlands) sampled in early fall with the intent of approximately standardizing soil temperature at time of sampling (SI Table 8274
dx.doi.org/10.1021/es400484t | Environ. Sci. Technol. 2013, 47, 8273−8279
Environmental Science & Technology
Article
Kiel Carbonate Device III + Finnigan MAT253 isotope ratio mass spectrometer (ThermoFinnigan, Dreieich, Germany). Data are reported in δ13C VPDB and sample measurement precision was better than ±0.10‰. The day 20 data for McLean Bog were not collected due to accidental loss of the samples. CH4−C Assimilation in Peat. After headspace gas removal, vials for each headspace treatment were opened for peat organic C isotopic analysis. Peat was extracted using a spatula, pH of the slurry was recorded, and 2 N HCl was subsequently added to acidify peat to pH 2. Then, samples were dried at 60 °C and ground using a Wiley Mill (Thomas Instruments Swedesboro, NJ, USA). Ground peat samples were analyzed at the Keck Paleoenvironmental & Environmental Stable Isotope Laboratory using a Costech 4010 elemental analyzer (EA) in conjunction with a Thermo Finnigan MAT 253 IRMS. Data were reported as δ13C VPDB, and sample measurement precision was better than ±0.22 ‰. Electron Acceptor Addition. Peat from Michigan Hollow (SI Table S1) was used to test the influence of potential electron acceptors (NO3−, SO42−, and Fe3+) on AOM following the general isotope tracer methods described above except that incubations were only run for 21 days following additions, isotope tracer control incubations received 12CH4 (not N2), and sampling occurred only at day 21. A control consisting of 1 mL of DDI H2O or 1 mL of degassed solutions containing NO3− (100 μM NO3− as Ca(NO3)2), SO42− (100 μM SO42− as Na2SO4), or Fe3+ (200 μM Fe3+as Fe(OH)3)19 was added to triplicate vials as above, however only one composite barium carbonate sample was run for isotope analyses for each treatment. Calculations and Data Analyses. Net CH4 and CO2 production rates were calculated based on GC measurements of headspace gas (no CH4 added) and expressed per mass dry peat following Gupta et al.20 For isotopic analysis, δ13C VPDB values were converted to 13C atom percent (13C AP) following this equation: 13C AP = ((δ13C VPDB + 1000)/(δ13C VPDB + 1000 + (1000/RStandard)))*100, where RStandard = 0.0118. The average 13C AP for the three replicates from each treatment and time interval was then corrected for CO2 in the dissolved phase using Henry’s Law in order to determine how much headspace CO2 was unaccounted for due to equilibration with porewater.18 It was determined after the first set of incubations that there were no significant differences in C isotope fractionation between the N2 and the 12CH4 incubations, and the latter was not included in the second batch (SI Figure 2; data not shown for peat phase). For both the CO2 and peat phase, respective 13 C AP values from N 2 controls (illustrating natural fractionation) were subtracted from those of the 13CH4 additions. The differences represented rates of CH4 oxidation to CO2 or C assimilation under anoxic conditions, and only values significantly (p < 0.05) greater in the 13CH4 additions than in the control were accepted as indicative of real AOM. Gross CH4 production was calculated as the sum of net CH4 production plus total anaerobic CH4 consumption.18 Statistical Analysis. Production, oxidation, and assimilation values were tested to confirm that the data had a normal distribution. Student t tests were conducted to determine significant differences (P ≤ 0.05) between treatments at the respective sampling time. As noted earlier, if significant 13C AP differences were not observed between the N2 additions and 13 CH4 additions, they were removed from further analysis. Variability is expressed as standard deviation, calculated using
Figure 2. Patterns of gross anaerobic CH4 oxidation measured using added 13C−CH4 tracers where subsequent 13CO2 concentration was measured over time in (a) fens and (b) bogs over time.
each set of three replicates from each sampling time. Correlations between site characteristics and measured rates were explored and correlation coefficients and resulting P values were calculated. Single t tests were also used to characterize the differences between categorical sets of sites (e.g., bog and fen).
■
RESULTS AND DISCUSSION Using 13C−CH4 tracers (2.0 kPa headspace CH4) and anaerobic metabolism in peat from 15 North American peatlands spanning from a permafrost site in northern Ontario to one of the most southerly known nonalpine Sphagnum peatlands in North America (SI Table S1), we found that after 20 and 40 days there was significant production of 13CO2, and therefore AOM, from all 15 peatlands (Figure 2). Amending incubations with 12C−CH4 resulted in a 13CO2 production identical to that of incubations only receiving N2 (SI Figure S2); 13CO2 in the headspace of the incubations that received 13 CH4 additions could only be the result of AOM. However, relative CO2 production rates were similar among all treatments (SI Table S2). The 5A′ Fen in Canada’s vast James Bay lowland region had the fastest AOM rate by day 20 (7.09 ± 0.37 nmol CH4 kg dry peat−1 s−1) and day 40 (4.74 ± 0.49 nmol CH4 kg dry peat−1 s−1). Site-averaged AOM rates ranged from 0.28 ± 0.03 to 7.09 ± 0.37 nmol CH4 kg−1 s−1 with an overall average 8275
dx.doi.org/10.1021/es400484t | Environ. Sci. Technol. 2013, 47, 8273−8279
Environmental Science & Technology
Article
Table 1. Mean Rates of Anaerobic CH4 Oxidation, Assimilation, and Total Consumption (±SD) over 20 and 40 Daysa CH4−C assimilation in peat
anaerobic oxidation of CH4 to CO2 study site
0−20 d
0−40 d
Michigan Hollow White River Rich Fen White River Int. Fen Channel Fen Big Run Bog Bog Lake Fen SA460 Fen 5A′ Fen
0.87 ± 0.21 2.08 ± 0.07 1.62 ± 0.17 2.53 ± 1.66 1.42 ± 0.1 1.12 ± 0.37 1.33 ± 0.07 7.09 ± 5.16
1.56 ± 0.1 1.57 ± 0.09 1.31 ± 0.03 1.77 ± 0.39 1.36 ± 0.16 2.54 ± 0.43 1.2 ± 0.1 4.74 ± 0.49
White River Poor Fen Buckles Bog Bog on Permafrost Mclean Bog Dryden Bog S2 Bog S1 Bog
0.98 ± 0.23 0.28 ± 0.03 0.12 ± 0.04 b 1.05 ± 0.16 0.91 ± 0.04 0.65 ± 0.04
1.15 0.47 0.26 0.88 0.83 0.83 1.07
± ± ± ± ± ± ±
0.02 0.05 0.19 0.06 0.27 0.11 0.08
0−20 d Fens/Higher pH 0.22 ± 0.02 0.29 ± 0.13 0.34 ± 0.08 3.45 ± 1.35 NSD NSD NSD NSD Bogs/Lower pH NSD NSD NSD 0.52 ± 0.16 0.45 ± 0.21 NSD 4.55 ± 3.08
total CH4 consumption
0−40 d
0−20 d
0−40 d
0.25 ± 0.1 0.21 ± 0.04 0.36 ± 0.07 NSD NSD 5.96 ± 3.37 0.45 ± 0.24 NSD
1.09 ± 0.22 2.38 ± 0.15 1.96 ± 0.19 5.98 ± 2.14 1.42 ± 0.1 1.12 ± 0.37 1.33 ± 0.07 7.09 ± 5.16
1.81 ± 0.14 1.79 ± 0.09 1.66 ± 0.08 1.77 ± 0.39 1.36 ± 0.16 8.49 ± 3.4 1.65 ± 0.26 4.74 ± 0.49
NSD NSD NSD 0.3 ± 0.07 0.36 ± 0.17 NSD 3.5 ± 1.84
0.98 ± 0.23 0.28 ± 0.03 0.12 ± 0.04 0.52 ± 0.16 1.5 ± 0.26 0.91 ± 0.04 5.2 ± 3.08
1.16 0.47 0.26 1.18 1.19 0.83 4.56
± ± ± ± ± ± ±
0.02 0.05 0.19 0.09 0.32 0.11 1.84
All rates are nmol CH4 kg dry peat−1 s−1. NSD indicates no significant difference (p > 0.05) in 13C concentrations between controls and samples with 13CH4 additions. Study sites are arranged in order from sites with higher pH porewater and connectivity to groundwater (fens) to lower pH sites where surface peat is largely isolated from regional groundwater (bogs). bSamples lost. a
40-day AOM rate of ∼1.5 nmol CH4 kg−1 s−1 (Table 1). Our reported rates are faster than those reported in marine continental margins21 and freshwater systems (reviewed in ref 7), but slower than marine CH4 seeps21 and the average rate reported by Smemo and Yavitt (∼17 nmol CH4 kg−1 s−1).18 This is not surprising as AOM rates could be a reflection of overall system CH4 dynamics,6,16 and Smemo and Yavitt18 derived that rate based on multiple studies in a nutrient-rich, high CH4-producing peatland without Sphagnum moss. Thus, the relatively nutrient-poor Sphagnum-dominated peat environment found in most of our sites could explain the somewhat lower average rates we report here. Moreover, the stableisotope tracer technique used in this study is similar to those used in many marine AOM studies.22,23 This approach provides more robust rate estimates than the first-order kinetic isotope pool dilution technique used by Smemo and Yavitt18 and represents a direct measurement of complex syntrophic interactions (such as with marine AOM consortia)24 that do not follow first-order kinetic principles. Previous studies from marine and freshwater sediments have reported that bacterial and archaeal communities responsible for AOM assimilate some CH4-derived C into biomass,12,25,26 but such patterns have not been explored in peatlands. We measured bulk assimilation of 13C that entered soil organic matter through the microbial community via 13CH4 oxidation. Because microbial C generally represents
