Ecosystems

, Volume 10, Issue 7, pp 1148–1165

Regulation of Decomposition and Methane Dynamics across Natural, Commercially Mined, and Restored Northern Peatlands

Authors

    • Department of Geography and Centre for Climate and Global Change ResearchMcGill University
    • Department of GeographyUniversity of Toronto, Mississauga
  • Christian Blodau
    • Department of Geography and Centre for Climate and Global Change ResearchMcGill University
    • Limnological Research Station and Department of HydrologyUniversity of Bayreuth
  • Charlotte Roehm
    • Department of Geography and Centre for Climate and Global Change ResearchMcGill University
    • Limnological Research Station and Department of HydrologyUniversity of Bayreuth
    • Département des Sciences BiologiquesUniversité du Québec à Montréal
  • Per Bengtson
    • Department of Forest SciencesUniversity of British Columbia
  • Tim R. Moore
    • Department of Geography and Centre for Climate and Global Change ResearchMcGill University
Article

DOI: 10.1007/s10021-007-9083-2

Cite this article as:
Basiliko, N., Blodau, C., Roehm, C. et al. Ecosystems (2007) 10: 1148. doi:10.1007/s10021-007-9083-2

Abstract

We examined aerobic and anaerobic microbial carbon dioxide (CO2) and methane (CH4) exchange in peat samples representing different profiles at natural, mined, mined-abandoned, and restored northern peatlands and characterized the nutrient and substrate chemistry and microbial biomass of these soils. Mining and abandonment led to reduced nutrient and substrate availability and occasionally drier conditions in surface peat resulting in a drastic reduction in CO2 and CH4 production, in agreement with previous studies. Owing mainly to wetter conditions, CH4 production and oxidation were faster in restored block-cut than natural sites, whereas in one restored site, increased substrate and nutrient availability led to much more rapid rates of CO2 production. Our work in restored block-cut sites compliments that in vacuum-harvested peatlands undergoing more recent active restoration attempts. The sites we examined covered a large range of soil C substrate quality, nutrient availability, microbial biomass, and microbial activities, allowing us to draw general conclusions about controls on microbial CO2 and CH4 dynamics using stepwise regression analysis among all sites and soil depths. Aerobic and anaerobic decomposition of peat was constrained by organic matter quality, particularly phosphorus (P) and carbon (C) chemistry, and closely linked to the size of the microbial biomass supported by these limiting resources. Methane production was more dominantly controlled by field moisture content (a proxy for anaerobism), even after 20 days of anaerobic laboratory incubation, and to a lesser extent by C substrate availability. As methanogens likely represented only a small proportion of the total microbial biomass, there were no links between total microbial biomass and CH4 production. Methane oxidation was controlled by the same factors influencing CH4 production, leading to the conclusion that CH4 oxidation is primarily controlled by substrate (that is, CH4) availability. Although restoring hydrology similar to natural sites may re-establish CH4 dynamics, there is geographic or site-specific variability in the ability to restore peat decomposition dynamics.

Keywords

carbon dioxideFTIR spectroscopylipidsmethane oxidationmicrobial biomassnitrogennutrientspeatphosphorusroots

Introduction

Northern peatland ecosystems cover just 3% of the earth’s surface, yet have sequestered 270–455 Gt of carbon (C) from the atmosphere (Turunen and others 2002), or up to one-third of total global soil C, and are contemporary atmospheric C sinks (Roulet and others 2007). This is primarily related to low carbon dioxide (CO2) emission rates arising from a combination of cold temperatures and partially waterlogged, anoxic conditions and unique soil organic properties arising from endemic plant tissues that constrain soil microbial respiration (Moore and Basiliko 2006). Peatlands are large contemporary sources of atmospheric methane (CH4) (Vasandar and Kettunen 2006), as methanogenic production under anoxic conditions is often greater than CH4 oxidation rates under oxic conditions. Long-term feedbacks between hydrologic and ecological dynamics are largely responsible for the unique biogeochemistry of peatlands, although controls on microbial activities responsible for net efflux of CO2 to the atmosphere and CH4 production and consumption are poorly understood. There is a need to better understand controls on greenhouse gas exchange in peatlands to predict how anthropogenic disturbances, including land-use, climate, and environmental changes such as increased atmospheric deposition, impact these processes that include large potential climate and environmental change feedbacks.

Commercially mined peat is a valuable agronomic and horticultural resource (Cleary and others 2005). Mining involves drainage and removal of vegetation, which results in lower and more variable water tables and eliminates primary production. Most contemporary peat mining is conducted using large scale milling and vacuum removal of recently dried peat, although a small number of peatlands are harvested using more traditional block-cutting techniques that otherwise became substantially less common in the 1970s (Cleary and others 2005). Several studies have explored natural and active restoration of hydrologic and vegetation dynamics in eastern Canada (for example, Lavoie and others 2005; McNeil and Waddington 2003; Girard and others 2002; Robert and others 1999; Grosvernier and others 1997; also see http://www.gret-perg.ulaval.ca/en_publications.html). Previous work has also explored the effects of mining and restoration on CO2 and CH4 dynamics and the results depend on the method of mining used and subsequent success of restoring hydrological dynamics and establishing vegetation and eventually a new peat profile (Marinier and others 2004; Petrone and others 2003, 2001; Waddington and others 2002; Sundh and others 2000; Tuittila and others 2000, 1999; Moore and others unpublished). Mining creates sites that are net sources of CO2 (for example, Waddington and others 2002) and may have reduced CH4 emissions (for example, Tuittila and others 2000), whereas the effects of revegetation and peat regeneration on fluxes of CO2 and CH4 are variable among sites (for example, Marinier and others 2004; McNeil and Waddington 2003; Tuittila and others 1999). The role of microorganisms and the decomposition of peat substrate that have been drastically altered through mining and restoration are clearly important factors driving greenhouse gas fluxes, yet controls on activities are poorly understood (Glatzel and others 2004).

Through mining, deep, well-decomposed peat is exposed at the surface. Croft and others (2001) reported that mining decreased the number of cultivatable bacteria and decreased the microbial biomass C relative to natural sites, although they did not measure microbial activity other than increased nitrogen (N) mineralization. Glatzel and others (2004) and Waddington and others (2001) reported that peat in mined sites had lower potential (that is, measured in vitro) CO2 and CH4 production than nearby natural sites, presumably due to poor C substrate and nutrient availability for microorganisms. Andersen and others (2006) confirmed that poor substrate quality was the main reason for slow respiration rates, and 3 years after restoration of a vacuum-mined site, respiration rates remained low. Glatzel and others (2004) reported that newly formed peat, 30 years after restoration of a block-cut peatland, had much larger potential CO2 and CH4 production than natural sites, causing them to be net sources of CO2 to the atmosphere (Glatzel and others 2004; T. R. Moore and others, unpublished). Andersen and others (2006) and Croft and others (2001) have begun to uncover underlying controls on microbial communities and their biogeochemical activities and establish linkages with impacts of contemporary vacuum mining and recent active restoration. However, control of microbial greenhouse gas dynamics following block-cut mining has not previously been explored in detail, despite the substantial post-disturbance peat accumulation at these sites (Glatzel and others 2004), in contrast to vacuum harvesting methods that are considerably more challenging to restore regarding hydrologic and vegetation dynamics (for example, Chirno and others 2006; Girard and others 2002; Robert and others 1999).

In this work, we explore the hypothesis that commercial peat mining, post-mining abandonment, and restoration of block-cut peatlands alters C substrate and nutrient availability that leads to changes in microbial biomass and, in turn, microbial CO2 and CH4 exchange. In particular, we predicted that decreased C substrate quality and nutrient availability supports less microbial biomass that leads to slower rates of peat mineralization and CH4 oxidation. To test these hypotheses, we determined C, N, and phosphorus (P) concentrations in the peat and microbial biomass, other extractable ions, and peat organic chemistry using Fourier-transformed infrared (FTIR) spectroscopy and solvent- and water-extraction based organic matter stability indices in 120 samples of peat from soil profiles of two sets of peatlands in eastern Canada where field fluxes of CO2 and CH4 have been measured over three years (T. R. Moore and others, unpublished). We measured potential microbial CO2 production and CH4 production and oxidation. The large range of peat substrate and nutrient availability across the sites provided a unique “natural laboratory” to analyze the importance of controls on microbial greenhouse gas production and consumption in peatlands that are otherwise difficult to discern within a single undisturbed site.

Methods

Study Sites and Sampling

Two regions of commercial mining in eastern Canada were chosen. Four sites were sampled near Rivière-du-Loup, Quebec, a natural site, an actively mined site, a site that has been mined and abandoned for 30 years, and a block-cut mined site that had been restored (Figure 1). In the natural site samples were taken from hummocks dominated by Sphagnum capillifolium, S magellanicum, and Sphagnum russowi, with a shrub overstory of Kalmia spp. and Chamaedaphne calyculata. The site was a dry bog, with summer water tables often below 40 cm depth. We sampled hummocks because regeneration in nearby block-cut sites is generally with hummock forming species and current restoration processes use hummock species. The actively vacuum-mined portion of the peatland was bare of vegetation and the water table was below 70 cm during the summer. The abandoned site was once mined with block-cutting techniques and then abandoned in 1970. It had some Eriophorum spissum tussocks, but samples were taken from regions of bare peat that had an average summer water table depth of approximately 40 cm. The restored site was block-cut until 1975, and drainage ditches closed by natural collapse and then completed mechanically in 1992 to raise the water table (Robert and others 1999). Samples were taken from hummocks in the center of the previously mined trench region dominated by S. capillifolium and Sphagnum angustifolium with an overstory of Kalmia angustifolia. There was approximately 30 cm of newly formed peat on top of a humified old peat profile thicker than 2 m (Figure 1), and average summer water table depth was approximately 35 cm.
https://static-content.springer.com/image/art%3A10.1007%2Fs10021-007-9083-2/MediaObjects/10021_2007_9083_Fig1_HTML.jpg
Figure 1.

Peatlands near Rivière du Loup, QC (left) and Shippagan, NB (right). A Natural bogs. B Actively mined sites. C Mined sites that were abandoned approximately 30 years prior. D Mined sites that were actively re-flooded and restored approximately 30 years prior to sampling. E The peat profiles cut away from hummocks in the restored sites with dotted line marking the previous surface after mining. The alternating red and white segments on the marking stick are 1 cm long.

The second set of sites was located near Shippagan, New Brunswick, and four sites similar to those at Rivière-du-Loup were chosen. Samples at the natural site (Figure 1) were taken from hummocks dominated by S. capillifolium, with a shrub over-story of C. calyculata. The site had less shrub overstory and less pronounced hummock-hollow topography than the Rivière-du-Loup site. The average summer water-table depth was about 30 cm. The actively vacuum mined site at Shippagan was similar to that in Rivière-du-Loup, however there was less woody material in the surface peat. The abandoned site was mined using block-cutting techniques, and then vacuum mined before abandonment in about 1970. Samples were taken from an area of bare peat with an average summer water table depth of approximately 40 cm. As at Rivière-du-Loup, there were sparsely distributed E. spissum plants in this site. The restored site was also previously block-cut and had accumulated approximately 35 cm of new peat above a residual peat deposit of about 1 m since mining ceased in 1970 (Figure 1). Based on nearby trench sites studied by Robert and others (1999), trenches that had been likely partially sealed through natural breakdown were actively blocked completely in 1984. Samples were taken from hummocks dominated by S. capillifolium, S. magellanicum, with K. angustifolia and C. calyculata. The average summer water table depth was approximately 35 cm.

In August 2001, triplicate 10 cm2 by approximately 50 cm deep cores were removed from each site at Rivière du Loup and Shippagan. Immediately following sampling, cores were cut into five 10-cm vertical segments beginning directly below photosynthetic regions of the Sphagnum mosses, where present, and segments were sealed in two heavy-gauge freezer bags, placed on ice in coolers, and transported to the laboratory. Each 10-cm3 segment of peat was homogenized by hand using a knife and scissor and mixed in a sterile plastic bag. Approximately 60 ml of peat was freeze-dried and ground for FTIR spectroscopy and solvent-extractable lipid analysis.

Chemical, Microbial Biomass, and Nutrient Analyses

Solid state FTIR spectroscopy was carried out on a Bruker Vector 22 with the ATR MIRacle technique (Bruker Optics, Ettlingen, Germany). Freeze-dried ground peat (ca. 0.1 g) was placed on the quartz crystal and pressed with a hand compressor to maximize surface area of the sample touching the cell. Absorbance spectra were run in the frequency region of 3,800–450 cm−1 using a resolution of 2 cm−1 and a data interval of 0.2 cm−1. The baselines of the resulting spectra were brought to zero to facilitate the qualitative comparison between samples. Peak ratios were determined semi-quantitatively from peak heights at the wavelengths of 1,050, 1,620 and 1,720 cm−1 (Niemayer and others 1992). A commonly used humification index is the 1,720:1,620 quotient (see Kalbitz and others 1999 for a more detailed review). The index provides information about the relative abundance of protonated COO groups and aromatics, and represents the relative ratio of organic acids to aromatic molecules. The abundance of polysaccharides to aromatic molecules is represented by the 1,050:1,620 quotient, and larger ratios are usually interpreted as characteristic for more easily biodegradable organic matter (Kalbitz and others 1999). A similar method has been recently demonstrated to differentiate organic chemical composition of different peats (Artz and others 2006).

Water extracts (3:1 deionized H2O:peat) were taken from each sample after shaking on an oscillating shaker for 1 h at 200 rpm. Sub-samples of extracts were filtered with 0.45-μm glass-fiber filters and dissolved organic C (DOC) concentrations measured on a Shimadzu 5050 TOC analyzer (Shimadzu, Kyoto, Japan). Extracts were then acidified to pH 2.0 with hydrochloric acid and humic acids were allowed to precipitate at 4°C for 24 h. Samples were filtered to 0.45 μm and DOC concentrations measured. Ratios of humic acids: total extractable DOC were calculated and expressed as percent humic acids in DOC. Humic acids are mainly high molecular weight molecules and compounds with little carbohydrate residue (Paul and Clark 1996) and may be more recalcitrant and resistant to microbial decomposition than non-precipitable fulvic acids.

Lipids extractable with di-ethyl ether (DEE) and chloroform (CHCl3) have been demonstrated to represent bioavailable and biorecalcitrant components of soil organic matter and composts and are commonly used in organic matter stability indices (Dinel and others 2001, 1996a; Dinel and Nolin 2000; Paré and others 1999). Lipids were sequentially extracted with DEE and CHCl3 from 1 g of freeze-dried ground peat using a Dionex 200 Accelerated Solvent Extractor (Dionex, Sunnyvale, CA, USA) at 6.9 Mpa and 100°C. Extracts were dried at 40°C for 2 days on sand and weighed. Ratios of CHCl3: total extractable lipids (the sum of both extracts) were calculated.

Degree of peat decomposition was characterized using the von Post index of humification (Stanek and Silc 1977). The index spans from 1 to 10, where a value of 1 represents intact plant structure with no decomposition, and a value of 10 represents complete humification with dark paste-like peat and no intact plant material. The pH of peat was measured potentiometrically in slurries used to determine aerobic CO2 production. Extractable organic and microbial C, N, and P and inorganic N and P were determined using a CHCl3 fumigation-extraction technique modified from Voroney and others (1993) following Basiliko and others (2006). Water-extractable (above) SO4 was measured with a Metrohm 690 ion chromatograph (IC, Metrohm Ltd. Herisau, Switzerland) after filtration to 0.2 μm with nylon filters and Na, K, Ca, Mg by atomic absorption spectroscopy with a VarianSpectrAA-20. Ion concentrations were expressed per g dry peat. For water-extractable ions and certain other variables, average values for each site were reported in table form without differentiating depths, whereas all data (n=120 for each variable) were used in correlation analyses.

Microbial CO2 and CH4 Exchange

Carbon dioxide and CH4 exchange was measured under aerobic and anaerobic conditions at 20°C, following methods modified from Basiliko and others (2004) and Glatzel and others (2004). Aerobic incubations were sampled initially and after 1 and 2 days for CO2 and CH4 analysis. Anaerobic flasks were incubated for 23 days and sampled two to three times each week for CO2 and CH4 analysis. By 20 days, anaerobic CO2 production rates had declined in certain flasks, presumably due to larger concentrations of CO2 in the headspace. Flasks were again evacuated and flushed with N2 and allowed to incubate for three additional days to capture linear increases in headspace CO2. To investigate the effects of residual roots on potential CO2 production, a second set of aerobic incubations with roots removed were made, and mass of fine (<2 mm diameter) and coarse (>2 mm diameter) roots were measured. Potential aerobic CO2 production of composite samples of roots removed from all samples (n = 3) were also measured. Carbon dioxide and CH4 concentrations were measured with gas chromatography methods (Basiliko and others 2004; Glatzel and others 2004). Production was calculated as the volume-corrected linear increase in CO2 or CH4 over time per g dry peat. Methane consumption was calculated as the volume corrected linear decrease in CH4 over time per g dry peat. The rate of aerobic CO2 production and CH4 consumption over the 2-day incubation and the rate of anaerobic CH4 and CO2 production after 20 days were chosen to represent potential aerobic and anaerobic fluxes, respectively. Moisture contents were determined by oven drying sub samples of each peat sample at 70°C.

Statistical Analyses

Analyses of variance with Tukey post-hoc tests were performed on SYSTAT 10 (SPSS Inc. Chicago, IL, USA) to compare inter-site fluxes, chemical, and microbial characteristics within Rivière du Loup or Shippagan, to similar sites between both regions, and to different depths within a site with site or depth as the factor, and production potential rate or chemical or microbial characteristics as the dependant variable. Characteristics across sites at certain individual depths were also compared. Non-normally distributed datasets (skewness/SE and kurtosis/SE of resulting residuals not sufficiently near zero) were log-transformed and re-analyzed. Methane production rates of zero were substituted with 0.0001 μg CH4 g−1 peat day−1 prior to log-transformation. Links between measured variables were explored with Pearson correlation coefficients with Bonferroni probabilities using SYSTAT 10. Resulting P values less than 0.05 are assumed to represent significant differences. To address issues involved with independent variable co-variation, stepwise multiple regression analysis (forward procedure, F to enter = 4.000, F to remove = 3.996) was used to test the dependence of the aerobic CO2 production, anaerobic CO2 production, CH4 production, and CH4 oxidation rates on the different soil properties in Table 2. The four treatments (natural, mined, abandoned, and restored) were analyzed independently, and then all treatments were analyzed together to find global controls on the CO2 and CH4 dynamics. The two regions were included in the analyses as a dummy variable, and the average soil depth (5, 15, 25, 35 and 45 cm) was a continuous independent variable.

Results

Peat Chemistry and Physical Properties

At both Rivière du Loup and Shippagan among all depths the natural sites had significantly larger ratios of organic acids to aromatics than the mined and abandoned sites but not the restored sites, there were no other differences between sites, and depth patterns were generally not clear (Figure 2). The relative abundance of polysaccharides to aromatics among all depths at Rivière du Loup was significantly larger at the natural site than at all other sites and significantly smaller at the abandoned site than at all others (Figure 2). If the top two depths, consisting of large woody material at the mined site, were excluded, ratios decreased in the order of natural, restored, mined, and abandoned (Figure 2). At surface depths from Shippagan, the restored site had the largest ratios, although they were not significantly greater than the natural site, and mined and abandoned sites had significantly lower ratios at these depths (Figure 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs10021-007-9083-2/MediaObjects/10021_2007_9083_Fig2_HTML.gif
Figure 2.

Fourier-transformed infra-red spectroscopy-derived ratios of carboxyl groups (from organic acids) and aromatic molecules, and polysaccharides and aromatics, the humic acid fraction of extractable DOC, total DEE and CHCl3 sequentially extracted lipids, von Post humification index, microbial biomass C, and total K2SO4-extractable organic and microbial N and P of peat profiles from sites at Rivière du Loup (RDL) and Shippagan (SHP). Sites are: natural bogs (NAT), actively mined (MIN), post-mining abandoned sites without vegetation (ABD), and post-mining, block-cut sites that were actively accumulating peat (RST). Horizontal bars represent standard deviation of 3 replicates.

Total water-extractable DOC concentrations did not vary systematically between sites and depths (data not shown). However the humic acid fraction of extractable DOC was significantly lower in natural and restored than mined and abandoned among all depths at both Rivière du Loup and Shippagan (Figure 2). Differences between natural and restored and mined and abandoned sites were not significant among all depths, but at individual depths restored sites had significantly smaller humic acid fractions than natural, and abandoned had significantly larger fractions than mined sites. Total extractable lipid concentrations were larger at Rivière du Loup than at Shippagan among all sites and depths representing 5 and 2.5% of peat gravimetrically. At Rivière du Loup, lipid patterns were similar to those of the humic acid fraction of DOC, with natural and restored sites having lower lipid concentrations than mined and abandoned sites (Figure 2). Likewise, at certain individual depths, the restored site had significantly smaller concentrations than the natural site. At surface depths at Shippagan, average lipid concentrations increased in the order of restored, natural, and mined and abandoned, although among all depths, differences were not significant (Figure 2). The CHCl3-extractable fraction of total lipids did not illustrate any clear patterns among sites or depths (data not shown).

In natural and restored sites at Rivière du Loup, the von Post humification index illustrated less humified peat at the surface and significantly increasing humification with depth (Figure 2). In contrast, mined and abandoned sites were humified throughout depth (Figure 2). At three of five depths, the abandoned site was more humified than the mined site, although among all depths, the two sites were not significantly different. The Shippagan sites had similar patterns of humification to those at Rivière du Loup, except that differences between natural and restored and mined and abandoned sites were only significant in the surface three depths (Figure 2). Among all depths, pH ranged from 3.7 to 3.9 and was not significantly different among the eight sites (Table 1). Peat at Rivière du Loup was generally drier than at Shippagan, and there were no significant differences between natural, mined, and abandoned sites, whereas the restored site had a larger moisture fraction (Table 1). At Shippagan, natural and restored sites both had significantly larger moisture fractions than mined or restored sites.
Table 1.

Peat Properties among Sites and Regions

 

Rivière du Loup

Shippagan

NAT

MIN

ABD

RST

NAT

MIN

ABD

RST

pH

3.8 (0.1)a

3.9 (0.2)a

3.8 (0.2)a

3.9 (0.2)a

3.7 (0.1)a

3.8 (0.1)a

3.9 (0.1)a

3.8 (0.1)a

Moisture

0.83 (0.04)a,c

0.83 (0.04)a

0.83 (0.04)a,c

0.89 (0.02)b,e

0.93 (0.01)d

0.85 (0.03)a,f

0.86 (0.02)c,e,f

0.92 (0.01)b,d

NO3

30 (8.7)a

27 (5.5)a

38 (12)a

49 (14)b

75 (11)c

30 (5.0)a

30 (4.2)a

35 (11)a

NH4

65 (17)a

100 (43)a

35 (22)a

47 (62)a

66 (72)a

95 (23)a

53 (31)a

137 (121b

Organic N

72 (16)a,d

47 (12)b,c

55 (9.2)b,d,e,f

94 (28)a

77 (9.1)a,f

42 (9.2)c,e

42 (10)c,e

170 (43)g

Microbial N

291 (89)a

13 (12)b

20 (16)b

293 (157)a

202 (171)a

9.6 (7.4)b

24 (19)b

286 (163)a

Organic C:N

17 (4.3)a,c

11 (1.6)b

8.7 (0.5)b

19 (3.4)a,d

25 (3.9)c,e

22 (5.7)d,e,f

19 (4.1)a,f

16 (2.7)a

Microbial C:N

7.8 (0.8)a

21 (11)b,e

14 (5.0)a,b,c

7.1 (0.8)a

10 (2.9)a,d

44 (20)f

18 (8.4)c,d,e

7.6 (1.6)a

PO4

6.8 (6.1)a

2.9 (0.9)a

3.0 (0.7)a

6.7 (2.7)a

8.9 (1.3)a

4.3 (0.9)a

4.7 (0.9)a

36 (24)b

Organic P

4.2 (1.2)a

4.4 (1.2)a

2.8 (1.1)a

5.6 (2.3)a,b

10 (2.8)b,c

6.7 (2.5)a,c

6.1 (1.7)a,c

24 (12)d

Microbial P

110 (26)a

2.1 (1.7)b

8.8 (6.1)b

87 (36)a

37 (33)c

0.42 (1.12)b

1.31 (3.51)b

109 (61)a

Na

0.50 (0.28)a,b,c

0.43 (0.10)b,c

0.35 (0.06)b

0.69 (0.22)a,d

1.03 (0.17)e

0.60 (0.12)c,d

0.77 (0.24)d

1.07 (0.36)e

K

0.11 (0.04)b

0.03 (0.01)a

0.02 (0.01)a

0.16 (0.14)b,c

0.11 (0.07)b

0.03 (0.01)a

0.03 (0.01)a

0.25 (0.18)c

Ca

0.04 (0.02)a,b

0.03 (0.01)a

0.03 (0.01)a

0.04 (0.02)a,b

0.05 (0.02)b

0.04 (0.01)a,b

0.03 (0.01)a

0.04 (0.02)a,b

Mg

0.02 (0.01)ab

0.01 (0.00)a

0.01 (0.00)a

0.02 (0.01)a,b

0.03 (0.01)b,c,d

0.03 (0.01)c,e

0.02 (0.02)b,e

0.02 (0.01)a,b,d

SO4

0.02 (0.02)a,b

0.01 (0.01)a

0.01 (0.00)a

0.04 (0.02)b,c,d

0.04 (0.01)c,e

0.02 (0.01)a,d,f

0.03 (0.01)b,e,f,g

0.04 (0.01)c,g

Values are average pH, moisture fraction of wet peat, K2SO4-extractable and microbial N and P, C:N ratios, and water-extractable ions among all depths at Rivière du Loup and Shippagan. Sites were natural (NAT), actively mined (MIN), once mined and then abandoned (ABD), and restored block-cut mined sites that had new peat accumulation (RST). Concentrations are μg g−1 peat for N and P and mg g−1 peat for ions other than NO3, NH4, and PO4. Standard deviations of three replicates × five depths (n = 15) are in parentheses. Values that are not significantly different (P > 0.05) between sites are indicated with the same superscripted letter.

Concentrations of total extractable organic N (including microbial N) at both Rivière du Loup and Shippagan were largest at natural and restored sites among all depths, were five and seven times larger than at mined and abandoned sites and differences were significant (Figure 2). At Shippagan, the restored site had significantly larger concentrations than the natural site, and both sites had a clear pattern of decreasing concentrations with depth (Figure 2). Patterns of total extractable P were similar to total extractable N in both regions (Figure 2), although average concentrations in mined and abandoned sites were 12 and 17 times smaller than natural and restored sites at Rivière du Loup and Shippagan, respectively.

Inorganic nutrients extracted in K2SO4 were occasionally in larger concentrations in natural or restored sites, although patterns were not consistent between nutrients or regions. The natural site at Shippagan had significantly larger NO3 concentrations than all other sites, whereas the restored site had larger NH4 and PO4 concentrations (Table 1). Extractable organic N concentrations at Rivière du Loup were significantly larger at the restored site than at mined or abandoned sites, although the natural and abandoned sites were not significantly different (Table 1). At Shippagan, the restored site had the largest concentrations, significantly greater than the natural site, whereas mined and abandoned sites had significantly lower concentrations than either natural or restored sites (Table 1). Extractable organic P was not significantly different between sites at Rivière du Loup, whereas the restored site at Shippagan had significantly larger concentrations than natural, mined, and abandoned sites (Table 1). Of water-extractable ions, only K concentrations were significantly larger in natural and restored than mined and abandoned sites in both regions, and was significantly larger in the restored site at Shippagan than all other sites at Shippagan (Table 1). Among all sites and depths, Na concentrations were larger at Shippagan than Rivière du Loup and the restored site at Rivière du Loup and restored and natural sites at Shippagan had larger concentrations than both mined and abandoned sites within each region (Table 1).

Microbial Biomass

At both Rivière du Loup and Shippagan, microbial biomass C was largest at natural and restored sites, and among all depths these sites were not significantly different (Figure 2). However at certain depths, the natural site at Rivière du Loup had larger microbial C than the restored site and at Shippagan the restored site had larger microbial C than the natural. Microbial C at natural and restored sites exhibited concentration-depth dependence, with the smallest average values at the lowest depth. At both locations, mined and abandoned sites had less than 1 mg microbial C g−1 peat, were not significantly different within or between locations, and did not exhibit a pattern of decreasing microbial C with depth (Figure 2). Microbial biomass N concentrations in natural and restored sites at both Rivière du Loup and Shippagan were at least ten times greater than in mined and abandoned sites and differences were not significant between both natural and restored or mined and abandoned sites (Table 1). Microbial biomass C, N, and P were strongly and significantly correlated (Table 2).
Table 2.

Relationships between Peat Properties and Microbial Biomass and Activities

Peat property

Microbial biomass C

Aerobic CO2 production

Anaerobic CO2 production

Aerobic CH4 oxidation

Anaerobic CH4 production

Moisture content

0.322

0.342

0.181

0.662

0.545

PH

−0.068

−0.002

0.037

0.279

−0.087

Carboxyl:aromatic

0.366

0.240

0.268

−0.202

−0.216

Polysaccharide: aromatic

0.432

0.382

0.296

−0.160

−0.091

H2O-extractable DOC

−0.067

−0.126

−0.117

−0.150

0.068

Humic fraction of DOC

−0.738

−0.757

−0.586

−0.441

−0.471

Total lipids

−0.496

−0.641

−0.513

−0.444

−0.495

CHCl3:total lipids

0.094

−0.069

−0.068

−0.178

−0.209

von Post humification

−0.780

−0.788

−0.655

−0.233

−0.310

Total extract. organic N

0.952

0.888

0.751

0.173

0.517

Total extract. organic P

0.906

0.896

0.774

0.105

0.472

Microbial C

1.000

0.849

0.667

0.126

0.396

NO3

0.320

0.018

−0.038

0.493

0.230

NH4

−0.121

0.046

0.062

−0.031

0.077

Organic N

0.579

0.659

0.519

0.303

0.683

Microbial N

0.964

0.849

0.729

0.115

0.408

PO4

0.597

0.692

0.428

0.259

0.496

Organic P

0.542

0.607

0.384

0.381

0.626

Microbial P

0.902

0.878

0.783

0.050

0.411

Na

0.506

0.462

0.264

0.548

0.506

K

0.781

0.821

0.622

0.161

0.389

Ca

0.423

0.331

0.283

0.446

0.369

Mg

0.147

0.277

0.156

0.292

0.287

SO4

0.486

0.442

0.334

0.431

0.414

Values are Pearson correlation coefficients between peat moisture, pH, substrate and nutrient chemistry, and microbial biomass C, N, and P and microbial biomass C and potential CO2 and CH4 exchange among all sites and depths. Significant relationships (P < 0.05) are bold.

Microbial CO2 and CH4 Exchange

At Rivière du Loup, the fastest aerobic CO2 production rates generally occurred in the natural site, decreased in the restored site, and were slowest in the mined and abandoned sites. At the Shippagan sites, the fastest rates were measured in the restored site, decreased in the natural site, and were slowest in the mined and abandoned sites (Figure 3). Rates ranged from 0.02 to 0.56 mg COg−1 peat day−1. At both Rivière du Loup and Shippagan, aerobic CO2 production was not significantly different between mined and abandoned sites, although rates between all other sites within each region were significantly different (Figure 3). In both regions, natural and restored sites had the largest potential aerobic CO2 production in the surface peat and rates generally decreased and varied significantly with depth. Root removal resulted in decreased rates of aerobic CO2 production, although production rates were strongly correlated (root-free CO2 production = 0.72 × total CO2 production −0.01, r2 = 0.72, P < 0.001). Potential aerobic CO2 production of fine roots removed from peat was 0.08 mg COg−1 day−1, equal to 40% of total average CO2 production in all bulk peat samples that had roots. Anaerobic CO2 production potentials ranged from 0.06 to 0.83 mg COg−1 day−1 (Figure 3). At both Rivière du Loup and Shippagan, aerobic and anaerobic potentials were significantly and strongly correlated (r2 = 0.70 and 0.75, P < 0.001, respectively).
https://static-content.springer.com/image/art%3A10.1007%2Fs10021-007-9083-2/MediaObjects/10021_2007_9083_Fig3_HTML.gif
Figure 3.

Rates of potential CO2 production in laboratory incubations of peat profiles from sites at Rivière du Loup (RDL) and Shippagan (SHP) under (I) aerobic conditions after two days and (II) anaerobic conditions after 20 days at 20°C and rates of potential CH4 flux in laboratory incubations of peat profiles under (I) aerobic conditions (consumption) after 2 days and (II) anaerobic conditions (production, note log scale) after 20 days at 20°C. Sites are: natural bogs (NAT), actively mined (MIN), post-mining, abandoned, block-cut sites without vegetation (ABD), and restored, block-cut sites that were actively accumulating peat (RST). Horizontal bars represent standard deviations of 3 replicates.

Anaerobic CH4 production ranged from 0 to 476 μg CHg−1 peat day−1 (Figure 3). At Rivière du Loup, natural and restored sites had significantly faster rates than the mined and abandoned sites and the two lowest depths in the restored site had significantly faster rates of CH4 production than the natural site (Figure 3). At Shippagan the restored site had significantly faster rates than the natural site, and both sites had greater potential production than the mined and abandoned sites (Figure 3). The natural and restored sites at Shippagan had significantly greater potential production, with rates up to three and four orders of magnitude faster, than at Rivière du Loup. Potential aerobic CH4 consumption rates ranged from 21 to 140 μg CHg−1 peat day−1 (Figure 3). At Rivière du Loup, the restored site had significantly faster rates than all other sites, which were not significantly different from each other. At Shippagan, the natural and restored sites had significantly faster rates than the mined and abandoned sites. Natural and restored sites at Shippagan had greater CH4 oxidation potential than similar sites at Rivière du Loup, whereas there were not significant differences in mined or abandoned sites between the two regions. Sites with greatest CH4 consumption potentials (restored site at Rivière du Loup and natural and restored sites at Shippagan) also had greatest CH4 production potentials and among all sites CH4 consumption correlated positively and significantly with CH4 production.

Interactions Among Variables

Among all sites and depths, microbial biomass C correlated most strongly and positively with total extractable N and P and K and negatively with von Post humification, humic acid fraction of DOC, and total lipids (Table 2). Microbial CO2 production was also strongly and significantly correlated with microbial biomass C (Figure 4A for aerobic). It exhibited a threshold relationship, particularly with biomass C:N, but also N:P to a lesser extent, where production was constrained to slow rates above ratios of 8 and 3, respectively (Figure 4B, C). Anaerobic CO2 production potential correlated strongly with the same variables, although in all cases, values were slightly closer to zero than with aerobic potentials (Table 2). Microbial biomass and CO2 production correlated significantly with many of the same peat chemical and physical properties (Table 2). Microbial biomass C, N, or P did not correlate significantly with CH4 production or oxidation (Table 2). CH4 production varied positively and most strongly with organic N and P, moisture content, and Na and negatively with total extractable lipids and the humic acid fraction of extractable DOC (Table 2). Potential CH4 oxidation correlated significantly, positively, and most strongly to peat moisture fraction, Na and NO3 and negatively to total lipids and the humic acid fraction of DOC (Table 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs10021-007-9083-2/MediaObjects/10021_2007_9083_Fig4_HTML.gif
Figure 4.

Scatter plots of microbial biomass (A) C, (B) C:N, and (C) N:P and aerobic CO2 production rates across all sites and depths at Rivière du Loup (black symbols) and Shippagan (grey symbols).

In many cases there was significant co-variance between variables with which gas flux potentials were significantly correlated. For example, total extractable N and P and K were strongly positively correlated and correlated with microbial C, whereas Na was correlated positively with moisture content and negatively correlated with lipid concentrations and the humic acid fraction of DOC (results not shown). No single factor determining the rate of aerobic and anaerobic CO2 production could be identified by means of multiple regression analysis. Rather, the rate of these processes appeared to be dependent on a number of the soil properties. However, microbial biomass (C and P) and factors related to soil organic matter quality (for example, N and P content, humic fraction of DOC, von Post humification index, and so on) were consistent determinants of the CO2 production rate (Tables 3, 4). The exceptions were the actively mined and restored block-cut sites. At the actively mined sites Mg was the only soil property significantly related to the aerobic CO2 production rate, whereas both the Mg and organic P concentrations determined the anaerobic CO2 production (Table 3). At the restored block-cut sites, aerobic CO2 production was positively dependent on the Mg, PO4, and K concentration (Table 3). Generally, the stepwise regression analysis suggested that the same factors, that is, the size of the microbial biomass and the quality of organic matter, or in the case of mined sites, Mg, control the rate of aerobic and anaerobic CO2 production. This conclusion is supported by the strong correlation between aerobic and anaerobic CO2 production
Table 3.

Stepwise Regression among Land-Use “Treatments”

Treatment

Dependent

Factors in model

Coefficient

R2

P

Natural

Aerobic CO2 production

von Post

−0.801

0.765

<0.001

NO3

−0.472

  

Anaerobic CO2 production

Microbial P

0.766

0.571

<0.001

Aerobic CH4 oxidation

Region

0.804

0.587

<0.001

H2O- extractable DOC

−0.358

  

Log10 CH4 production

Region

0.860

0.669

<0.001

H2O- extractable DOC

−0.327

  

Mined

Aerobic CO2 production

Depth

0.560

0.480

<0.001

Mg

0.429

  

Anaerobic CO2 production

Mg

0.519

0.672

<0.001

Organic P

0.487

  

Aerobic CH4 oxidation

Ca

−1.34

0.445

<0.001

Na

1.06

  

Log10 CH4 production

Ca

−1.34

0.438

<0.001

Na

1.11

  

Abandoned

Aerobic CO2 production

Polysaccharide: aromatic

−0.726

0.892

<0.01

Na

0.894

  

Microbial P

0.457

  

CHCl3:total lipids

0.222

  

Anaerobic CO2 production

Humic fraction of DOC

−0.735

0.621

<0.001

Aerobic CH4 oxidation

Depth

0.856

0.743

<0.001

von Post

−0.584

0.743

<0.001

Moisture content

0.354

  

Total lipids

0.369

  

Log10 CH4 production

CHCl3:total lipids

−0.343

0.727

<0.001

Moisture content

0.391

  

Depth

0.418

0.727

<0.001

Restored

Aerobic CO2 production

K

0.508

0.876

<0.001

PO4

0.479

  

Mg

0.200

  

Anaerobic CO2 production

Total extractable organic N

0.439

0.725

<0.001

CHCl3: total lipids

−0.367

  

SO4

0.338

  

Aerobic CH4 oxidation

Moisture content

0.581

0.312

<0.01

Log10 CH4 production

Moisture content

0.546

0.270

<0.01

Values are stepwise regression model parameters for natural, mined, abandoned, and restored block-cut peatlands across Rivière du Loup and Shippagan.

Table 4.

Stepwise Regression across All Sites

Dependent

Factors in model

Coefficient

R2

P

Aerobic CO2 production

Total extractable organic P

0.566

0.882

<0.001

K

0.214

  

Humic fraction of DOC

−0.218

  

NH4

0.108

  

Carboxyl:aromatic

0.086

  

Anaerobic CO2 production

Total extractable organic P

1.173

0.852

<0.001

Humic fraction of DOC

−0.421

  

NH4

0.107

  

PO4

−0.212

  

Microbial C

−0.427

  

H2O-extractable DOC

0.181

  

Aerobic CH4 oxidation

Moisture content

0.520

0.512

<0.001

H2O- extractable DOC

−0.316

  

Ca

0.284

  

Polysaccharide: aromatic

−0.226

  

Log10 CH4 production

Moisture content

0.706

0.513

<0.001

H2O-extractable DOC

−0.220

  

Values are stepwise regression model parameters across all sites (natural, mined, abandoned, and restored block-cut peatlands) and regions (Rivière du Loup and Shippagan).

Soil moisture content and H2O-extractable DOC were the main factors determining the rate of CH4 production (multiple regression, Table 4). Larger moisture content consistently resulted in faster production rates (Tables 3, 4). The only exceptions were the mined sites, where Na and Ca alone could explain variations in the CH4 production rate, and the natural sites where the CH4 production rate was negatively related to the H2O-extractable DOC concentration (Table 3). Contrary to the soil moisture content, the negative relationship between H2O-extractable DOC concentrations and CH4 production rates was not as consistent when the treatments were analyzed separately. When the four treatments were analyzed separately the negative relationship between CH4 production and H2O-extractable DOC concentrations was not apparent, except at the natural site where this was the only soil property that was related to the CH4 production rate. Generally, mining, abandonment, and restoration resulted in alterations of the factors controlling CH4 production. This was indicated by the lack of relationship between soil moisture content and the CH4 production rate at the natural sites while the same relationship was found at abandoned and restored sites (Table 3). As well, the strong relation between Na and Ca and the CH4 production rate at actively mined sites was not found at other sites (Table 3). The CH4 oxidation rate was dependent on the same soil properties as the CH4 production rate, regardless of site (Tables 34).

Discussion

Land-Use Change Effects and Inter-Region Variability

Mining led to decreased CO2 and CH4 production in the near-surface peat and reduced the depth-dependency of rates, similar to previous studies by Glatzel and others (2004) and Waddington and others (2001). Potential CO2 production was rarely significantly different between mined and abandoned sites. Mining exposes well-decomposed peat that is thousands of years old to the atmosphere, and 30 years of post-mining abandonment did not generally affect decomposition patterns of this already very recalcitrant peat. Patterns between both abandoned sites were similar, despite differences in mining method. However it is important to acknowledge that comparisons among sites included many other important potential variables that we could not characterize including pre ‘treatment’ differences in vegetation and peat substrate chemistry and hydrology, variable lengths of time between harvesting and work in the present study, differences in mining methods in mined sites and one of the abandoned and the restored sites, and in the case of the restored sites, time since restoration.

Across 13 sites near Rivière du Loup, Glatzel and others (2004) reported that restoration, including formation of new Sphagnum peat, led to larger CO2 and CH4 production potentials than at nearby natural sites. In the present study, the restored site at Rivière du Loup had smaller CO2 production potentials than at the natural site, which was dry, perhaps partially owing to the surrounding, drained and mined area, and thereby leading to enhanced decomposition. At Shippagan, the restored sites had the largest aerobic CO2 production encountered in the study, 50% larger than at the natural site. Large production potentials indicate peat that is more bioavailable and less likely to contribute to long-term C sequestration, particularly as rates of photosynthetic CO2 uptake were not exceptionally fast at this site (T. R. Moore and others, unpublished). These contrasting patterns provide evidence that successful restoration of organic matter stability in a previously mined peatland may depend on physical and chemical site characteristics (see below), and that geographic variability in the ability to restore a site may be large. It is important to recognize that this interpretation is based on two sets of sites with three sampling locations in each and that we did not characterize potential differences that could have existed among sites prior to mining. It is also important to consider the role of potentially varying soil bulk densities among sites when extrapolating the present findings to the field scale. For example, more rapid rates of CO2 production per gram of peat in the restored and natural than mined and abandoned sites may not directly translate to more rapid CO2 efflux in the field if bulk density in the former is substantially smaller.

Potential aerobic and anaerobic CO2 production rates were strongly correlated, which has been suggested to result from similar peat properties controlling the activities of both aerobic and anaerobic microbial communities (Glatzel and others 2004). In many samples, potential anaerobic CO2 production was larger than aerobic, in contrast to all previous work comparing production rates under aerobic and anaerobic conditions (for example, Glatzel and others 2004; Moore and Dalva 1997; Yavitt and others 1997). Evacuation of anaerobic flasks after 20 days would have eliminated any product concentration limitations to anaerobic CO2 production and likely led to rapid rates. This finding supports Blodau and Moore (2003a, b), who reported that incubation methods could play a profound role in determining absolute magnitude of production. Caution should be used in interpreting magnitudes determined under different conditions, although these methods can provide important information about relative rates of decomposition between samples incubated under the same conditions.

Restored sites at both Shippagan and Rivière du Loup had the largest CH4 production potentials within regions, in agreement with Glatzel and others (2004). Drier conditions at the natural site than at the restored site could have limited methanogen biomass and activity in situ and resulted in lower production potentials even after 23 days of anaerobic incubation. At Shippagan, peat from the restored site produced CH4 up to 3 orders of magnitude faster than the natural site, despite no differences in moisture content, perhaps suggesting that larger CH4 production potentials resulted from enhanced substrate and nutrient availability. However stepwise regression analysis more consistently implicated moisture content as the most important control on CH4 production. That restored sites at Shippagan had substantially greater CH4 production potentials than natural sites is at least in part due to altered peat substrate and nutrient availability and indicates that after 30 years of vegetation restoration these sites could be larger sources of CH4 than natural sites. Methane consumption potentials were generally largest in the restored sites at each region with the largest CH4 production potentials, however differences in CH4 oxidation potential were much smaller between sites than CH4 production potentials. Although it is speculative to compare rates of CH4 production and consumption, because consumption assays involved the addition of substrate, smaller differences in potential consumption than production between the natural and mined sites may imply that enhanced production in the restored sites may result in larger CH4 fluxes to the atmosphere. Indeed a concomitant study of in situ CH4 fluxes indicated that the block-cut restored site had the largest efflux in the region (T. R. Moore and others, unpublished).

In Sphagnum peat, relative abundances of organic acids to aromatic compounds may not necessarily be related to microbial activity. Aromatic compounds are generally difficult to decompose and require unique oxidative enzymes. Organic acids, however, can include straight chain organic acids derived from fermentations, but also phenolic and uronic acids (Verhoeven and Liefveld 1997). The relative abundance of polysaccharides to aromatics may be a better index of bioavailability, and in the surface peat at Shippagan, the restored and natural sites had larger ratios, whereas the mined and abandoned sites had low ratios, consistent with patterns of CO2 production potentials. This pattern was similar at Rivière du Loup, with the exception of the mined site. The surface 20 cm at the mined site contained more coarse woody material, which is relatively rich in polysaccharides (Paul and Clark 1996), than any of the other sites.

Dissolved organic C may be an important source of C for microbial mineralization (Fenchel and others 1998), and the bioavailability of DOC could play a role in CO2 and CH4 production in peat. Mining resulted in surface peat that had a greater fraction of humic acids in DOC, and at certain depths, abandonment led to even larger fraction in both regions. At Rivière du Loup, the restored site had the smallest fraction of humic acids, perhaps supporting large CH4 production potential, although not CO2, where potentials were significantly larger at the natural site. The fraction of humic acids likely is an important control on peat mineralization, as it repeatedly appeared as a significant dependent variable in the stepwise regression analysis. The patterns of von Post humification were very similar to the humic acid fraction of extractable DOC, and had similar influence on CH4 and CO2 production rates.

Bioavailability indices based on DEE- and CHCl3-extractable lipids have not previously been applied to Sphagnum peat, although they have been applied successfully to a variety of other soils and composts to rapidly assess changes in organic matter stability through decomposition and as a result of land-use history (Dinel and Nolen 2000; Dinel and others 1996a). Molecules sequentially extracted with CHCl3 after DEE have been characterized as biorecalcitrant (see Dinel and others 2001, 1996b), and a larger fraction of CHCl3-extractable: total-extractable lipids corresponds to greater stability. The index could not demonstrate land-use change effects in the present study and generally did not relate to microbial activity or other bioavailability indices. Total-extractable lipids, however, increased with mining and abandonment and correlated to significantly lower microbial activity, in contrast to previous reports. It is likely that the organic chemistry of peat is unlike other soils and composts for which the method was developed. The total lipid fraction in peat represents either a recalcitrant fraction of plant-derived waxes left in greater concentration as labile material is mineralized or microbial metabolites that are resistant to decomposition (Moore and Basiliko 2006; Verhoeven and Liefveld 1997).

Concentrations of microbial biomass C were very low in mined and abandoned sites, with as little as 5% of biomass present in natural and restored sites at certain depths. Croft and others (2001) measured microbial biomass C and cultivable numbers of bacteria in natural, mined, and restored peatlands and reported that mining decreased microbial biomass substantially, and restoration caused increases, although restored sites were more similar to mined than natural sites. Andersen and others (2006) monitored the physicochemical and microbiological status of a Quebec bog restored after three years relative to natural and mined sites. It was shown that many nutrient and microbial biomass concentrations in a restored were more similar to a mined than a natural site, except for NH4, P, and K. In the present study, restoration returned concentrations to natural values, and at certain depths at Rivière du Loup, the natural site had larger microbial biomass C concentrations than the restored site, and at Shippagan, the restored site had larger biomass C concentrations than the natural site, consistent with patterns of CO2 production potentials. Differences in effects of restoration between the present study and Andersen and others (2006) and Croft and others (2001) clearly arise from the methods of mining (block-cutting vs. vacuum) and duration of time since restoration. Study of restoring vacuum mined sites in Canada and subsequent operational restoration endeavors, although progressing at a rapid rate, are relatively recent. As a result, there are not yet restoration examples of vacuum-mined sites in eastern Canada with substantial post-mining Sphagnum peat accumulation equivalent to restored block-cut sites. It is quite likely that beyond importance of time since restoration, the mining methods are inherently different regarding ease of restoring hydrologic, microclimatic, and vegetation dynamics (Lavoie and others 2005) that might eventually lead to subsequent restoration of microbial biogeochemistry. Therefore the present study compliments more than contrasts Andersen and others (2006) and Croft and others (2001), owing to different mining and restoration techniques.

Linking Peat Properties to Exchange Potentials

Among all samples, many physical, chemical, and biological properties that correlated strongly with exchange potentials co-varied, making precise characterization of peat-property controls on microbial activity difficult. Therefore we used stepwise regression analysis to help determine the most potentially important controlling factors. The most traditionally accepted determinants of CO2 production in soils are water availability, temperature, and microbial activity (Paul and Clark 1996). However, in laboratory assays of CO2 production, limitation by water and temperature are relieved, and variations in activity per biomass unit are minimized. Under those conditions, the size of the microbial biomass and the quality of the substrate they are growing on become the main determinants of CO2 production. The observation that variations in microbial biomass was one of the main causes for variations in the CO2 production rate is therefore in agreement with the literature that suggests that the size of the microbial biomass, its activity, and the quality and quantity of organic matter are the main determinants for the rate of C and N turnover (compare Moore and Basiliko 2006). It is also supported by the fact that natural and restored sites, which had significantly faster CO2 production rates, also had significantly larger biomass. This supports our hypothesis that decreased resource availability supports smaller microbial biomass and in turn slower rates of activity.

In addition to the microbial biomass and the quality of soil organic matter, the concentration of one or several inorganic nutrients was significantly related to potential CO2 production. Inorganic nutrients are generally poor indicators of nutrient availability in peatlands (Basiliko and others 2006, 2005). Accordingly, relationships between inorganic nutrient concentrations and the rate of aerobic CO2 production were inconclusive, suggesting that inorganic nutrient concentrations were not the major factor determining aerobic CO2 production rates. The exception might be the mined and restored sites, where Mg seems to be important in regulating the rates. Mg is required for the activity of many enzymes, and especially those involved in phosphate transfer (Madigan and others 1997). Across sites and treatments CO2 production was positively related to the concentration of P, which might suggest that P is a limiting factor for CO2 production. If P is a limiting factor and mining removes Mg, which is needed for the functioning of the enzymes involved in P metabolism, it is logical to find a positive relationship between the Mg concentration and CO2 production rates as found at disturbed sites in this study. These results seem to suggest that mining might increase the effect of nutrient limitation by removing cations necessary for the functioning of the enzymes involved in the metabolism of that nutrient, and not only by removing the nutrient itself. Our conclusion of direct or indirect P limitation to microbial activity contrasts with a report by Thormann and others (2001), who demonstrated a strong negative correlation between decomposition of S. fuscum plants and total P in bog water, although the method of measuring decomposition was different from the present study. Nitrogen and P are potentially limiting nutrients to microbial production and activity in peatlands, although previous work has not conclusively identified the mechanisms of limitation, with many nutrient addition studies yielding conflicting results (compare Basiliko and others 2006).

Although in most cases, stepwise regression analysis implicated moisture and anaerobism and certain cations as the primary controls on CH4 production, the ratio of polysaccharides to aromatics also likely played a role, albeit less important at Shippagan. Specific organic C substrates are commonly implicated as the major chemical control of CH4 production in the absence of aerobic conditions (Yavitt and others 2000, 1997; Yavitt and Lang 1990). Polysaccharides and organic acids could serve as substrates for fermentative bacteria and could supply products for methanogenesis. The lack of correlation between the microbial biomass and CH4 production and consumption probably occurred because methanogen and CH4-oxidizing bacterial biomass represent a very small fraction of total soil microbial biomass. The CH4 oxidation potential rate was dependent on the same soil properties as the CH4 production rate and CH4 production and consumption were correlated. This supports observations by, for example, Moore and Dalva (1997), and suggests that the CH4 oxidizers were substrate limited. High moisture content stimulates CH4 production probably by creating anaerobic microsites. The CH4 then diffuses to oxic microsites where it is consumed by methane oxidizers (Blodau and Moore 2003a, b). Moisture content was not the main controlling factor on methane dynamics in natural and mined sites. In the natural environment, DOC seems to be the main factor controlling the potential rates, with larger concentrations resulting in slower rates. Fractionation of DOC to humic and fulvic acids is only a very coarse separation and both fraction could potentially contain ranges of compounds that are relatively recalcitrant to, or potential inhibitors of, methanogenesis. Nevertheless, it is counterintuitive that this relationship existed across the natural sites; with total DOC more dominated by fulvic acids that could in principle serve as C or redox substrate and are less likely to inhibit methanogenesis than mined or abandoned sites. At the mined sites, Na seems to limit both CH4 and CO2 dynamics, in agreement with Basiliko and Yavitt (2001) who found Na to stimulate CH4 production. Sodium limitation of CH4 production, taken together with Mg limitation of CO2 production at the mined sites, seems to suggest that the effect of removing cations necessary for the function of key metabolic enzymes or processes might be as strong or stronger than the direct effect of removing potentially limiting macronutrients such as N and P.

Conclusions

Microorganisms play a key role in the flux of trace gases between the atmosphere and soils, but determining the environmental and edaphic controls on microbial gas exchanges is complex. Methodological challenges present a significant obstacle in measuring microbial activities that relate to patterns at field-scales and accurately characterizing the soil physical and chemical environment relevant to microbial communities involved in decomposition and greenhouse gas exchanges (compare Madsen 1998). We examined potential aerobic and anaerobic microbial greenhouse gas exchange of a wide range of peat samples representing different profiles at natural, mined, abandoned, and restored peatlands and used a large suite of techniques to characterize the nutrient and substrate chemistry of these soils. We present detailed analyses of microbial activities in peat profiles accumulated post-mining. Mining and abandonment led to reduced nutrient and substrate availability and occasionally drier conditions in surface peat resulting in drastic reduction in CO2 and CH4 production, in agreement with previous studies. Owing mainly to wetter conditions, CH4 production and oxidation were faster in restored relative to natural sites, whereas in one restored site increased substrate and nutrient availability led to much more rapid rates of CO2 production. This appears to imply that, for block-cut sites, restoring hydrologic dynamics to a more natural state might reestablish microbial CH4 dynamics, although there is geographic or site-specific variability that may influence the successfulness of restoring bulk peat decomposition. The present work compliments recent study of restoration efforts in vacuum-mined peatlands, however a longer time period until restoration (that is, water table fluctuations occurring within an acratelm accumulated post-mining) will postpone direct comparison of restoration effectiveness regarding microbial greenhouse gas dynamics. In the future, comparative studies of these mining methods may guide peat resource management and select practices that reduce the net atmospheric greenhouse gas burden. The sites we examined presented an exceptionally large range of soil C substrate quality, nutrient availability, microbial biomass, and microbial activities, allowing us to draw more global conclusions about controls on microbial CO2 and CH4 dynamics using stepwise regression analysis among all sites and soil depths. In summary, aerobic and anaerobic decomposition of bulk peat was constrained by organic matter quality, particularly P and C chemistry and closely linked to the size of the microbial biomass supported by these limiting resources. Methane production was more dominantly controlled by field moisture content (a proxy for anaerobism), even after 20 d of anaerobic laboratory incubation, and to a lesser extent by C substrate availability. Methane oxidation was controlled by the same factors influencing CH4 production, leading us to conclude that CH4 oxidation is primarily controlled by CH4 substrate availability.

ACKNOWLEDGMENTS

We thank Mike Dalva, Helénè Lalande, Dr. Stephan Glatzel, Michelle Marinier (McGill University), Louise Florent, Eglantine Imbeault, Dr. Judith Frégeau-Reid, and Dr. Henri Dinel (Agriculture and Agrifood Canada- Eastern Cereal and Oilseed Research Centre) for intellectual and technical support. Premier Horticulture and Sun Gro Horticulture graciously allowed site access. We are grateful for funding from the Natural Sciences and Engineering Research Council of Canada (TRM), McGill University (NB), and the Deutsche Forschungsgemeinschaft (CB). The comments of Mike Waddington and an anonymous reviewer greatly improved the manuscript.

Copyright information

© Springer Science+Business Media, LLC 2007