Hydrobiologia (2012) 692:41–55 DOI 10.1007/s10750-011-0934-7
WETLAND SERVICES AND MANAGEMENT
Emissions of CO2, CH4 and N2O from undisturbed, drained and mined peatlands in Estonia Ju¨ri-Ott Salm • Martin Maddison • Sille Tammik • Kaido Soosaar • Jaak Truu ¨ lo Mander U
Received: 8 December 2010 / Accepted: 19 October 2011 / Published online: 4 November 2011 Springer Science+Business Media B.V. 2011
Abstract The aim of this study is to estimate emissions of greenhouse gases CO2, CH4 and N2O, and the effects of drainage and peat extraction on these processes, in Estonian transitional fens and ombrotrophic bogs. Closed-chamber-based sampling lasted from January to December 2009 in nine peatlands in Estonia, covering areas with different land-use practices: natural (four study sites), drained (six sites), abandoned peat mining (five sites) and active peat mining areas (five sites). Median values of soil CO2 efflux were 1,509, 1,921, 2,845 and 1,741 kg CO2-C ha-1 year-1 from natural, drained, abandoned and active mining areas, respectively. Emission of CH4-C (median values) was 85.2, 23.7, 0.07 and 0.12 kg ha-1 year-1, and N2O-N -0.05, -0.01, 0.18 and 0.19 kg ha-1 year-1, respectively. There were significantly higher emissions of CO2 and N2O from abandoned and active peat mining areas, whereas CH4 emissions were significantly higher in natural and drained areas. Significant Spearman rank correlation was found between soil temperature and CO2 flux at all sites, and CH4 flux with high water level at Guest editor: Chris B. Joyce / Wetland services and management J.-O. Salm (&) M. Maddison S. Tammik ¨ . Mander K. Soosaar J. Truu U Department of Geography, Institute of Ecology and Earth Sciences, University of Tartu, 46 Vanemuise St., 51014 Tartu, Estonia e-mail: [email protected]
natural and drained areas. Significant increase in CH4 flux was detected for groundwater levels above 30 cm. Keywords Carbon dioxide Drainage Greenhouse-gas emission Methane Nitrous oxide Peat extraction
Introduction Natural peatlands are important sinks of atmospheric carbon dioxide (Moore, 1994). Peatlands play an important role in carbon accumulation and long-term sequestration, and undrained peatlands are a net sink for C (Minkkinen et al., 2002). This property is also attributed to ecosystem services in the form of carbon storage, with nearly 30% of the soil carbon pool being found in northern peatlands (Gorham, 1991), stored at an estimated rate of 23 g C m-2 year-1 (Gorham, 1995) (19.4 g C m-2 year-1 in western Canada; Vitt et al., 2000; 18.5 g C m-2 year-1 in Finland; Turunen et al., 2002). Peatlands contain the equivalent of 40–60% of the roughly 730 Gt of carbon currently held in the atmosphere as CO2 (IPCC, 2001). There have been several investigations in order to estimate greenhouse-gas (GHG) fluxes and the C balance on national scale, e.g. in Finland (Minkkinen et al., 2002; Alm et al., 2007; Saarnio et al., 2007), Sweden (Nilsson et al., 2001; von Arnold et al., 2005a, b) and North America (Bridgham et al., 2006). An important factor is the estimation of the
proportion of emissions from peat mining areas, including life-cycle analysis to determine the net GHG emissions from the peat industry (Cleary et al., 2005; Kirkinen et al., 2007). No investigations based on field measurements of GHG from peatlands have yet been performed in Estonia. The total area of peat-covered land in Estonia is about 1,010,000 ha (Orru, 1992), corresponding to 22.4% of the country’s mainland. This area includes both drained and natural peats. Approximately 70% of peatlands have been affected by drainage, probably to the extent that peat accumulation has ceased and mineralization of organic matter has replaced carbon accumulation. The total area of Estonian transitional fens and ombrotrophic bogs is estimated to be 339,772 ha, of which at least 51,978 ha has been drained. On the basis of literature data, annual efflux from drained areas was estimated to be 419,000–676,000 t CO2 equivalents (eq) year-1, and -141,000 to 380,000 t CO2 eq year-1 from the undrained area. We calculated the global warming potential if Estonian peatlands were to be hydrologically restored. This is 2.3–2.7 times lower than the present total emission of 278,000–1,056,000 t CO2 eq year-1 (Salm et al., 2009). This excludes peat mining areas, which are mainly established on transitional fens and ombrotrophic bogs. Although their areal coverage is small, there is significant potential for GHG emissions, namely CO2 efflux. According to the revision of peat extraction sites, the area of abandoned and active peat mining sites is 9,371 and 19,574 ha, respectively (Ramst & Orru, 2009). Mining alters the fluxes of CH4 and CO2 (Sundh et al., 2000) and creates sites that are net sources of CO2 (Waddington et al., 2002; Basiliko et al., 2007; Ma¨kiranta et al., 2007) but have reduced emissions of CH4 (Tuittila et al., 2000). Mining activities remove the upper peat layers and expose well-decomposed (Basiliko et al., 2007) and low-substrate-quality recalcitrant peat (Waddington et al., 2001). This provides unfavorable conditions—presumably poor C substrate and nutrient availability—for peat-decomposing microorganisms, as the quality of the available substrate is different from areas under natural conditions (Basiliko et al., 2007). Therefore, the bacterial population has been found to be lower in postvacuum-extracted peatland than in natural peatland (Croft et al., 2001), which may lead to lower CO2 emissions compared with areas in natural condition.
Hydrobiologia (2012) 692:41–55
Gilmer et al. (1998) suggested that oxidation rates decrease to zero after 10–15 years post drainage, due to a decrease in peat substrate quality and supply of labile C over time. Nevertheless, aerobic conditions may support higher CO2 emissions and reduction of CH4 emissions. Emission of N2O from natural mires has been reported to be relatively low, varying on average from 0.01 to 0.3 kg N2O-N year-1 (Martikainen et al., 1995; Nyka¨nen et al., 1995; Laine et al., 1996; Regina et al., 1996; Alm et al., 1999; Minkkinen et al., 2002; Turunen et al., 2002). In a Scottish peatland catchment, terrestrial emissions of CH4 and N2O combined returned only 4% of CO2 eq captured by net ecosystem exchange (NEE) of CO2 to the atmosphere (Dinsmore et al., 2010). Drainage and altered water regime significantly increase N2O emissions from soils (Martikainen et al., 1993), raising the median values from peatlands to 5.6–6.6 (Martikainen et al., 1995; Nyka¨nen et al., 1995; Laine et al., 1996; Minkkinen et al., 2002; Alm et al., 2007) and up to 29 kg N2ON year-1 in some deciduous forests on peat soils (Von Arnold et al., 2005b). The aim of this study is to estimate emissions of GHG CO2, CH4, and N2O, and the effects of drainage and peat extraction on these processes, in Estonian transitional fens and ombrotrophic bogs. Field investigations using the closed-chamber method provide additional information on the effect of drainage and peat extraction on local scale, and the GHG balance in Estonian peatlands.
Materials and methods Study sites Gas fluxes were investigated at four peat extraction sites in eastern and central Estonia, and in natural and drained areas located in central Estonia (Fig. 1; Table 1). The following management types were distinguished: natural mire (N), drained peatland (D), abandoned peat extraction area (A) and active peat extraction area (M). Natural and drained areas were located in two different peatlands in Soomaa National Park: Valgeraba and Kuresoo. Within these areas, nine study sites were established in order to cover areas with different vegetation (e.g. presence of Eriophorum vaginatum),
Hydrobiologia (2012) 692:41–55
Fig. 1 Distribution of transitional fens and oligotrophic bogs and location of study sites in Estonia. 1 Areas affected by drainage; 2 natural areas or areas not included in databases of drained areas
uliginosum, Rubus chamaemorus, Polytrichum strictum and different Sphagnum species.
microreliefs (hummocks, hollows) and effects caused by land use. In Valgeraba there were four study sites: •
Valgeraba 1N and 2N (coordinates N58260 11.8400 , E25140 12.3200 ) is a pristine bog belonging to the wooded hummock bog subtype. Study areas are next to each other and less than 3 m apart. Gas emissions were measured from the hummock microsite and from level patches between plenty of hummocks. Typical plants are Calluna vulgaris, Empetrum nigrum, Andromeda polifolia, Oxycoccus palustris, Pinus sylvestris and different Sphagnum species. Valgeraba 6D and 7D (coordinates N58260 06.2600 , E25140 10.6500 ) belong to the mesotrophic (mixotrophic) bog forest site type. Study areas are situated 30 m apart, and both areas are under strong drainage influence. The area was drained for forestry about 40 years ago. The area is covered with tall pines (Pinus sylvestris) and birches (Betula pubescens). Other typical plants are Calluna vulgaris, Melampyrum pratense, Vaccinium
There were five study sites in Kuresoo: •
Kuresoo 8D (coordinates N58280 26.0900 , E25120 47.4600 ) belongs to the wooded hollow– ridge bog subtype. The area is under strong drainage influence. The study area is in a hollow with very few Sphagnum mosses, but large patches are covered by Rhynchospora alba. The site (also sites 9D and 10D) is located in a drained part (70 ha) of Kuresoo mire (11,000 ha). The area is covered with sparse tree vegetation (Betula sp., Pinus sylvestris) mainly along the ditches apart from measurement site. The area was drained for peat mining purposes in the 1960s and 1970s but activities never started. Kuresoo 9D (coordinates N58280 25.4800 , E25120 46.3300 ) belongs to the wooded hollow–ridge bog subtype and is under strong drainage influence. Gas emissions were measured from hummocks
Site types are numbered and determined based on land use: natural (N), drained (D), abandoned (A) and active (mined) peat extraction (M). Water table levels are measured on monthly basis; average, minimum and maximum levels are presented, indicating water level below the ground surface. Peat decomposition describes the upper 50 cm of the peat profile at the study sites (values according to von Post index)
0.95 0.95 1.05
2.20 2.20 2.15
3–5 1–4 2–5
1.70 Peat depth (m)
1–5 2–5 1–5
5 4 6 0 -3 0 0 Max.
3 25 50 Ave.
20M 19M Water table depth (cm)
Site type no. and code
Number of sites
5D 15A 14A
7D 6D 2N 11N 11A 18M 17M 13A 10D 4N 3N
N 5980 3000 E 27400 500 N 59210 500 E 2760 2300
Kasesoo Hiiesoo Peatland
Table 1 Main characteristics of investigated peatlands
N 58300 1000 E 2590 4400
N 58260 4800 E 25140 2300 N 58190 3500 E 26130 800 N 59150 2100 E 27380 5500
Hydrobiologia (2012) 692:41–55
with sparse vegetation. Typical plants are Calluna vulgaris and Empetrum nigrum, with lichens (Cladonia sp.) also being common. Kuresoo 10D (coordinates N58280 25.4100 , E25120 44.8000 ) also belongs to the wooded hollow–ridge bog subtype, and is under strong drainage influence. Gas emissions were measured from hummocks with Eriophorum vaginatum. Kuresoo 3N (coordinates N58280 28.4300 , E025120 30.2600 ) belongs to the hollow–ridge bog subtype and is an undisturbed area. Gas emissions were measured from a hollow. The study area is in a hollow with lush Sphagnum species. Other typical plants are Drosera anglica, Rhynchospora alba and Scheuchzeria palustris. Kuresoo 4N (coordinates N58280 29.2800 , E025120 30.2800 ) also belongs to the hollow–ridge bog subtype and is an undisturbed area. Gas emissions were measured from hummocks with Calluna vulgaris, Empetrum nigrum, Polytrichum strictum and Sphagnum species.
Peat extraction areas were located in three different peatlands: Kasesoo, Puhatu and Hiiesoo. Ten different study sites were established within these areas. Environmental conditions in these areas were strongly affected by peat extraction activities: removal of vegetation, and establishment of a drainage system to lower the water level (except for the Kasesoo 5D study site, which was classified as a drained area). Some of the patches were abandoned peat extraction sites with different times since abandonment. In peat extraction areas, the vacuum mining method was used. Study sites were chosen according to their usage (a drained bog was used as a reference site, and there were also active peat extraction areas and abandoned peat extraction areas) and the degree of peat decomposition on the surface layer in peat extraction areas. The degree of decomposition was determined according to the von Post method (Stanek & Silc, 1977). In Kasesoo there were three study sites: •
Kasesoo 5D (coordinates N59080 30.400 , E27 400 40.300 ) is a bog affected by drainage. The description of the study area is based on the classification of the Estonian vegetation site types by Paal (1997). This area belongs to the wooded hummock bog subtype group. The vegetation mostly consists of Sphagnum species and species such as Pinus sylvestris, Betula pubescens, Andromeda polifolia, Calluna
Hydrobiologia (2012) 692:41–55
vulgaris, Chamaedaphne calyculata, Eriophorum vaginatum, Oxycoccus palustris, Empetrum nigrum and Drosera rotundifolia. The height of Pinus sylvestris is around 2 m. The thickness of the peat layer is 6.6 m, of which the upper layer of 1.5 m is very slightly decomposed peat. Kasesoo 16M (coordinates N59080 45.500 , E27390 49.100 ) is an active peat extraction area and has no vegetation. The thickness of the peat layer is 2.6 m, of which the upper layer of 0.8 m is almost entirely undecomposed, very slightly decomposed or slightly decomposed peat. This area is under strong drainage influence and under the influence of wind and water erosion. In 2009, the area had not been mined for 20 years when it was once again put into operation. Kasesoo 12A (coordinates N59080 28.300 , E27390 11.200 ) is an area that had been abandoned for 20 years. The area was sparsely covered with trees a few meters high (Betula pubescens, Pinus sylvestris) and Calluna vulgaris, but in the summer of 2009 the vegetation was removed and the drainage system was renovated. The thickness of the peat layer is 3.8 m. In Puhatu there were also three study sites:
Puhatu 17M (coordinates N59150 57.400 , 0 00 E2737 53.3 ) is an active peat extraction area with no vegetation. The thickness of the peat layer is 2.7 m, of which the upper 2.2 m is very slightly decomposed peat. Puhatu 18M (coordinates N59150 38.700 , E27390 04.300 ) is an active peat extraction area with no vegetation and highly decomposed peat. The thickness of peat layer is 0.6 m. The area is under strong drainage influence, and the water level is well below the peat layer. Puhatu 13A (coordinates N59150 57.000 , E27 370 54.400 ) is an inactive peat extraction area that has not been used for 5 years. The area is mainly covered with Calamagrostis epigeios and Calamagrostis neglecta. The thickness of the peat layer is 2.8 m, of which the upper 2.2 m is almost entirely undecomposed or slightly decomposed peat.
In Hiiesoo there were four study sites: •
Hiiesoo 14A (coordinates N59210 06.700 , 0 00 E2707 26.5 ) has been unused for 10 years. The vegetation consists mostly of Calamagrostis
epigeios, and in the hollows there is Eriophorum vaginatum. The thickness of the peat layer is 1.5 m. Hiiesoo 15A (coordinates N59210 30.900 , E27050 55.600 ) has been unused for 25 years. Vegetation is sparse and consists mostly of Betula pubescens, Pinus sylvestris, Calluna vulgaris, Tussilago farfara, Juncus articulatus and Carex flava. The thickness of the peat layer is 1.7 m. Hiiesoo 19M (coordinates N59210 06.700 , E27070 26.500 ) is an active peat extraction area and is under great drainage influence. The thickness of the peat layer is 1.45 m, of which the upper layer of 0.75 m is very slightly decomposed peat. Hiiesoo 20M (coordinates N59210 07.600 , E27060 10.300 ) is an active peat extraction area. The thickness of the peat layer is 2.6 m, of which the upper layer of 1.75 m is almost entirely undecomposed peat or slightly decomposed peat. In Sangla, one study site was chosen for analysis:
Sangla 11A (coordinates N58190 37.3700 , E26 130 5.500 ) belongs to the mesotrophic (mixotrophic) bog forest site type and is under strong drainage influence. The study area is located between an active peat mine (250 m apart) and a cultivated grassland (Phleum pratense, Trifolium sp.) (40 m uphill and separated by a ditch). The area is covered with tall trees (Betula sp., Pinus sylvestris, Populus tremula), some of which were selectively cut a few years ago. Other typical plants are Frangula alnus, Fragaria vesca, Rubus idaeus and Molina caerulea.
Field measurements The closed-chamber method (Hutchinson & Livingston, 1993) was used for measurement of CO2, CH4 and N2O fluxes. Gas samplers [closed chambers with a cover made of polyvinyl chloride (PVC), height 50 cm, Ø 50 cm, volume 65 l, sealed with a waterfilled ring on the soil surface, painted white to avoid heating during application] were installed in five replicates. Gas sampling was carried out on a monthly basis from January to December 2009. Measurements consisted of three gas samples which were collected during 1 h of measurement (at 0, 30 and 60 min). In winter season, with snow coverage (up to 25 cm) chambers were placed on the snow. During each gas
Hydrobiologia (2012) 692:41–55
sampling session at each microsite, the depth of the groundwater table (cm) in the observation wells (Ø 50 mm, up to 1.5 m deep PVC pipes perforated and sealed in the lower 0.5 m part) was determined and soil temperature were measured at four depths (10, 20, 30 and 40 cm). Soil chemical content (total carbon, total nitrogen, total sulphur and total phosphorus), water level, water pH, redox potential and oxygen content were measured at all sites. No statistically significant correlations for soil chemical properties were found, so these data are not included in the results.
mean air temperature during the vegetation period varied from 10.1C to 13.1C in Tartu and Viljandi, and was lower in the Eastern part of Estonia (10.2–13.1C) (Table 2; Fig. 2). In 2009, annual mean air temperatures were slightly higher (0.2–0.3C) than the 30-year average at all weather stations; temperatures at vegetation period were similar as well (only 0.1C higher in Tartu). Yearly precipitation varied from 414 to 1,001 mm during the period 1980–2009, and in the years 2008 and 2009 was significantly higher (104–163 mm) than average precipitation during this period at all sites.
Calculation and statistical analyses
The gas concentration in the collected air was determined using a Shimadzu GC-2014 gas chromatography system (electron capture detector and flame ionization detector; Loftfield et al., 1997) in the laboratory of the Department of Geography, Institute of Ecology and Earth Sciences, University of Tartu, Estonia. Emission rates for one site were calculated as the average of the results of five subsamples taken.
Normality of variable distributions was checked using the Kolmogorov–Smirnov, Lilliefors and Shapiro– Wilk tests. In the case of gas analyses, the distribution differed from normal, and hence non-parametric tests were performed. Medians, 25% and 75% percentiles, and minimum and maximum values of variables are presented. We used the Kruskal–Wallis analysis of variance (ANOVA) test and multiple comparison of mean ranks test to check the significance of differences between gas fluxes for different land-use categories, and the Spearman rank correlation and non-linear regression to analyse the relationship between GHG fluxes and environmental conditions. Statistical analysis was carried out using Statistica 7.1 (StatSoft Inc.). The level of significance of P = 0.05 was accepted in all cases. Additionally, redundancy analyses (RDA) were applied to relate gas emission data to environmental parameters (Legendre & Legendre, 1998). The soil temperature and depth of groundwater data were used in RDA as explanatory variables, and land-use categories were considered as a categorical value. A forward selection procedure with 1,000 permutations
Meteorological data Meteorological analyses are based on air temperature, measured hourly and calculated daily, as well as daily, monthly and annual precipitation from the Meteorology Station of Tartu Observatory (N58150 5500 , E26270 5800 ), corresponding to weather conditions at the Sangla study site (16 km from the station); from Viljandi Meteorology Station (N58220 4000 , E25360 0100 ), corresponding to the Soomaa study sites (25 km from the station); and at Jo˜hvi Meteorology Station (N59190 4400 , E27230 5400 ), corresponding to the Hiiesoo (17 km from the station), Puhatu (16 km from the station) and Kasesoo (25 km from the station) study sites. In the period from 1980 to 2009, the
Table 2 Annual precipitation and mean air temperatures during the vegetation period (April–October) Year
Annual precipitation (mm) Viljandi
Mean annual air temperature during vegetation period (April–October) (C) Jo˜hvi
Based on data from the Meteorology Stations of Viljandi (corresponding to study sites in Soomaa National Park), Tartu Observatory (corresponding to the Sangla study site) and Jo˜hvi (corresponding to Hiiesoo, Puhatu and Kasesoo study sites)
Hydrobiologia (2012) 692:41–55
efflux values were registered in the winter period (January–March and December 2009), when measurements were taken at temperatures below 0C. There was no significant difference between the four site categories (Kruskal–Wallis ANOVA test), and median values of CO2-C fluxes were similar in natural, drained and active peat extraction areas; emissions were highest at abandoned peat extraction sites (Fig. 3). Emissions correlated with soil temperature measured at different (0, 10, 20, 30 and 40 cm) depths (Spearman r = 0.79; P \ 0.05) (Fig. 4). RDA analysis indicated that water level and soil temperature at 10 cm from the ground surface explained 68.9% of CO2-C flux (Fig. 5). Results showed that correlations were similar at drained, abandoned, mined and natural sites, with R2 values between 0.57 and 0.73 (P \ 0.05). The strongest correlation was in natural areas with R2 value of 0.73 (P \ 0.05). As the results of the analysis did not show significant differences, data are combined into one figure (Fig. 4). Median values of cumulative annual soil efflux of CO2-C at natural, drained, abandoned and active extraction sites were 1,509, 1,921, 2,845 and 1,741 kg CO2-C ha-1 year-1, respectively. Fig. 2 Mean monthly air temperature (upper part) and monthly precipitation (lower part) in 2009 at Viljandi weather station (corresponding to Soomaa study sites), Jo˜hvi weather station (corresponding to Kasesoo, Hiiesoo and Puhatu study sites) and Tartu weather station (corresponding to the Sangla study site)
was applied for selection of statistically significant explanatory variables. For RDA, the CANOCO 4.52 program was used. In terms of CO2 emissions, we were unable to estimate NEE values, and soil heterotrophic respiration was measured. However, at active peat extraction sites as well as at abandoned sites with no or very sparse vegetation cover, soil efflux values measured in chambers can be considered to be NEE values.
Results Soil CO2 efflux The averaged soil efflux of carbon dioxide varied between 0 and 138.0 mg CO2-C m-2 h-1. Low CO2
Emissions of CH4 The averaged emissions of CH4-C varied between -82 and 12,037 lg CH4-C m-2 h-1. Negative values indicate methane consumption and were registered from sites affected by peat extraction and from active peat mining sites. There was a significant difference between the site categories (Kruskal–Wallis ANOVA test), with the highest emissions being measured at natural and drained sites (Fig. 6). Comparison of CH4-C fluxes at natural and drained versus abandoned and mined sites showed significant differences between these two land-use groups (multiple comparison of mean ranks; P \ 0.05). Emissions correlated negatively with water level depth (Spearman rank correlation r = -0.59); additionally there was also significant correlation at soil temperatures above 10C (Fig. 7). Also, a weak but significant correlation with CH4 emission and soil temperature measured at different depths and with air temperature at weather stations was found (Spearman rank correlation r = 0.20; P \ 0.05). Analysing areas
Hydrobiologia (2012) 692:41–55
Fig. 3 Soil efflux of CO2-C from natural (N), drained (D), abandoned (A) and active (mined) peat extraction (M) sites. Median, average (in brackets) and interquartile range (mg C m-2 h-1) are given. For site type abbreviations see Table 1
Fig. 4 The relationship between CO2-C efflux and average (±standard deviation, SD) soil temperature (at 10, 20, 30 and 40 cm) at all study sites
with high water table (1N, 2N, 3N, 4N, 5D, 6D, 7D, 8D, 9D and 10D; see also Table 1), strong correlation was found (Spearman rank correlation r = 0.65; P \ 0.05) between CH4 emission and soil temperature (Fig. 8). Emissions increased for water table above 30 cm (Fig. 7). In this respect, correlation was not significant at other study sites with low water table on average, even though on many occasions it fluctuated up to the surface (e.g. Hiiesoo study sites). RDA analysis showed that CH4-C emissions were well explained by site type (48.9%) and to lesser extent by environmental variables (36.1%) (Fig. 5).
Median values of cumulative annual flux of CH4-C at natural, drained, abandoned and active extraction sites were 85.2, 23.7, 0.07 and 0.12 kg ha-1 year-1, respectively. Emissions of N2O Average emissions of N2O-N varied between -22.7 and 328.8 lg N2O-N m-2 h-1 (Fig. 9). There was a significant difference between natural and drained areas versus abandoned and active peat mining sites (Kruskal–Wallis ANOVA test; multiple
Hydrobiologia (2012) 692:41–55
Fig. 7 Relationship between CH4-C emission (lg C m-2 h-1) and water level depth (cm) for all study sites. Regression line was fitted only for CH4-C emission values recorded at average soil temperature values above 10C. Emissions increase for water table above 30 cm Fig. 5 Ordination diagram based on redundancy analysis of GHG flux data with respect to environmental variables. Dependent variables are indicated by solid arrows, and explanatory environmental variables are indicated by dashed arrows. Qualitative variables (land management types) are indicated by triangles: Nat area natural area; Drained drained area; Aband abandoned peat extraction site; Active active peat extraction site; Temp 10 soil temperature at 10 cm depth
comparison of mean ranks; P \ 0.05); the highest emissions were measured at active peat extraction sites (Fig. 9). Significantly higher values were
registered at three active peat extraction areas (16M, 17M and 18M) and from an area severely affected by peat mining (11A). The latter is located in an active mining area with an efficient drainage system (water level has been 55–78 cm below the surface and fluctuation is small compared with other drained sites). Therefore, the flora has no mire vegetation left and the area is used for forestry. Emissions correlated negatively with water level depth (Spearman rank correlation r = -0.36; P \ 0.05) but not with soil
Fig. 6 CH4-C emissions from natural (N), drained (D), abandoned (A) and active (mined) peat extraction (M) sites. Median, average (in brackets) and interquartile range (lg C m-2 h-1) are given. a and b, significantly differing values (Kruskal– Wallis ANOVA test and multiple comparison of mean ranks test). For site type abbreviations see Table 1
CH4-C mg C m-2 h-1
Fig. 8 Relationship between CH4-C emission (lg C m-2 h-1) and average (±SD) soil temperatures (at 10, 20, 30 and 40 cm) at all study sites with high water table (1N, 2N, 3N, 4N, 5D, 6D, 7D, 8D, 9D and 10D)
Hydrobiologia (2012) 692:41–55
y = 140.92e(0.20*x) R² = 0.45 p < 0.05
10000 8000 6000 4000 2000 0 0
Average temperature in soil (°C; 10-40 cm) Natural and drained areas
Fig. 9 N2O-N emissions at natural, drained, abandoned and active (mined) peat extraction sites. Median, average (in brackets) and interquartile range (lg N m-2 h-1) are given. a and b, significantly differing values (Kruskal– Wallis ANOVA test and multiple comparison of mean ranks test). For site type abbreviations see Table 1
temperature measured at different depths. RDA analysis showed that N2O emissions are explained by environmental variables (22.1%) and to lesser extent by site type (10.0%) (Fig. 5). Median values of cumulative annual flux of N2O-N from natural, drained, abandoned and active extraction sites were -0.05, -0.01, 0.18 and 0.19 kg N2ON ha-1 year-1, respectively.
Discussion Soil CO2 efflux Our results closely correspond with those of other studies in which CO2 fluxes were well explained by
soil temperature (e.g. Waddington et al., 2001; Koh et al., 2009; Ojanen et al., 2010). No significant correlation was found between CO2-C efflux and water level. Nevertheless, the RDA analysis indicated that water level and soil temperature at 10 cm from the ground surface explained 67.1% of CO2-C flux (Fig. 5). There was a significant difference in water table depth between mined and natural areas; the water regime was severely altered in mining areas, and it was considerably lower than in other drained areas and natural ones (Wilcoxon matched pair test, P \ 0.05). In natural and drained areas the water table was registered at no lower than 20 cm, whereas at abandoned and active extraction sites it was mostly below that level, exceeding 120 cm in depth. Mining areas were also characterized by a fluctuating water table;
Hydrobiologia (2012) 692:41–55
e.g., in the case of the autumn rainy season, water level exceeded 20 cm at some sites. It is estimated that peatland drainage and mining operations increase the depth of the aerobic zone, and therefore CO2 emissions could increase by 250–300% (Nyka¨nen et al., 1995). The results of our study showed a considerable increase in emissions in mining areas, although the increase was not as high at active extraction sites. Emissions of CO2-C from mining areas were higher than in other areas (Fig. 3). However, some in vitro experiments have shown higher CO2 production rates from natural sites compared with abandoned block-cut sites (Waddington et al., 2001). This was explained by the fact that the fibric surface layer contains a substantial supply of labile carbon. Importance was given to the top layer (*60 cm) of peat in CO2 production in natural peatlands, as this had higher substrate quality due to its proximity to organic matter inputs and better access to oxygen (Waddington et al., 2001). This may also explain why in certain drained areas (8D, 9D) CO2-C emissions were registered lower in comparison with natural areas—Sphagnum moss had disappeared from these sites and the area was covered with sparse vegetation, and also a lot of the area was open peat surface with high mineralization rates of the upper 0–5-cm layer of peat. At these study sites, substrate quality probably also fell over the years, and fresh organic matter input was low. At the same time, due to the high water level conditions, the circumstances were unfavourable for further peat oxidation. This was noticed especially at site 8D, the hollow without mosses and rare coverage of Rhynchospora alba, and water level not below 10 cm. Drained areas with tree coverage and dense vegetation (5D, 6D, 7D), and Sphagnum moss coverage (6D, 7D), had higher emissions. Due to water level drawdown, total litter input could have increased, especially due to higher tree litter input which could have caused also higher emissions. However, it is hard to predict the proportion of CO2-C originating from increased input of fresh litter and decomposition of peat layer, as shown by Strakova et al. (2010). The availability of fresh organic matter in active peat extraction areas was lower than in mined areas. Therefore, CO2-C emissions from active mining sites were lower than those from abandoned mining areas with considerable plant coverage. Nevertheless, there was one exception (16M) that had higher emission rates. Mining operations were recommenced in the
area in 2008, after it had been abandoned for 20 years. This was probably related to site preparation, which revealed a surface with less decomposed peat and higher substrate quality. In consideration of NEE and longer time periods, natural bogs and fens act as sinks for carbon and CO2 (Salm et al., 2009), but this situation may have interannual variations. Due to the drier climate and lower water table, aerobic conditions are developing, and these are favourable for decomposition of organic material (e.g. Moore & Dalva, 1993; Bubier et al., 2003). During some dry years, natural peatlands have been found to lose carbon (Alm et al., 1999; Waddington & Roulet, 2000), meaning that also net ecosystem emissions are negative. Therefore, weather conditions, especially summer temperatures, have a strong impact on soil CO2 efflux (Ma¨kiranta et al., 2007). Weather conditions in 2009 were wetter than in the previous decades, but temperatures were not considerably different (Table 2). It could be only assumed that relatively high water tables and reduced aerobic conditions determined emissions. Based on measurements in peat mining areas, Sundh et al. (2000) have determined CO2-C emissions from northern and southern Sweden to be from 628 to 1,967 and 1,092 to 2,787 kg CO2-C ha-1 year-1, respectively. In Finland, the study carried out by Alm et al. (2007) gives estimates of 2,672 kg CO2-C ha year-1. The results of these studies are comparable to the measurement results of this investigation. In addition, the net CO2 exchange approach does not incorporate peat loss due to wind and water erosion (estimated to take place at a rate of 6 mm year-1) (Waddington & McNeil, 2002) and ditches, which contribute an additional 1–3% (Sundh et al., 2000). Emissions of CH4 Our results show that hydrological regime of the study sites was severely affected in mining areas, which also explained low emissions at these sites. In the case of groundwater level deeper than 30 cm from the surface, no significant CH4-C emissions appear. A similar trend has been discovered by several investigations (e.g. Werner et al., 2003; Pelletier et al., 2007; Soosaar et al., 2010), and could be explained by the greater thickness of the potential CH4 production zone and the lesser thickness of the potential CH4 oxidation zone associated with the higher water table (Lai,
2009). Even in the case of water level increase, however, the weak correlation with CH4 net flux could be explained by enhanced CH4 oxidation in the oxygenated water column (Bubier et al., 1995). The potential for CH4 oxidation by methanotrophs also found to be of a magnitude greater than potential CH4 production by methanogens (Segers, 1998). Weak correlation was found with temperature at study sites with lowered water tables, which corresponds to the results of other studies (e.g. Sundh et al., 2000). On the other hand, sites with high water tables correlated strongly with temperatures, supporting the results of Dinsmore et al. (2009). Mining areas with deep water levels showed significantly lower emissions and methane consumption (e.g. area 11A). This could be attributed to the presence of highly recalcitrant organic matter in deeper peat layers that remained after peat harvesting, therefore providing less favourable conditions for CH4 production (Lai, 2009). Cleary et al. (2005) and Hyvo¨nen et al. (2009) reported low CH4-C fluxes at active peat extraction sites of 14.0 and 6.8 kg CH4C ha year-1, respectively. Our results confirm the significant reduction of emissions from peat excavation areas and average values of 5.0 kg CH4-C ha-1 year-1 from peat mining areas. In addition, although altered circumstances in peat mining areas do not necessarily result in complete cessation of CH4 emission, there are additional sources for emissions—fluxes from drainage ditches (Sundh et al., 2000) and stockpiles (Alm et al., 2007). Alm et al. (2007) reported 54.2 kg CH4-C ha-1 year-1, which includes emissions from ditches. According to Sundh et al. (2000), high emissions from ditches in mining areas may emit similar amounts of CH4 to those from pristine peatlands. In this respect, values reported by Alm et al. (2007) are close to emissions from the pristine areas in this study. However, this investigation does not include measurements performed at ditches and stockpiles. Data from the literature give an interquartile range for CH4-C emissions from natural and drained transitional fens and ombrotrophic bogs of 15–75 and 7.5–15 kg CH4-C ha-1 year-1, respectively (Salm et al., 2009). Although the emissions in this study fall in the range of emissions from natural areas, there is a difference in comparison with drained areas. The most likely reason for this was the unexpectedly high water table determined at the drained sites in this study.
Hydrobiologia (2012) 692:41–55
There were site-specific differences between natural sites (two different areas: 1N and 2N, of the wooded hummock bog subtype; 3N and 4N, of the wooded hollow–ridge bog subtype). High emissions from the drained site (10D) could be explained by the presence of E. vaginatum. Similar results are reported from an investigation by Minkkinen & Laine (2006). Plants can transport CH4-C from the anaerobic zone to the atmosphere via the roots and stem, thereby eliminating the possibility of aerobic oxidation (Sundh et al., 2000). Emissions of N2O The negative rates of N2O flux show that, in some soils of natural and drained study sites, denitrification would be complete and result in N2 emission. On the other hand, negative as well as small positive values measured in the study may be an error around zero. Therefore, small values may also show an absence of either N2O-N production or consumption. Similar conclusions have been reached by Hayden & Ross (2005) in estimating N2O-N emission from ombrotrophic peatland. In natural ombrotrophic bogs, N2O-N emissions have been estimated to be small or inconsequential (Martikainen et al., 1993; Regina et al., 1996; Alm et al., 2007), but these may be higher from areas drained for forestry purposes (Minkkinen et al., 2002). In addition, data from the literature give small flux values; the interquartile range for N2O-N emissions from natural and drained transitional fens and ombrotrophic bogs is 0.02–0.06 and 0.2–2.0 kg N2O-N ha-1 year-1, respectively (Salm et al., 2009). There is not much data available on N2O-N emissions from peat mining areas. Hyvo¨nen et al. (2009) reported N2O-N fluxes in the range 0.03–0.12 kg N2O-N ha-1 year-1 at an active peat extraction site, which is similar to the results presented in this study. In addition, it is estimated that the majority of N2O in peat harvesting areas is released from ditches and stockpiles (Alm et al., 2007). Alm et al. (2007) reported 2.0 kg N2O-N ha-1 year-1, including ditch emissions. One of the reasons for high emissions from active mining areas may be the lack of vegetation. Vegetation is dependent on N availability (Aerts et al., 1995) and is the main competitor for denitrifying microorganisms (Silvan et al., 2005). According to a study by Silvan et al. (2005), rapidly growing vegetation is a
Hydrobiologia (2012) 692:41–55
major control of N2O flux in sedge-dominated peatlands during the growing season, reducing the amount of available NO3 for denitrification and decreasing the N2O emissions from these ecosystems.
53 Production, Water Quality and Food Security (D1.20.10). We are grateful for the assistance provided by the Institute of Ecology and Earth Sciences of the University of Tartu, especially Anto Aasa and Marko Kohv, AS Tootsi Turvas, the Estonian Environmental Board, especially Madis Oras and Meelis Leivits, and the Estonian Fund for Nature.
Conclusions Our study showed that peat mining alters the fluxes of CO2, CH4 and N2O compared with those of natural peatlands. The median cumulative annual fluxes of CO2-C, CH4-C and N2O-N in active peat mining areas were 1741, 0.12 and 0.19 kg ha-1 year-1, respectively. Due to the lowered water table and the removal of peat from the upper layer, conditions remained similar to mining sites even after abandonment of these areas. Median values were similar to emissions from mining areas; cumulative annual fluxes of CO2C, CH4-C and N2O-N were 2,845, 0.07 and 0.18 kg ha-1 year-1, respectively. Therefore, abandoned as well as active mining sites are continuous net emitters of CO2 and N2O. Compared with natural areas, emissions from drained sites did not differ significantly, although the vegetation confirms the effects of drainage. This could be attributed to the similar water level, which was unexpectedly high in drained areas. Median values of N2O-N were significantly lower in natural and drained sites compared with mining areas, whereas CH4-C emissions were higher. CO2-C emissions did not differ significantly between the sites. Cumulative annual fluxes of CO2-C, CH4-C and N2O-N were 1,921, 23.7 and -0.01 kg ha-1 year-1 in drained, and 1,509, 85.2 and -0.05 kg ha-1 year-1 in natural areas. The results show that further studies should be undertaken to improve the estimation of C balance. These include C uptake by photosynthesis, emissions from ditches and stockpiles and C erosion. Further analysis of microbial community composition may also be a main factor in explaining the variation of GHG emission between microsites. Acknowledgments This study was supported by the Ministry of Education and Science of Estonia (grant SF0180127s08), the Estonian Science Foundation (grant 7527), a grant EE0012 from Iceland, Liechtenstein and Norway through the EEA Financial Mechanism and the Norwegian Financial Mechanism, and a grant through the IAEA Coordinated Research Project on Strategic Placement and Area-Wide Evaluation of Water Conservation Zones in Agricultural Catchments for Biomass
References Aerts, R., R. Logtestijn, M. Staalduinen & S. Toet, 1995. Nitrogen supply effects on productivity and potential leaf litter decay of Carex species from peatlands differing in nutrient limitation. Oecologia 104(4): 447–453. Alm, J., S. Saarnio, N. Nykanen, J. Silvola & P. J. Martikainen, 1999. Winter CO2, CH4 and N2O fluxes of some natural and drained peatlands. Biogeochemistry 44: 163–186. Alm, J., N. J. Shurpali, K. Minkkinen, L. Aro, J. Hyto¨nen, T. Laurila, A. Lohila, M. Maljanen, P. J. Martikainen, P. Ma¨kiranta, T. Penttila¨, S. Saarnio, N. Silvan, E.-S. Tuittila & J. Laine, 2007. Emission factors and their uncertainty for the exchange of CO2, CH4 and N2O in Finnish managed peatlands. Boreal Environment Research 12: 191–209. Basiliko, N., C. P. Blodau, P. C. Bengtson, C. Roehm & T. R. Moore, 2007. Regulation of decomposition and methane dynamics across natural, commercially mined, and restored northern peatlands. Ecosystems 10: 1148–1165. Bridgham, S. D., J. P. Megonigal, J. K. Keller, N. B. Bliss & C. Trettin, 2006. The carbon balance of North American wetlands. Wetlands 26: 889–916. Bubier, J. L., T. R. Moore & S. Juggins, 1995. Predicting methane emission from bryophyte distribution in northern Canadian peatlands. Ecology 76: 677–693. Bubier, J. L., G. Bhatia, T. R. Moore, N. T. Roulet & P. M. Lafleur, 2003. Spatial and temporal variability in growing-season net ecosystem carbon dioxide exchange at a large peatland in Ontario, Canada. Ecosystems 6: 353–367. Cleary, J., N. T. Roulet & T. R. Moore, 2005. Greenhouse gas emissions from Canadian peat extraction, 1990–2000: a life-cycle analysis. Ambio 34: 456–461. Croft, M., L. Rochefort & C. J. Beauchamp, 2001. Vacuumextraction of peatlands disturbs bacterial population and microbial biomass carbon. Applied Soil Ecology 18: 1–12. Dinsmore, K. J., U. M. Skiba, M. F. Billett & R. M. Rees, 2009. Effect of water table on greenhouse gas emissions from peatland mesocosms. Plant and Soil 318: 229–242. Dinsmore, K. J., M. F. Billett, U. M. Skiba, R. M. Rees, J. Drewer & C. Helfter, 2010. Role of the aquatic pathway in the carbon and greenhouse gas budgets of a peatland catchment. Global Change Biology 16: 2750–2762. Gilmer, A. J., S. M. Ward, N. Holden, A. Brebreton & E. P. Farrell, 1998. Restoration peatland carbon sequestration model: an aid in carbon cycling research. In Malterer, T., K. Johnson & J. Stewart (eds), Peatland Restoration and Reclamation: Proceedings of the 1998 International Peat Symposium, Duluth (14–18 July). International Peat Society, Duluth, MN: 197–201.
54 Gorham, E., 1991. Northern peatlands: role in the carbon cycle and probable responses to climatic warming. Ecological Applications 1: 182–195. Gorham, E., 1995. The biogeochemistry of northern peatlands and its possible response to global warming. In Woodwell, G. M. & F. T. Mackenzie (eds), Biotic Feedback in the Global 129 Climate System: Will the Warming Feed the Warming?. Oxford University Press, New York: 169–187. Hayden, M. J. & D. S. Ross, 2005. Denitrification as a nitrogen removal mechanism in a Vermont peatland. Journal of Environment Quality 34: 2052–2061. Hutchinson, G. L. & G. P. Livingston, 1993. Use of chamber systems to measure trace gas fluxes. In Harper, L. E., A. R. Mosier, J. M. Duxbury & D. E. Rolston (eds), Agricultural Ecosystems Effects on Trace Gases and Global Climate Change. ASA Special Publication No. 55. American Society of Agronomy, Madison, MI, USA: 1–55. Hyvo¨nen, N. P., J. T. Huttunen, N. J. Shurpali, N. M. Tavi, M. E. Repo & P. J. Martikainen, 2009. Fluxes of nitrous oxide and methane on an abandoned peat extraction site: effect of reed canary grass cultivation. Bioresource Technology 100: 4723–4730. IPCC, 2001. The carbon cycle and atmospheric carbon dioxide. In Houghton, J. T., et al. (eds), Climate Change 2001: The Scientific Basis. Cambridge University Press, New York: 183–237. Kirkinen, J., K. Minkkinen, T. Penttila¨, S. Kojola, R. Sieva¨nen, J. Alm, S. Saarnio, N. Silvan, J. Laine & I. Savolainen, 2007. Greenhouse impact due to different peat fuel utilization chains in Finland – a life-cycle approach. Boreal Environment Research 12: 211–223. Koh, H. S., C. A. Ochs & K. Yu, 2009. Hydrologic gradient and vegetation controls on CH4 and CO2 fluxes in a spring-fed forested wetland. Hydrobiologia 630: 271–286. Lai, D. Y. F., 2009. Methane dynamics in northern peatlands: a review. Pedosphere 19: 409–421. Laine, J., J. Silvola, K. Tolonen, J. Alm, H. Nyka¨nen, H. Vasander, T. Sallantaus, I. Savolainen, J. Sinisalo & P. J. Martikainen, 1996. Effect of water-level drawdown in global climatic warming: northern peatlands. Ambio 25: 179–184. Legendre, P. & L. Legendre, 1998. Numerical Ecology, 2nd English edn. Elsevier Science BV, Amsterdam, xv ? 853 pp. Loftfield, N., H. Flessa, J. Augustin & F. Beese, 1997. Automated gas chromatographic system for rapid analysis of the, atmospheric trace gases methane, carbon dioxide, and nitrous oxide. Journal of Environment Quality 26: 560–564. Ma¨kiranta, P., J. Hyto¨nen, L. Aro, M. Maljanen, M. Pihlatie, H. Potila, N. J. Shurpali, J. Laine, A. Lohila, P. J. Martikainen & K. Minkkinen, 2007. Soil greenhouse gas emissions from afforested organic soil croplands and cutaway peatlands. Boreal Environment Research 12: 159–175. Martikainen, P. J., H. Nyka¨nen, P. Crill & J. Silvola, 1993. Effect of a lowered water table on nitrous oxide fluxes from northern peatlands. Nature 366: 51–53. Martikainen, P. J., H. Nyka¨nen, J. Alm & J. Silvola, 1995. Change in fluxes of carbon dioxide, methane and nitrous oxide due to forest drainage of mire sites of different trophy. Plant and Soil 169(8): 571–577.
Hydrobiologia (2012) 692:41–55 Minkkinen, K. & J. Laine, 2006. Vegetation heterogeneity and ditches create spatial variability in methane fluxes from peatlands drained for forestry. Plant and Soil 285: 289–304. Minkkinen, K., R. Korhonen, I. Savolainen & J. Laine, 2002. Carbon balance and radiative forcing of Finnish peatlands 1900–2100 – the impact of forestry drainage. Global Change Biology 8: 785–799. Moore, T., 1994. Trace gas emissions from Canadian peatlands and the effect of climatic-change. Wetlands 14: 223–228. Moore, T. R. & M. Dalva, 1993. Influence of temperature and water table position on carbon dioxide and methane emissions from columns of peatland soils. Journal of Soil Science 44: 651–664. Nilsson, M., C. Mikkela¨, I. Sundh, G. Granberg, B. H. Svensson & B. Ranneby, 2001. Methane emission from Swedish mires: national and regional budgets and dependence on mire vegetation. Journal of Geophysical Research 106(D18): 20847–20860. Nyka¨nen, H., J. Silvola, J. Alm & P. J. Martikainen, 1995. Emissions of CH4, CO2 and N2O from a virgin fen and a fen drained for grassland in Finland. Journal of Biogeography 22: 351–357. Ojanen, P., K. Minkkinen & J. Alm, 2010. Soil–atmosphere CO2, CH4 and N2O fluxes in boreal forestry-drained peatlands. Forest Ecology and Management 260: 411–421. Orru, M., 1992. Estonian Peat Resources. Teaduste Akadeemia Kirjastus, Tallinn, Estonia (in Estonian). Paal, J., 1997. Eesti taimkatte kasvukohatu¨u¨pide klassifikatsi¨ RO keskkonnaprooon. Eesti Keskkonnaministeerium, U gramm, Tallinn (in Estonian). Pelletier, L., T. R. Moore, N. T. Roulet, M. Garneau & V. Beaulieu-Audy, 2007. Methane fluxes from three peatlands in the La Grande Rivie`re watershed, James Bay lowland, Canada. Journal of Geophysical Research 112: G01018. Ramst, R. & M. Orru, 2009. Eesti mahaja¨etud turbatootmisalade taastaimestumine. Eesti Po˜levloodusvarad ja -ja¨a¨tmed 1–2: 6–7 (In Estonian). Regina, K., H. Nykanen, J. Silvola & P. J. Martikainen, 1996. Fluxes of nitrous oxide from boreal peatlands as affected by peatland type, water table level and nitrification capacity. Biogeochemistry 35: 401–418. Saarnio, S., M. Morero, N. J. Shurpali, E.-S. Tuittila, M. Ma¨kila¨ & J. Alm, 2007. Annual CO2 and CH4 fluxes of pristine boreal mires as a background for the lifecycle analyses of peat energy. Boreal Environment Research 12: 101–113. ¨ . Mander, 2009. Global Salm, J.-O., K. Kimmel, V. Uri & U warming potential of drained and undrained peatlands in Estonia: a synthesis. Wetlands 29: 1081–1092. Segers, R., 1998. Methane production and methane consumption: a review of processes underlying wetland methane fluxes. Biogeochemistry 41(1): 23–51. Silvan, N., E. S. Tuittila, V. Kitunen, H. Vasander & J. Laine, 2005. Nitrate uptake by Eriophorum vaginatum controls N2O production in a restored peatland. Soil Biology and Biochemistry 37: 1519–1526. ¨ . Mander, M. Maddison, A. Kanal, A. Kull, K. Soosaar, K., U Lo˜hmus, J. Truu & A. Augustin, 2011. Dynamics of gaseous nitrogen and carbon fluxes in riparian alder forests. Ecological Engineering 37: 40–53. Stanek, W. & T. Silc, 1977. Comparisons of four methods for determination of degree of peat humification
Hydrobiologia (2012) 692:41–55 (decomposition) with emphasis on the von Post method. Canadian Journal of Soil Science 57: 109–117. Strakova, P., J. Anttila, P. Spetz, V. Kitunen, T. Tapanila & R. Laiho, 2010. Litter quality and its response to water level drawdown in boreal peatlands at plant species and community level. Plant and Soil 335: 501–520. Sundh, I., M. Nilsson, C. Mikkela¨ & G. Granberg, 2000. Fluxes of methane and carbon dioxide on peat-mining areas in Sweden. Ambio 29: 499–503. Tuittila, E. S., V. M. Komulainen, H. Vasander, H. Nyka¨nen, P. Martikainen & J. Laine, 2000. Methane dynamics of a restored cut-away peatland. Global Change Biology 6: 569–581. Turunen, J., E. Tomppo, K. Tolonen & A. Reinikainen, 2002. Estimating carbon accumulation rates of undrained mires in Finland – application to boreal and subarctic regions. Holocene 12: 69–80. Vitt, D. H., L. A. Halsey, I. E. Bauer & C. Campbell, 2000. Spatial and temporal trends in carbon storage of peatlands of continental western Canada through the Holocene. Canadian Journal of Earth Sciences 37: 683–693. Von Arnold, K., B. Ha˚nell, J. Stendahl & L. Klemedtsson, 2005a. Greenhouse gas fluxes from drained organic forests
55 land in Sweden. Scandinavian Journal of Forest Research 20: 400–411. Von Arnold, K., M. Nilsson, M. Ha˚nell, P. Weslien & L. Klemedtsson, 2005b. Fluxes of CO2, CH4 and N2O from drained organic soils in deciduous forests. Soil Biology and Biochemistry 37: 1059–1071. Waddington, J. M. & P. McNeil, 2002. Peat oxidation in an abandoned cutover peatland. Canadian Journal of Soil Science 82: 279–286. Waddington, J. M., K. D. Warner & G. F. Kennedy, 2002. Cutover peatlands: a persistent source of atmospheric CO2. Global Biogeochemical Cycles 16: 1002. doi:10.10292001 GB001398. Waddington, J. M. & N. T. Roulet, 2000. Carbon balance of boreal patterned peatland. Global Change Biology 6: 87–97. Waddington, J. M., P. A. Rotenberg & F. J. Warren, 2001. Peat CO2 production in a natural and cutover peatland: implications for restoration. Biogeochemistry 54: 115–130. Werner, C., D. Kenneth, P. Bakwin, C. Yi, D. Hurts & L. Lock, 2003. Regional-scale measurements of CH4 exchange from a tall tower over a mixed temperate/boreal lowland and wetland forest. Global Change Biology 9: 1251–1261.