Methane emissions in two drained peat agro-ecosystems with high and low agricultural intensity
Methane (CH4) emissions were compared for an intensively and extensively managed agricultural area on peat soils in the Netherlands to evaluate the effect of reduced management on the CH4 balance. Chamber measurements (photoacoustic methods) for CH4 were performed for a period of three years in the contributing landscape elements in the research sites. Various factors influencing CH4 emissions were evaluated and temperature of water and soil was found to be the main driver in both sites. For upscaling of CH4 fluxes to landscape scale, regression models were used which were specific for each of the contributing landforms. Ditches and bordering edges were emission hotspots and emitted together between 60% and 70% of the total terrestrial CH4 emissions. Annual terrestrial CH4 fluxes were estimated to be 203 (±48%), 162 (±60%) and 146 (±60%) kg CH4 ha−1 and 157 (±63%), 180 (±54%) and 163 (±59%) kg CH4 ha−1 in the intensively managed site and extensively managed site, for 2006, 2007 and 2008 respectively. About 70% of the CH4 was emitted in the summer period. Farm based emissions caused per year an additional 257 kg CH4 ha−1 and 172 kg CH4 ha−1 for the intensively managed site and extensively managed site, respectively. To further evaluate the effect of agricultural activity on the CH4 balance, the annual CH4 fluxes of the two managed sites were also compared to the emissions of a natural peat site with no management and high ground water levels. By comparing the terrestrial and additional farm based emissions of the three sites, we finally concluded that transformation of intensively managed agricultural land to nature development will lead to an increase in terrestrial CH4 emission, but will not by definition lead to a significant increase in CH4 emission when farm based emissions are included.
KeywordsPeat soils Ditches Temperature Upscaling
The assessment of the contribution of the trace gas methane (CH4) to the greenhouse gas effect and the related global warming is of great importance. Northern peatlands are significant sources of CH4, and are estimated to emit between 20 and 50 Tg yr−1 (Mikaloff Fletcher et al. 2004a, b). In northern oligotrophic and eutrophic managed peatland systems, net uptake and emission rates of CH4 have been found to depend on groundwater level, soil moisture content, temperature, and grassland management (Van den Pol-van Dasselaar et al. 1998a; Hargreaves and Fowler 1998; Christensen et al. 2003; Blodau and Moore 2003; Hendriks et al. 2007; Pelletier et al. 2007). CH4 emissions are difficult to estimate because of their large spatial and temporal variability (e.g. Smemo and Yavitt 2006). Therefore there are two major challenges when upscaling of CH4 fluxes from small scale to landscape scale: 1) selecting the correct ecosystem variables for stratification and 2) developing robust predictive relationships (Groffman et al. 2000).
The majority of fen meadow areas in Atlantic Europe are intensively managed. In the Netherlands, eutrophic peatlands have been drained for centuries and in the last 50 years peatlands have been drained even more deeply to make modern agriculture possible. These peatlands are therefore major carbon sources of CO2 as a result of peat oxidation (Langeveld et al. 1997; Veenendaal et al. 2007). Burgerhart (2001) has suggested that peat oxidation can be reduced if agricultural peatlands are transformed into wetland nature by raising the water table and by reducing agricultural intensity. These measures alter the carbon cycle and probably turn carbon sources into carbon sinks. Large uncertainties exist, however, of such measures on the CH4 balance. Hendriks et al. 2007 studied a nature restoration site with no agricultural practices since 1997. They found that a sink of 71 g CO2-equiv m−2 yr−1 had developed and they attributed this to a decrease in CO2 emissions from fields and an increase in CH4 emissions from ditches and waterlogged soil after stopping agricultural activity. However, in this case the previous situation for CH4 emission in the time that this area was used for intensive farming is unknown.
The objective of this study is 1) to find empirical relationships that describe the CH4 emissions from the different landscape elements and to provide a spatially integrated flux for the intensively managed and extensively managed area and 2) to investigate the influence of agricultural activity on the CH4 balance. Firstly, terrestrial methane fluxes were studied in an intensively managed area, where manure and fertiliser are applied and frequent mowing is practiced. These fluxes are compared to terrestrial CH4 fluxes measured in an extensively managed area with reduced agricultural activity. Secondly, farm based methane fluxes were estimated using an emissions factor approach. Methane fluxes of the two managed sites were compared to methane fluxes in a former agricultural site which was converted to a nature reserve 11 years ago by stopping intensive dairy farming, raising the water table (>20 cm below field level for most of the year) (Hendriks et al. 2007).
Experimental sites and methodology
Main characteristics of the Oukoop and Stein site in the Netherlands
Soil class, topography and landscape element
Fibric rheic eutric Histosol
0–23 cm: anthropogenic topsoil
0–23 cm: anthropogenic
Flat, (alluvial) plain
Peat from ditches on edges
23–50 cm: clayey peat
Saturated for long periods during winter as a result of compaction
Application of fertiliser: 370, 350 and 260 kg ha−1in 2006, 2007 and 2008, resp.
>50 cm: peat, 70% discernible remnants of wood and reed
Mean highest WT: ca. 35 cm Mean lowest WT: ca. 60 cm
Application of cow manure: 55, 57 and 90 m3 ha−1in 2006, 2007 and 2008, resp
Fibric rheic eutric Histosol
0–23 cm: anthropogenic topsoil
0–29 cm: anthropogenic
Poorly drained since 2005
Flat, (alluvial) plain
Peat from ditches on edges
29–50 cm: clayey peat
Saturated for long periods every year
No application of fertiliser or cow manure
>50 cm: peat, few recognizable plant remnants
Mean highest WT: ca. 25 cm Mean lowest WT: ca. 50 cm
The management in Stein has been changed from intensive to extensive some 20 years ago and the area has been turned into a bird reserve. The research area was used as hayfield without application of manure and fertiliser during the measurement period and was mown three times each year after June 15 (2006, 2007 and 2008). The water table was adjusted since 2006, with a high water table (15–20 cm below the surface) in winter and a low water table in summer (50 cm below the surface). In Stein, Lolium perenne is also dominant, often with Poa trivialis co-dominant and no plants are present in the drainage ditches. Over time, Holcus lanatus, Anthoxanthum odoratum and Rumex acetosa have become more abundant. Characteristics of the Stein site are given in Table 1.
Terrestrial emissions of CH4 were determined using a closed chamber method (Hutchinson and Mosier 1981). Methane concentrations were measured using a Photo Acoustic Field Gas Monitor (INNOVA 1412 sn, 710-113, ENMO services, Belgium) connected by Teflon tubes to a PVC chamber (e.g. van Huissteden et al. 2005). Samples were taken from the headspace of a closed cylindrical, dark chamber (30 cm diameter, 25 cm height) that was placed on a collar. A small fan was installed in the chamber to homogenize the inside air and a water lock was placed to control inside pressure. On land we used water between the chamber and the collar to seal the chamber from the ambient air during the measurement. For the ditches we used floaters and a lever system to gently lower the chamber onto the water surface, carefully avoiding the effect of pressure differences and the disturbance of the water surface. We used external silica gel and soda lime filters to minimize cross-interference of CO2 and water vapour when CH4 was measured at high CO2 concentrations. The gas analyser was annually calibrated and tested for drift at the NMI (The Netherlands Institution of Standards, Delft, the Netherlands). In addition, the equipment was occasionally cross-checked for drift with a standard calibration gas. During the measurement period however, the equipment was not found to drift. All measurements were taken during the day, between 9am and 4pm. Each flux measurement consisted of five point-measurements taken at one-minute intervals from which dC/dt was calculated.
At the beginning of January 2005, 6 PVC collars (diameter 30 cm) were installed randomly in the footprint area of the eddy covariance systems at both experimental sites, so that gas emissions could be sampled. The measurements were carried out from January 2005 to November 2008 once a month up to twice a month or more during intensive field campaigns.
From February 2006 until November 2008, 19 additional sampling points, distributed over the experimental sites, were sampled every month at both research locations to study spatial variability. At both locations three landscape elements were distinguished with different soil/water temperatures and soil moisture conditions. They were: permanently water-filled ditches, ditch edges, and the field area with a fluctuating water table. Acreage of the landforms was determined by measurements in the field (Nol et al. 2008), the use of GIS maps and use of aerial photographs. In each of the two fields, four sample points were located on the water surface of the ditches, four points on the ditch edges and 11–14 sample points in the fields. In total over 1200 measurements were taken.
In addition to each flux measurement, soil or water temperature was measured at 5 cm depth and soil moisture content was determined in the top 5 cm of soil at the sample points, using a HH2 Delta-T device (Delta T Devices, Llandindrod Wells, USA) calibrated for the soil type. Soil and water temperatures were measured every half hour by sensors installed at a depth of 4 cm (Campbell scientific, USA and e+ sensor L-50, Eijkelkamp, the Netherlands, respectively). Water table depth was recorded hourly with pressure sensors installed in a steel frame to a depth of 70 cm into the soil at one or two places in the field (e+ sensor L-50, Eijkelkamp Agrisearch Equipment BV, Giesbeek, the Netherlands). In the ditch water of both experimental sites, electrical conductivity (EC) (Multi 350i, Fisher Scientific, the Netherlands) and pH (Cond340i-SET, Fisher Scientific, the Netherlands) was measured in 2008. In Stein, water tables were determined during each field CH4 measurement in 2008 at the location of the measurement.
Wind speed, air temperature and water vapour pressure were measured with an eddy covariance system consisting of a Campbell Csat C3 Sonic anemometer (Campbell Scientific, Logan, Utah, USA) directed into the main wind direction and a Licor 7500 open path Infrared gas analyser (LI-COR Lincoln, NE, USA). The height of the mast was 305 cm and was located in the middle of the field sample points in both areas (Veenendaal et al. 2007; Schrier-Uijl et al., accepted AFM). For continuous measurements of CH4 in the intensively managed site, we used the eddy covariance dataset of Kroon et al. (2007) where CH4 concentrations were measured with a system consisting of a three-dimensional sonic anemometer and a QCL spectrometer (model QCL-TILDAS-76, Aerodyne Research Inc., Billerica MA, USA).
Calculations and statistical analyses
Each flux measurement consisted of five points taken at one-minute intervals. The slope dC/dt of the gas concentration curve at time t = 0 s was estimated using linear regression and the slope-intercept method as described by Kroon et al. (2008). The average flux values for CH4 estimated by the slope intercept method were not significantly different from those estimated by the linear method and therefore, linear regression was used to calculate CH4 fluxes. The short measurement period of 240 seconds, the rejection of the last point in the case of levelling, and additional measures taken to prevent leakage, mixing and temperature artefacts made it possible to use linear regression (Schrier-Uijl et al., accepted AFM). First, the data quality was assessed: outliers resulting from disturbances, chamber leakage or instrument failures were removed from the data set. Annual mean net CH4 emissions were estimated by linear regression models of natural-logarithm transformed CH4 data using Tsoil/Twater as explanatory variables (model-based approach, e.g. Hargreaves and Fowler 1998). The reliability of this model based approach has been tested previously for the Oukoop experimental site by performing a comparison of the chamber method measurements with eddy covariance measurements of CH4. The methods agreed well and the difference in cumulative emissions was 17% over a three months period (Schrier-Uijl et al. accepted AFM).
The statistical significance of differences between landscape elements within sites was calculated with one-way ANOVA. Analysis of covariance, with temperature as covariate, was used to ascertain the statistical significance of differences in the emissions from the landscape elements of the two sites. Correlations between natural-logarithm transformed CH4 emissions and independent variables were calculated using step-wise multiple linear regression analysis (case-wise elimination of variables). Paired T-tests were used to calculate the statistical difference in emission between sites at the same day. Uncertainties per landscape element were estimated with a temperature dependent approach and were weighted for the proportion of each landscape element in the total landscape. Statistical analyses were carried out with SPSS. For the calculations of the contribution of animals and manure in the total CH4 balance, we used the method as described by Hensen et al. (2006) which uses simple emission factors for dairy cows, heifers, calves, manure and farmyard manure.
Matrix of Pearson correlations (r) of ln transformed methane emissions and related parameters for the different landscape elements in Oukoop (intensively managed) and Stein (extensively managed) measured by chambers (left) and eddy covariance (right)
Oukoop chamber measurements
Oukoop EC measurements
0.265** (n = 195)
0.545** (n = 134)
−0.235* (n = 119)
0.2665(n = 40)
0.481** (n = 2540)
−0.0815(n = 77)
0.063**5(n = 2545)
0.397** (n = 80)
0.350** (n = 74)
−0.237* (n = 73)
0.422** (n = 90)
0.478** (n = 97)
0.1731(n = 30)
0.503**3(n = 26)
Stein chamber measurements
Stein EC measurements
0.474** (n = 186)
0.279* (n = 190)
−0.166* (n = 146)
0.1775(n = 36)
0.500** (n = 64)
0.473** (n = 60)
0.05 (n = 59)
0.424** (n = 59)
0.496** (n = 76)
0.0352(n = 30)
−0.0724(n = 30)
Correlation of field CH4 emissions with soil moisture were significant (P < 0.05) in both the intensively and the extensively managed sites (Table 2), but when using non-linear multiple regression with temperature as explanatory variable, adding moisture did not significantly improve the predictive power of the regression equation. However, comparing monthly datasets revealed an exception for April in 2006 and 2007, when soil moisture was a stronger predictor than soil temperature for ln(CH4) at both sites. The water table did not show a correlation with ln(CH4) neither measured by eddy covariance in Oukoop nor measured by chambers in Stein (Table 2). Eddy covariance measurements show that after temperature correction variability in fluxes over the total area seem to be significantly correlated to wind velocity (Table 2). However, because fluxes measured by eddy covariance are integrated over a large area it is not clear if this dependence on turbulence appears in fields, edges, ditches or in a combination of these landscape elements. Ditch water temperature and electrical conductivity (EC) in Oukoop were positively correlated with ln(CH4) fluxes and pH was not (Table 2). Correlation statistics of chamber measurements (over three years) and eddy covariance measurements (over three months) are given in Table 2.
Water temperature was the main driving variable for CH4 emission in ditches and soil temperature in soils and therefore, water and soil temperature based regressions were used to estimate annual CH4 balances.
Annual terrestrial methane balances
Mean soil temperatures, water temperatures (°C) and temperature based estimates of annual terrestrial methane emissions (kg CH4 ha−1) for the intensively (Oukoop) and extensively (Stein) managed areas
Terrestrial CH4emission Oukoop
Terrestrial CH4emission Stein
Discussion and conclusions
The objective of this research was to investigate the influence of reduced agricultural activity on the CH4 balance in two distinct areas. Therefore, the CH4 emission from a drained, intensively managed peat area was compared to the CH4 emission from a drained, extensively managed peat area.
First, the spatial and temporal variation of fluxes was investigated and a regression based approach was used to get spatially integrated annual CH4 emissions. Both areas were stratified into fields, ditches and ditch edges and in both sites the permanently water-filled, 30–60 cm deep ditches and ditch edges turned out to be CH4 hotspots that together emit 60% and 68% of the total terrestrial emission in Stein and Oukoop, respectively. Apparently, in the procedure of upscaling of the chamber based fluxes to landscape scale it is of great importance to determine the flux associated with each landscape element because they contribute significantly different to the CH4 balance.
The edges of the ditches were saturated for most of the year, with soil moisture contents >60%. These conditions can give rise to a differently vegetation than in the drier fields and in some places Iris pseudacorus and Typha angustifolia are present. These aerenchymatic plants may cause additional emissions in the edges because CH4 diffuses rapidly through their stems (e.g. Hendriks et al. 2007).
The CH4 emission rates from all three landscape elements vary seasonally. In the extensively managed site, high CH4 emissions were concentrated in the summer period, while in the intensively managed site they were concentrated in early spring and summer, partly associated with field applications of slurry. Van den Pol-van Dasselaar et al. (1999) also reported higher CH4 emissions after manure application suggesting the combination of higher temperatures in summer, wet soil, the application of easily decomposable organic material and the anaerobic conditions in the slurry as the reason for enhanced CH4 production. In both sites, over 70% of the total terrestrial CH4 flux is emitted in summer. The exceptionally high emission peaks from ditches (up to 800 mg m−2 hr−1 in the summer) may have resulted from ebullition events, during which CH4 quickly passes through the top layer in the water column. These events mainly occurred when both the water temperature and the wind velocity were high. Ebullition can play an important role during turbulent conditions when bubbles are triggered to escape from the ditch bottom and also low air pressure can cause an increase in ebullition (Hendriks et al. 2008). Besides, during turbulent conditions the aquatic boundary layer thickness is decreased and the potential diffusion rate increases (Kremer et al. 2003; Minkkinen and Laine 2006).
Because an integration-based approach may lead to overestimation due to exceptionally high emissions eg. in the summer months, a regression based approach was used to calculate annual CH4 fluxes (eg. Mikkelä et al. 1995; Chanton et al. 1995). Therefore, we aimed to find empirical relationships that describe the CH4 emissions from the landscape elements of the two drained peat areas and we found temperature to be the main driver. Except for temperature, no other measured variables did contribute significantly to the predictive power of the chamber based regression. Electrical conductivity was significantly correlated to CH4 emission from ditches, but to less data was available to include this in the regression. The only month for which we found a strong significant correlation between volume fraction of water in the soil and CH4 emission rates was April—the period when the field begins to dry out after being waterlogged in winter and when air temperature may rise rapidly from 10 to 25°C. Water table did not influence CH4 emissions significantly, but the highest emissions occurred at intermediate and sometimes high water tables. In the study areas the water table fluctuated seasonally and within seasons also due to rainfall and polder level regulation by the Dutch water board with high (−15 cm) average water table depths in winter and low (−50 cm) in summer in Stein. In Oukoop water tables depths were low (−65 cm) in summer and high (−20 cm) in winter, and in the winter period perched water tables occurred after heavy rains. Most strong correlations between water table and CH4 emission in other studies have been found at water tables between 0 and 20 cm below field level (eg. Furukawa et al. 2005; Hargreaves and Fowler 1998; Strack et al. 2004). The mostly lower water table (>20 cm below field level) in our site could be the cause for the lack of correlation between water table and CH4 emission.
Eddy covariance measurements showed that no significant differences in CH4 flux occurred between day and night after correction for temperature. Thus, empirical relations based on temperature can be used to predict 24-hour fluxes of CH4 per landscape element. The predictions of monthly to annual CH4 emission have found to be in good agreement with values as measured by eddy covariance (Schrier et al, accepted AFM). However, considering the daily variability in CH4 fluxes a better understanding of the influence of variables such as wind, precipitation, air pressure and application of manure and fertiliser on CH4 emission is needed.
Comparison between the CH4 emission rates in this study and the CH4 emission rates reported in other studies on peatland ecosystems. Mean CH4 emission rates are in mg CH4 m−2 hr−1 and the three last columns represent the landscape elements
Edge/ saturated land
Minkkinen and Laine (2006)
15.72 up to 25 in summer
Hendriks et al. (2007)
Eutrophic fen abandoned agriculture
Bubier et al. (1993)
5.8 up to 38.2 in summer
Bellisario et al. (1999)
Less eutrophic fen
Pelletier et al. (2007)
Liblik et al. (1997)
Van den Pol-van Dasselaar et al. (1998b)
Less eutrophic fen
Waddington and Day (2007)
Less eutrophic fen
Adrian et al. (1994)
Huttunen et al. (2003)
up to 8.0
Hamilton et al. (1994)
Less eutrophic fen
Chanton et al. (1993)
Less eutrophic fen
5.3–12.4 (aerenchym plants)
Schrier et al. (this study)
Eutrophic fen (intensively managed)
Schrier et al. (this study)
Eutrophic fen (extensively managed)
The field fluxes found in this study (water tables of 20–60 cm below field level) are lower compared to fluxes found in some other eutrophic systems with water tables between 0 and 20 cm below field level for most of the year (eg. Bellisario et al. 1999; Hendriks et al. 2007; Furukawa et al. 2005; Hargreaves and Fowler 1998; Strack et al. 2004). The extremely high emission rates from ditches found by Minkkinen and Laine (2006) and by Bubier et al. (1993) were similar to the extreme values we found at turbulent water conditions in summer, but average summer fluxes of 18.72 mg m2 hr−1 were higher.
Estimates of annual terrestrial and annual total (terrestrial + farm based) methane emissions (kg ha−1) for the intensively (Oukoop) and extensively (Stein) managed areas
Terrestrial CH4emission Oukoop
Total CH4emission Oukoop
Terrestrial CH4emission Stein
Total CH4emission Stein
To investigate the influence of agricultural activity on the CH4 balance, CH4 fluxes of our two managed sites were also compared to a former agricultural peat site which was converted to a nature reserve 11 years ago by raising the water table (>20 cm below field level for most of the year) and cessation of any agricultural management (Hendriks et al. 2007). Without farm based emissions, the intensively managed area and extensively managed area are a net source of CH4 ranging from 146 to 203 kg ha−1 and from 156 to 180 kg ha−1, respectively and the natural site is a larger source of 417 kg ha−1. When we took account of the farm-based emissions, the estimated total annual CH4 flux was 427 kg ha−1 in the intensively managed site and 339 kg ha−1 in the extensively managed site averaged over the three years compared to 417 kg ha−1 in the natural site of Hendriks et al. (2007). This suggests that transformation of intensively managed agricultural land to nature development will lead to an increase in terrestrial CH4 emission, but not by definition lead to a significant increase in CH4 emission when farm based emissions are included.
This research project was funded by Wageningen University, The Province of North Holland, CarboEurope IP (contract number: GOCE-CT-2003-505572) and the Dutch National Research Programme Climate Changes Spatial Planning (www.klimaatvoorruimte.nl). Many thanks are due to Jan van Walsem and Frans Möller for their help in the field and in the lab and to the van Eijk and van Eeuwijk families for their logistical support. We are grateful to the State Forestry Service for access to the site and to Nico de Bruin for his collaboration. J. Burrough advised on the English and E.J. Bakker advised on the statistical analyses.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
- Bubier J, Costello A, Moore TR, Roulet NT, Savage K (1993) Microtopography and methane flux in boreal peatlands, northern Ontario, Canada. Can J Bot-Revue Canadienne De Botanique 71:1056–1063Google Scholar
- Burgerhart N (2001) Mogelijkheden voor koolstofopslag in Nederlandse ecosystemen, Dutch National Inventory Report for LULUCF sector, category 5D1Google Scholar
- Hendriks DMD, van Huissteden J, Dolman AJ, van der Molen MK (2007) The full greenhouse gas balance of an abandoned peat meadow. Biogeosciences 4:411–424Google Scholar
- Hendriks DMD, Dolman AJ, van der Molen MK, van Huissteden J (2008) A compact and stable eddy covariance set-up for methane measurements using off-axis integrated cavity output spectroscopy. Atmospheric Chemistry and Physics 8(2):431–443Google Scholar
- Kroon PS, Hensen A, Jonker HJJ, Zahniser MS, van ’t Veen WH, Vermeulen AT (2007) Suitability of quantum cascade laser spectroscopy for CH4 and N2O eddy covariance flux measurements. BG 4:715–728Google Scholar
- Mikaloff Fletcher SE, Tans PP, Bruhwiler LM, Miller JB, Heimann M (2004a) CH4 sources estimated from atmospheric observations of CH4 and 13C/12C isotopic ratios: 1. Inverse modeling of source processes. Glob Biogeochem Cycles 18(4), doi:10.1029/2004GB002223
- Mikaloff Fletcher SE, Tans PP, Bruhwiler LM, Miller JB, Heimann M (2004b) CH4 sources estimated from atmospheric observations of ch4 and 13C/ 12C isotopic ratios: 2. Inverse modeling of CH4 emissions from geographical regions. Glob Biogeochem Cycles, 18Google Scholar
- Pelletier L, Moore TR, Roulet NT, Garneau M, Beaulieu-Audy V (2007) Methane emissions from three peatlands in the La Grande Riviera watershed, James Bay lowland, Canada. J Geophys Res-Biogeosci, 112, noG1, Note(s): G01018.1-G01018.12] (3/4 p)Google Scholar
- Strack M, Waddington JM, Tuittila ES (2004) Effect of water table drawdown on northern peatland methane dynamics: Implications for climate change. Global Biochemical Cycles, doi:10.1029/2003GB002209
- Van Amstel A, Kroeze C, van Eerdt M, Dumont M, Both D (2003) National emissions of methane and nitrous oxide from agriculture in The Netherlands: current estimates and future trends, Workshop on Inventories and projections of greenhouse gas and ammonia emissions from agriculture, 27–28 February 2003, European Environmental Agency, CopenhagenGoogle Scholar
- Waddington JM, Day SM (2007) Methane emissions from a peatland following restoration, J Geophys Res-Biogeosci, doi:10.1029/2007JG000400