Advertisement

Biogeochemistry

, Volume 102, Issue 1–3, pp 265–279 | Cite as

Release of CO2 and CH4 from lakes and drainage ditches in temperate wetlands

  • A. P. Schrier-Uijl
  • A. J. Veraart
  • P. A. Leffelaar
  • F. Berendse
  • E. M. Veenendaal
Open Access
Article

Abstract

Shallow fresh water bodies in peat areas are important contributors to greenhouse gas fluxes to the atmosphere. In this study we determined the magnitude of CH4 and CO2 fluxes from 12 water bodies in Dutch wetlands during the summer season and studied the factors that might regulate emissions of CH4 and CO2 from these lakes and ditches. The lakes and ditches acted as CO2 and CH4 sources of emissions to the atmosphere; the fluxes from the ditches were significantly larger than the fluxes from the lakes. The mean greenhouse gas flux from ditches and lakes amounted to 129.1 ± 8.2 (mean ± SE) and 61.5 ± 7.1 mg m−2 h−1 for CO2 and 33.7 ± 9.3 and 3.9 ± 1.6 mg m−2 h−1 for CH4, respectively. In most water bodies CH4 was the dominant greenhouse gas in terms of warming potential. Trophic status of the water and the sediment was an important factor regulating emissions. By using multiple linear regression 87% of the variation in CH4 could be explained by PO4 3− concentration in the sediment and Fe2+ concentration in the water, and 89% of the CO2 flux could be explained by depth, EC and pH of the water. Decreasing the nutrient loads and input of organic substrates to ditches and lakes by for example reducing application of fertilizers and manure within the catchments and decreasing upward seepage of nutrient rich water from the surrounding area will likely reduce summer emissions of CO2 and CH4 from these water bodies.

Keywords

Biogeochemistry Drainage ditches Greenhouse gas emission Nutrients Peatlands Temperate lakes 

Introduction

Freshwater bodies such as ditches, streams, wetlands and lakes contribute appreciably to the processing of carbon and its transport to the atmosphere (e.g. Bastviken et al. 2004; Walter et al. 2006; Wang et al. 2006). It has been estimated that lakes annually emit 8–48 Tg methane (CH4), which is 6–16% of the global natural CH4 emissions (Bastviken et al. 2004; St. Louis et al. 2000), and 513 Tg C carbon dioxide (CO2) (Cole et al. 1994). Saarnio et al. (2009) have estimated that large lakes alone account for 24% of all wetland CH4 emissions in Europe. It has been shown that small water bodies also significantly contribute to the landscape-scale CH4 budgets in wetland regions (e.g. Schrier-Uijl et al. 2009a, b; Juutinen et al. 2009; Repo et al. 2007; Walter et al. 2007; Roulet and Moore 1995). Yet though it is likely that both lakes with organic-rich sediment and also eutrophic ditches contribute especially significantly to regional greenhouse gas balances, they are poorly studied and very little is known about their underlying biogeochemical processes (Saarnio et al. 2009).

CO2 is produced by respiration in sediments and throughout the water column and can also be a product of biological processes in the sediment (Fig. 1). As CO2 is highly soluble, high concentrations can accumulate near the sediment/water interface, which results in oversaturation and release to the atmosphere. It has been suggested that the transport of dissolved organic carbon (DOC) from terrestrial environments is an important source of carbon in aquatic environments. If this is the case, lakes in organic-rich peatlands have larger CO2 fluxes than lakes in mineral catchments (Rantakari and Kortelainen 2005; Huttunen et al. 2002).
Fig. 1

Simplified illustration of CO2 and CH4 dynamics in water bodies, with OM organic matter

CH4 emission is the balance of two counteracting processes: methanogenesis in anoxic conditions and the oxidation of the generated CH4 (Minkkinen and Laine 2006; Bastviken et al. 2002), (Fig. 1). CH4 is a major product of carbon metabolism in lakes; its production depends on the availability of alternative electron acceptors such as O2, NO3 , Fe3+ and SO42− (van Bodegom and Scholten 2001). After these electron acceptors have been used up, CH4 production becomes the dominating degradation process of organic matter and is the terminal microbial process in the anaerobic degradation of organic matter. The CH4 travels from the sediment through the water column to the atmosphere and on the way it can be oxidised into CO2 (Whiting and Chanton 2001). Most of the CH4 that remains unoxidised will be emitted by diffusive flux to the atmosphere.

The underlying microbial processes affecting CO2 and CH4 production and emission are regulated by variables such as sediment and water temperature, oxygen availability, organic matter availability and composition, sediment and water chemistry, the presence of electron acceptors (redox conditions), pH, electrical conductivity (EC) and factors such as water depth and lake size (e.g. Stadmark and Leonardson 2005; Juutinen et al. 2009; Repo et al. 2007; Frei et al. 2006; Loeb et al. 2007; Casper et al. 2003).

Most of the freshwater lakes in the Netherlands are in peat areas, are very shallow (<2 m), and were created by large-scale dredging and removal of peat during the early seventeenth century (Gulati and van Donk 2002). They vary in area, depth, hydrology and physico-chemical characteristics, but most of them are eutrophic, due to the application of fertilisers and manure within their catchments, the oxidation of peat and the upward seepage of nutrient-rich water from the surrounding area. Drainage since the Middle Ages has resulted in the typical landscape of narrow fields separated by drainage ditches (Fig. 2).
Fig. 2

The typical Dutch peat area landscape of lakes and narrow fields separated by drainage ditches (TOP10vector Product information 2007)

Shallow fresh water bodies are not very well understood in terms of their greenhouse gas emissions and have not been incorporated in previous regional or global greenhouse gas budgets. The emission from these water bodies is probably high and poses an international rather than a domestic problem because in many lowland regions of Europe agriculture continues to contribute appreciably to the nutrient loading of lakes and ditches (Gulati and van Donk 2002; Lamers et al. 1998).

About 16% of the total area (41,864 km2) of the Netherlands is covered by water, mostly classified as wetland (Gulati and van Donk 2002) and with over 300,000 km of drainage ditches. Much of this wetland is in peat areas. It is very important to quantify how these wetlands contribute to the greenhouse gas balance and which factors regulate the emission. In this study we focus on the high-emitting temperate lakes and drainage ditches in peat areas in the Netherlands and on many variables that can alter the emission of CH4 and CO2. The two aims of this study were (1) to quantify CH4 and CO2 fluxes from shallow lakes and drainage ditches in the Netherlands during a 3-week period in the summer season and (2) to identify the factors that regulate the emissions of CH4 and CO2 from lakes and ditches.

Materials and methods

Study sites

Measurements were performed in a 3-week period between June 16th and July 6th in the summer of 2009 in 5 shallow fresh water lakes and 14 drainage ditches at 7 locations in peat areas in the Netherlands (Fig. 3).
Fig. 3

Geographical distribution of the 5 sampled lakes (L1–L5) and 14 sampled drainage ditches (at 7 locations) (D1–D7) in peat areas in the Netherlands

The 5 lakes are located in peat areas in the Netherlands and differ in trophic status (de Haan et al. 1993) and depth (Table 1). L4 and L5 are located in the east of the Netherlands where the subsoil consists of mesotrophic to oligotrophic sedge-peat overlying sand. The other lakes are located in the southwest of the Netherlands where the subsoil consists of eutrophic to mesotrophic reed-sedge peat and alder carr peat.
Table 1

Characteristics of the lakes sampled

Lake

Abbreviation

Trophic status

Lake size (ha)

Average depth (m)

Reeuwijkse plas

L1

Eutrophic

927

2.1

Vinkeveense plas

L2

Eutrophic

1,079

2.4

Nieuwkoopse plas

L3

Eutrohpic

676

2.5

Belterwiede

L4

Mesotrophic to eutrophic

613

1.8

Schutsloterwiede

L5

Mesotrophic to eutrophic

141

1.2

Drainage ditches at 7 locations in different peat areas in the Netherlands were sampled. They differed in trophic status and water depth (Table 2). At each location 2 connected ditches were sampled and because there were no significant differences between them related to water quality they were treated as 1 location in the analyses. All the drainage ditches sampled contained some aquatic vegetation.
Table 2

Characteristics of the drainage ditches

Drainage ditch

Abbreviation

Trophic status

Ditch width (m)

Mean water depth (m)

Oukoop

D1

Eutrophic

8

0.25

Stein

D2

Mesotrophic

6

0.28

Horstermeer

D3

Eutrophic

3

0.90

Drie berken

D4

Mesotrophic

3

0.87

Koole

D5

Mesotrophic

3

0.80

Sint Jans

D6

Eutrophic

2

0.28

Doosje

D7

Mesotrohpic

6

0.43

Measurements

Flux measurements and calculation of fluxes

Detailed measurements of CH4 emission and CO2 emission were performed with floating chambers from a dinghy at different locations in the lakes and drainage ditches. We measured the emissions from each lake on two different days. On each of these days we measured at three different locations per lake, and repeated the measurements five times at each location. This yielded 30 measurements per lake. Each ditch was sampled on 1 day in the 3-week period, with 8 replicates per ditch. This yielded 16 measurements per location in the two connected ditches. All measurements were performed between 10.30 and 14.30 h. Data quality was assessed and outliers resulting from disturbances were removed from the dataset. Emissions of CH4, CO2 and N2O were determined using a closed dark chamber method and a Photo Acoustic Field Gas Monitor (INNOVA 1412 sn, 710-113, ENMO services, Belgium) connected to a PVC chamber by Teflon tubing (e.g. van Huissteden et al. 2005; Hendriks et al. 2007). Fluxes of N2O appeared to be too low to detect with the gas analyzer, therefore the N2O flux measurements were not included in the analyses. Samples were taken from the headspace of this closed cylindrical dark chamber (30 cm diameter, 25 cm height). Gas samples were taken every minute during a 5-min period and every single measurements was checked on linearity of the build up of the gas concentration in the chamber. This check eliminated about 30% of the measurements. The slope dC/dt of the gas concentration curve at time t = 0 was estimated using linear regression (e.g. van Huissteden et al. 2005; Schrier-Uijl et al. 2009b). A small fan was installed in the chamber to homogenise the inside air and a water lock was used to control pressure in the chamber. We used a floater to place the chamber onto the water surface, carefully avoiding the effect of pressure differences and the disturbance of the water surface (for details, see Schrier-Uijl et al. 2009a). Since the gas monitor software does not compensate fully for cross-interference of CO2 and water vapour at high concentrations, air was led through glass tubes filled with silica gel and soda lime before it entered the gas analyser, to remove water vapour. To cross-validate the chamber-based measurements, we also performed eddy covariance measurements on L1 at the same time and location and compared these with the chamber measurements within the footprint of the system. The eddy covariance system was located along a boardwalk in L1 and the footprint of the mast was on the lake. Within this footprint chamber measurements were performed on the lake during a period of 4 h. The two independent methods had previously been compared at different temporal scales in a heterogeneous landscape of fields and ditches (Schrier-Uijl et al. 2009b; Kroon et al. 2007).

Variables measured

At each lake and drainage ditch we measured water temperature and pH at two depths (10 and 30 cm and at 25 cm depth in D1 and D2), dissolved oxygen at 10 cm intervals from the water surface to the sediment surface, and the EC at 10 cm depth. Oxygen, pH, temperature and EC were measured with an HQ multiprobe with a luminescent dissolved oxygen sensor (Hach Company, Loveland, Colorado, USA). The variables investigated in the ditches were the dissolved CH4 concentrations at the water surface, the middle of the water column and in the water immediately above the sediment. Samples for dissolved methane analysis were taken using an airtight 20 ml glass syringe at three depths in the water column: at the sediment surface, at the water surface, and at a depth half-way in between. The water samples were transferred into airtight glass Exetainers® (Labco, high Wycombe, UK) containing 120 μl ZnCl2 to halt biological processes; to prevent air bubbles being trapped in these vials they were filled to overflowing before being capped. The samples were stored in water at 20°C until analysis. Dissolved methane was measured by membrane inlet mass spectrometry (MIMS) (Lloyd and Scott 1983) using an OmniStarTM Gas Analysis System (Pfeiffer Vacuum, Asslar, Germany), equipped with a quadrupole QMS 200 mass spectrometer with a Channeltron detector (Burle Industries). The MS was operated by Quadstar 32-bit software for data acquisition. The sample was pumped through a water bath at 20°C before passing through silicon membrane tubing in which gases were released to the MS. An inlet as described by Kana (1994) was used for the analysis, but without using a cryotrap, as this would have frozen out the methane. Instead, to prevent confounding effects of water vapour, the inlet at the MS side was heated to 180°C. Methane was measured at mass to charge ration (m/z) of 15, as a pre-calibration experiment had shown that this gave the most reliable results. Concentrations of methane were calculated by comparing the ion current at m/z 15 of the sample to the ion current at m/z 15 of air-saturated water at 20°C.

The water in each lake was sampled at three locations with 3 replicates (mixed sample). The water in each ditch was sampled at two locations with 3 replicates (mixed sample). Undisturbed sediment samples were taken from the sediment top layer (upper 10 cm) by means of a plastic cup perforated with holes 2 cm apart at the end of a length-adjustable pipe.

Two of the three water samples were filtered immediately with a Whatman 0.45 μm cellulose membrane filter (Whatman International Ltd, Maidstone, England); the third sample was not filtered. All samples were transported in coolers and stored frozen (−20°C) until analyses. The unfiltered water samples were analysed for organic matter (OM) content, %C, %N, Chlorophyll-a content, total N and total P; the filtered samples were analysed for NO3  + NO2 , NH4 +, SO4 2−, Fe2+ and PO4 3− using a SANplus autoanalyzer (Skalar Analytical, Breda, the Netherlands). Dissolved organic carbon (DOC) and dissolved inorganic carbon (DIC) were measured in filtered samples using a carbon analyser. Total N and total P were measured using a SANplus auto analyser with laser destructor. All these samples were measured in duplicate. Chlorophyll-a content in unfiltered samples from the microcosms was measured using a phytoPAM fluorometer (Heinz Waltz GmbH, Effeltrich, Germany). For the sediment samples a CaCl2 extraction was used to obtain the available PO4–P, NH4–N and NO3–N, and an ammonium oxalate extraction was used to obtain the active form of Fe.

Data analysis

Correlations between the measured variables and fluxes of CO2 and CH4 were first tested by using Pearson correlation analysis. Data were tested for normality. We used stepwise, multiple linear regression analyses to quantify the relationships between environmental variables and fluxes of CO2 and CH4 (SPSS 15.0). The variables that significantly enhanced the emissions of CO2 and CH4 were selected and were used to build regression models. Differences in the fluxes and variables between and within lakes and ditches were tested using one-way ANOVA (SPSS 15.0).

Results

Climatic variables

During the sampling period the mean day air temperatures ranged from 15 to 25°C, the average temperature at the surface of the water bodies studied ranged from 19.2 to 25.4°C and the wind speed at 3 m above water level ranged from 2.1 to 4.5 m s−1 (Table 3).
Table 3

Average (mean ± SD) water temperatures at 10 cm depth (T 10 in °C), at 30 cm depth (T 30 in °C) and average wind speed (U in m s−1) during

 

Mean T 10

Mean T 30

Mean U

D1

20.1 (± 0.1)

19.5 (± 0.9)

4.1

D2

25.4 (± 2.5)

25.2 (± 2.4)

3

D3

19.2 (± 0.6)

18.4 (± 0.4)

2.7

D4

25 (± 0.5)

25.8 (± 0.3)

3

D5

22.3 (± 0.3)

22 (± 0.1)

4.6

D6

20.2 (± 0.7)

19.1 (± 1.4)

2.1

D7

22.8 (± 0.2)

21.8 (± 1.2)

3.9

L1

23.6 (± 3.0)

23.7 (± 0.5)

3

L2

21.4 (± 1.1)

20.3 (± 0.5)

3.6

L3

23.3 (± 1.1)

23.2 (± 1.1)

3

L4

23.2 (± 1.4)

22.7 (± 1.2

3.9

L5

25 (± 1.2)

24.9 (± 0.6)

3.9

For the location codes see Tables 1 and 2

Characteristics of lakes and drainage ditches

The lakes and drainage ditches were humic, shallow and nutrient-rich. The sediment in D6 and D7 had the lowest organic matter content because these two ditches are located in an area with shallow peat on sand. The EC in all the lakes and ditches sampled ranged from 269–866 μS cm−1; the pH ranged from 6.8–9.0, with the highest values in the lakes (Table 4).
Table 4

General characteristics of the lakes and ditches sampled, with standard deviations for lake/ditch depth, EC, pH and percentage organic matter in the sediment

Site

Lake size (ha) or Ditch width (m)

Sediment type

Mean depth (cm)

Mean EC (μS cm−1)

Mean pH

Organic matter sediment (% dry weight)

L1

927

Peat

209.2 (± 42)

508.7 (± 53)

9.0 (± 0.3)

46.63 (± 9.5)

L2

1079

Peat

240.0 (± 42)

866.0 (± 13)

8.4 (± 0.1)

36.89 (± 12.9)

L3

676

Peat

253.3 (± 71)

392.5 (± 2)

8.2 (± 0.1)

63.08 (± 14.1)

L4

613

Peat on sand

178.0 (± 27)

398.0 (± 9)

8.4 (± 0.1)

58.72 (± 3.4)

L5

141

Peat on sand

120.0 (± 21)

393.0 (± 15)

8.4 (± 0.2)

59.81 (± 5.8)

D1

8

Peat

25.0 (± 3)

523.0 (± 38)

9.0 (± 0.0)

55.94 (± 3.0)

D2

6

Peat

27.5 (± 9)

269.3 (± 46)

6.8 (± 0.3)

57.37 (± 7.9)

D3

3

Peat

90.0 (± 14)

662.0 (± 48)

7.0 (± 0.4)

37.77 (± 3.7)

D4

3

Peat

86.7 (± 8)

386.7 (± 2)

7.2 (± 0.3)

74.22 (± 0.8)

D5

3

Peat

80.0 (± 17)

387.0 (± 2)

7.7 (± 0.1)

73.89 (± 0.8)

D6

2

Peat on sand

27.5 (± 3)

350.3 (± 26)

7.2 (± 0.3)

9.53 (± 3.1)

D7

6

Peat on sand

42.5 (± 3)

395.5 (± 6)

8.1 (± 0.3)

2.65 (± 2.4)

Lake area was determined from Top10vector maps: TDN (2006)

Emissions to the atmosphere

Lakes and drainage ditches studied acted as sources of CO2 and CH4 emissions to the atmosphere (Figs. 4, 5), except for L1 where a small uptake of CO2 was measured. The mean release of both gases to the atmosphere was significantly higher from the ditches than from the lakes (P < 0.001).
Fig. 4

Mean CO2 fluxes (mg m−2 h−1) and their standard deviations in the sampling period for the ditches (D1–D7) and lakes (L1–L5) at different locations in peat areas in the Netherlands. Positive flux values indicate release from the water to the atmosphere

Fig. 5

Mean CH4 fluxes (mg m−2 h−1) with their standard deviations for the ditches (D1–D7) and lakes (L1–L5) at different locations in peat areas in the Netherlands. Ditches were sampled on 1 day (n = 16 per location) in the period 16 June–6 July 2009; Lakes were sampled during 2 days in the same period (n = 24 per location). Positive flux values indicate release from the water to the atmosphere

The contribution of CO2 emission compared to CH4 emission in terms of warming potential is given in Fig. 6, where CH4 fluxes have been transformed to CO2 equivalents (CH4 is 23 times as potent as CO2).
Fig. 6

Contribution to greenhouse warming of CO2 emission (dark grey) compared to CH4 emission (light grey) in lakes and ditches, given in CO2 equivalents

Lakes

The emission of CO2 from the lakes (n = 93) ranged from −6.0 to 123.9 mg m2 h−1 and CH4 emission (n = 96) ranged from 1.4–18.1 mg m2 h−1. The CO2 fluxes from L1 were significantly lower than those from the other lakes (P < 0.01); the CO2 fluxes from L2 were significantly higher than those from L3, L1 and L5 (P < 0.05). The highest CH4 emission was measured from L5 and the lowest from L2, but the differences were not significant. The lakes acted as sources of emissions of both gases, except for L1 that acted as a very small sink for CO2. In terms of warming potential, in 3 lakes the dominant emitted greenhouse gas was CH4 and in 2 lakes it was CO2 (Fig. 6).

Ditches

The emission of CO2 from the drainage ditches (n = 80) ranged from 69.6 mg m2 h−1 to 199.0 mg m2 h−1 and CH4 emission (n = 79) ranged from 1.2 to 39.3 mg m2 h−1. The CO2 emission from D3 was significantly higher than the fluxes from D7, D1 and D5. The highest CH4 emission was measured from D4 and the lowest from D7, but the CH4 fluxes did not differ significantly because there was great variability among the ditches. In all ditches except D6 and D7, the dominant greenhouse gas in terms of warming potential was CH4. All ditches acted as sources of emissions of both gases.

Cross-validation

Large-scale CH4 flux measurements by eddy covariance were performed on one of the lakes (L1) to cross-validate flux values from this homogeneous landscape on a diurnal base. A cross-validation of chamber based CH4 and CO2 values and eddy covariance based values is also performed earlier for a more heterogeneous peat area (Schrier-Uijl et al. 2009b). Details for the used eddy covariance instruments have been reported in Veenendaal et al. 2007 and Kroon et al. 2007. In this study, CH4 fluxes within the footprint of the eddy covariance system were 5.8 ± 3.26 (mean ± SD, n = 24) measured by chambers compared to 4.6 ± 1.3 measured by eddy covariance over a 4-h period. It would be of great interest in the future to also use eddy covariance to capture temporal variability of greenhouse gas fluxes (CH4 and CO2) from water bodies and to explain more of the measured variability.

Dissolved oxygen and dissolved CH4

Typical vertical profiles of oxygen saturation during the measurements are shown in Fig. 7.
Fig. 7

Oxygen saturation (%) in the water columns of lakes (right) and drainage ditches (left) at different depths in the water column at the time of the measurements

On average, the lakes had a higher O2 saturation than ditches. In both types of waterbody, oxygen saturation decreased only slightly at the top of the water column, which suggests that there was hardly any respiration by aquatic organisms. Deeper in the water column the oxygen saturation fell rapidly to values close to 0% just above the sediment. Of the lakes, L2 had the highest O2 saturation throughout the profile, and of the ditches D3 and D6 had the lowest O2 saturation.

Dissolved CH4 concentrations were measured at three depths: at the top and middle of the water column and just above the ditch sediments. In all the ditches the dissolved CH4 concentrations increased with depth (Fig. 8).
Fig. 8

Concentrations of dissolved CH4 (μg l−1) in the water of ditches at the top, middle and bottom of the water column. The y-axis shows the depth of the ditch (cm)

Concentrations of dissolved CH4 (μg l−1) in the water of ditches at the top, middle and bottom of the water column. The y-axis shows the depth (cm).

D1 and D3 had a high dissolved CH4 concentration and also a high CH4 emission (Fig. 5). None of the following variables correlated significantly with the dissolved CH4 in the ditch water immediately above the sediment or with the difference between dissolved CH4 concentration at the water surface and at the sediment surface: nutrient content (NO3 , NH4 +, Fe, PO4 3−); sediment oxygen demand (SOD); O2 saturation of the water, organic matter content (% organic matter, %N, %C); amount of green algae and plants. We did not find any significant correlation between dissolved CH4 concentration at the water surface and CH4 release to the atmosphere. The oxygen saturation at the sediment surface correlated negatively with CH4 emission to the atmosphere (P = 0.065).

The variables measured and their correlation with CH4 and CO2 emission

Climate, depth, EC and pH

Climatic conditions in the 3-week sampling period were stable. No significant correlation was found between the CO2 and CH4 fluxes and the temperature or the wind velocity. Neither the depth of water in the ditches (range 0.28–0.90 m) or the depth of water in the lakes (range 1.20–2.53 m) correlated significantly with CO2 or CH4 release to the atmosphere, although the deepest lakes tended to have the lowest CH4 and CO2 emissions. A positive correlation was found between EC and CO2 flux and a significant negative correlation was found between CO2 emission and the pH of the water (r = −0.81; P = 0.001). Though the correlation between CH4 emission and pH was also negative, it was not significant (r = −0.23; P = 0.41).

Nutrients and organic matter in water and sediment

The percentage of N measured in the lake sediments was significantly positively correlated with the release of CO2 (P < 0.01); in ditches, the %N and the %OM measured in the sediments were significantly positively correlated with the release of CH4 (P = 0.02 and P = 0.05, respectively). The lowest organic matter contents of the sediments were found in D6, D7 and L2 (Table 4), which had the lowest CH4 fluxes (Fig. 5). In this study neither the DIC nor the DOC correlated significantly with the CO2 flux or the CH4 flux.

The ammonium (NH4 +) concentration in the water ranged from 0.1 to 478.6 μg NH4–N l−1: the highest concentrations were found in D3, D6 and L2. Ammonium concentration correlated positively with CO2 emission (r = 0.67; P < 0.05). The NO3 concentration in the water was around 0 mg N l−1—except for L2 and L4, where the mean concentrations were 0.43 and 0.12 mg N l−1, respectively. In the sediment of the lakes and ditches the NH4 + concentrations ranged from 12.3 to 478.1 mg/kg dry weight with the highest concentrations in D4 (324.2 mg kg−1 dry weight) and D5 (478.1 mg kg−1 dry weight). The NO3–N concentration ranged from 0.0 to 3.55 mg kg−1 dry weight, with the highest concentration in D4 and the lowest in L4. The only lake with high NO3 concentrations in the water and sediment was L2: it was also the only lake where measurable N2O emissions were observed (0.163 mg m−2 h−1, n = 23). See Table 5 for the concentrations of NH4 + and table 6 for the concentrations of NO3–N.
Table 5

Chemical composition of the water of lakes and ditches with means for sulphate (SO4 2−), iron (Fe2+), phosphate (PO4 3−), ammonium (NH4 +), nitrate (NO3 ) and total P

Location

SO4 2− (mg l−1)

Fe2+ (μg l−1)

PO4 3− (μg P l−1)

NH4 + (μg N l−1)

NO3 (mg N l−1)

Ptot (μg P l−1)

D1

47.1

324.3

98

13.8

0.00

153.4

D2

16.6

864.5

33

0.1

0.00

66.3

D3

3.4

1032.0

291

478.6

0.00

515.7

D4

21.6

58.5

11

8.8

0.00

18.3

D5

22.1

26.7

24

9.2

0.00

15.0

D6

7.1

470.5

240

166.4

0.00

201.0

D7

19.2

146.7

45

3.8

0.00

15.8

L1

34.8

48

53

12.2

0.00

11.3

L2

53.4

59.7

11

76

0.42

14.0

L3

21.0

31.7

25

11.5

0.00

10.3

L4

18.2

113.5

10

15.8

0.12

12.6

L5

16.9

640.0

18

6.1

0.00

25.4

For location codes see Tables 1 and 2

Table 6

Means for Fe, PO4–P, NH4–N, NO3–N in the top10 cm of the bottom sediments of the sampled lakes and drainage ditches

Location

Fe (mg kg−1)

PO4–P (mg kg−1)

NH4–N (mg kg−1)

NO3–N (mg kg−1)

D1

2.7

8.6

154.9

0.4

D2

3.0

0.6

215.4

0.7

D3

7.0

0.3

187.8

0.9

D4

6

20.7

478.1

3.6

D5

3.1

8.8

323.6

0.5

D6

2.0

0.2

51.9

0.2

D7

0.8

0.1

12.3

0.1

L1

1.5

2.6

106.0

0.4

L2

1.7

3.3

8.0

1.2

L3

0.5

5.2

123.7

0.3

L4

5.7

0.5

86.7

0

L5

na

na

na

na

A weak negative correlation was found between the SO4 2− concentrations in lake water (range 17.8–53.4 mg l−1) and ditch water (range 3.4–47.1 mg l−1) and CO2 and CH4 fluxes (r = −0.43, P = 0.16; r = −0.1, P = 0.81). The water of D3 had the lowest SO4 2− concentrations of all the lakes and ditches sampled; this ditch had a high Fe concentration in its water and sediment (Table 6), which suggests the binding of SO4 2− to iron. Overall, the SO4 2− concentrations did not significantly differ between the ditches and lakes. The Pearson correlations between CO2, CH4 and Fe2+ in the water were weakly positive; the Fe2+ concentrations we measured ranged from 33.6 to 1,032 μg l−1, with an average of 301.6 μg l−1. Methane emission correlated significantly positively with the PO4 3− concentration of the sediments of the lakes and ditches (r = 0.77, P = 0.81). The PO4 3− concentration of the sediments correlated positively both with the Fe concentration and the SO4 2− concentration of water. The total P concentration in the water correlated positively with CO2 emission, indicating the high availability of organic substrates. The nutrient concentrations are given in Tables 5 and 6.

Multiple regression analyses

Multiple regression with stepwise elimination of variables showed that for summer CH4 fluxes the PO4 3− concentrations in the sediment and the Fe2+ concentrations in the water explained 87% of the variation when Eq. 1 was used:
$$ F_{{{\text{CH}}_{4} }} = - 3.482 + 2.183\left[ {{\text{PO}}_{4} } \right]_{{{\text{sediment}}}} + 22.116\left[ {\text{FE}} \right]_{\text{water}} $$
(1)
where \( F_{{{\text{CH}}_{4} }} \) is the CH4 flux (mg m−2 h−1), \( \left[ {{\text{PO}}_{4} } \right]_{{{\text{se}}\dim {\text{ent}}}} \)is the PO4–P concentration in the sediment (mg kg−1) and \( \left[ {\text{FE}} \right]_{\text{water}} \) is the concentration of Fe2+ in the water (mg l−1). Table 7 presents statistical details of the model. Fig. 9 shows the measured CH4 fluxes versus the CH4 fluxes in the sampled lakes and ditches, modelled by means of Eq. 1.By performing regression analyses with ditches only, the fit of the regression improved to R2 = 0.94.
Table 7

Statistical details of the CH4 and the CO2 multiple regression models with PO4 concentration in the sediment, Fe concentration in the water, depth, EC and pH as explanatory variables: R square of the model, significance of the model, Pearson correlations and significance of the separate variables

 

R2 model

P model

Pearson corr.

P parameter

PO4sediment

Fewater

 

PO4sediment

Fewater

 

CH4 model

0.87

0.000

0.77

0.19

 

0.000

0.003

 

 

 

 

Depth

EC

pH

Depth

EC

pH

CO2 model

0.89

0.000

−0.62

0.17

−0.81

0.06

0.004

0.001

Fig. 9

Measured CH4 fluxes versus modelled CH4 fluxes (mg m−2 h−1) based on a multiple regression model with PO4–P concentration in the sediment and Fe2+ concentration in the water as explanatory variables. The grey line is the 1:1 line

For CO2, a regression model with mean depth of lake or ditch water and the EC and pH of the water as independent variables explained up to 89% of the variation in summer CO2 emission at the water–atmosphere interface when Eq. 2 was used:
$$ F_{{{\text{CO}}_{ 2} }} = 477.359 - 0.213{\text{depth}} + 0.171{\text{EC}} - 54.4{\text{pH}} $$
(2)
where \( F_{{{\text{CO}}_{ 2} }} \) is the CO2 flux, depth is the mean depth of the water in the sampled lake or ditch (cm), EC is the mean electrical conductivity and pH is the mean pH in the sampled lakes and ditches. Figure 10 shows the measured CO2 fluxes in lakes and ditches versus the modelled CO2 fluxes by means of Eq. 2. Table 7 presents statistical details of the model.
Fig. 10

Measured CO2 flux versus modelled CO2 flux (mg m−2 h−1) based on a multiple regression model with mean depth of water, EC and pH as explanatory variables. The grey line is the 1:1 line

Regression analyses on the data from the ditches only improved the predictive power of the regression to R2 = 0.91.

Discussion

In this study we determined the magnitude of CH4 and CO2 fluxes from 12 water bodies in Dutch wetlands during a 3-week period in the summer season and studied the factors that might regulate emissions of CH4 and CO2 from these lakes and ditches. During this period the lakes and ditches acted as CO2 and CH4 sources of emissions to the atmosphere; the fluxes from the ditches were significantly larger. One lake (L1) was in equilibrium with the atmosphere in terms of CO2 emission. Kosten et al. 2010 found that <10% of lakes worldwide are in equilibrium with the atmosphere in terms of pCO2, and they found that most other lakes are CO2 sources. Compared with other studies, the lake emissions founding our study were in the intermediate to high range (see Table 8 for CH4 fluxes). For example, Rantakari and Kortelainen (2008) found CO2 fluxes in the range 7.48–11.5 mg m−2 h−1 in 37 boreal Finnish lakes. The average CH4 emission from our drainage ditches was higher than the lake fluxes found in other studies (Table 7). As the CH4 emissions measured by the gas analyser within the footprint area of an eddy covariance system at location L2 agreed within the uncertainty limits with the EC system, we are confident that our measurement technique provided reliable flux estimates that are applicable to larger areas (Kroon et al. 2007; Schrier-Uijl et al. 2009b).
Table 8

Comparison between the CH4 emission rates in this study and the CH4 emission rates reported in other studies

Reference

System

Location

Sampling period

Flux CH4 (mg m−2 h−1)

Guerin and Abril 2007

Tropical lake

French Guiana

Late spring

4 ± 4.7

Juutinen et al. 2009

30 Eutrophic Boreal lakes

Finland

All seasons

Median of 0.137

Stadmark and Leonardson 2005

3 Ponds

South Sweden

Summer

10

Huttunen et al. 2002

Boreal lakes

Finland

Summer

1.0

Bastviken et al. 2004

11 Lakes

North America

Summer 2000

Range 0.15–3.2

Repo et al. 2007

3 Boreal lakes

Siberia

Summer

0.34

Present research

5 Temperate lakes

Netherlands

Early summer 2009

1.4–18.1, mean 5.0

Present research

7 Drainage ditches

Netherlands

Early summer 2009

1.2–39.3, mean 18.8

Mean CH4 emission rates are in mg CH4 m−2 h−1. The period of sampling and the location are given

For units: mg CH4 m−2 h−1 refer to mg CH4 emitted per m2 of water area per hour

The temporal variability of emissions of CH4 and CO2 from water bodies is normally found to be related to temperature and wind velocity when measuring over longer time spans (e.g. Stadmark and Leonardson 2005; Frei et al. 2006; Hendriks et al. 2007; Repo et al. 2007; Schrier-Uijl et al. 2009a, b; Kroon et al., in press). However, a large part of the variability of fluxes cannot be explained by temperature or wind velocity only. Our results refer to data collected during summer, a period in which around 70% of the annual ditch emissions are generated. The study did not last long enough to include seasonal patterns of CH4 and CO2 production and emission. Diurnal stratification and mixing due to day–night temperature differences may bias flux estimates if the only measurements available are from the daytime (Repo et al. 2007). In Schrier-Uijl et al. (2009a, b) the diurnal variation of CH4 fluxes over an area with fields and ditches was tested in October/November 2006. After correction for temperature dependency, the emission of CH4 did not differ significantly between day and night. Nevertheless, there could be diurnal variation of fluxes from water bodies, because less oxygen will be produced at night, which will result in a lower redox potential and higher CH4 production. In addition, less CO2 will be taken up by aquatic plants at night, because then they are not photosynthesising. These effects should be considered when estimating annual fluxes from water bodies by using continuous measurements such as eddy covariance. In our study, only diffusive fluxes of CO2 and CH4 were measured; however, ebullition can also contribute to the emission of CH4 from water bodies (Walter et al. 2006; Walter et al. 2007). While sampling ditches D5 and D4, we observed ebullition, so it is possible that we underestimated the release of CH4 fluxes. In a summer study done by Repo et al. (2007) in Siberian water bodies, ebullition was observed in two of the three lakes sampled (depth <1.5 m) and accounted for 19–37% and 11–40% of the total CH4 emissions from these two lakes.

The fact that the lakes and ditches acted as sources for CH4 and CO2 indicates that CO2 production exceeded CO2 uptake during photosynthesis by plants and that CH4 production exceeded CH4 oxidation. Our observation that the deeper, mostly less eutrophic lakes with low EC had the smallest fluxes agrees with findings reported for lakes in the boreal zone in Finland (depth range 3.8–26.5 m) (Juutinen et al. 2009). Deeper water bodies usually have less degradable organic matter and more oxidation of CH4 than shallow lakes, because the transport pathway is longer (e.g. Borges et al. 2004). The EC, which is an indicator of trophic status in fresh water lakes, and the depth of the water body were two of the three significant predictors in the regression analyses for CO2 fluxes.

The pH correlated negatively with emissions from both gases, yet at lower pH values (pH < 7) the correlations are usually positive (e.g. Inubushi et al. 2005). CO2 enters the water as a result of the biological processes of organic carbon degradation and respiration by plants. In our ecosystems it is likely that through uptake of CO2 by plants during the day in the growing season, HCO3 is transformed to CO2, causing the HCO3 concentration to decline and diminishing the buffering effect. This reduced buffering effect can result in pH values above 9.0 and in a negative relation between CO2 flux and pH. Incorporating pH in the regression equation for CO2 significantly improved the equation’s predictive power. In peat soils in temperate areas the optimum pH for methanogenesis is between 5.5 and 7.0, which explains the slightly negative correlation we found between CH4 emission and pH (Le Mer and Roger 2001).

Water turbulence due to wind can increase mixing of oxygen in the water. This is illustrated by the higher oxygen concentration throughout the water column of L2, which had been subjected to high wind speeds on the day before sampling. The high O2 saturation in L2 corresponded with higher CO2 fluxes and very low CH4 fluxes, illustrating the oxidation of CH4 to CO2. The opposite can be seen in D3, D6 and L5, where the oxygen concentrations at the water surface were low and the CH4 fluxes were high, illustrating the low turnover of CH4 carbon to CO2. The fast decrease in dissolved CH4 from the sediments to the water surface in D3 and D1 indicates that most of the dissolved CH4 is oxidised during transport or passes through the water column and escapes to the atmosphere very quickly. As found in other studies, dissolved CH4 poorly predicted the diffusive fluxes at the water–air interface (e.g. Huttunen et al. 2006; Juutinen et al. 2009). The factors responsible for this finding could be variation in duration of storage, release of CH4 to the atmosphere and complex processes during transport of CH4 through the water column (e.g. Kankaala et al. 2003).

The input of organic matter as a substrate in the lake or drainage ditch system increases the availability of substrates and this can increase the production of CO2 and CH4 (e.g. Casper 1992) and increases the possibility of minimising the competition for electron donors between methanogenesis and other anaerobic processes (Scholten et al. 2002; Scheid et al. 2003). As long as O2 reaches the sediments, it will act as the primary oxidant of organic matter.

Permanently anaerobic conditions in the sediment may hamper nitrification of NH4 + to NO3 , but stimulate denitrification of NO3 to N2 by microorganisms, leading to a high NH4 + to NO3 ratio, as was found in this study. Also, the greater availability of NH4 + compared to NO3 suggests the occurrence of dissimilatory reduction of NO3 to NH4 + (DNRA) under anaerobic conditions in these ditches. DNRA is likely to occur in the organic sediments that we sampled, as this process usually occurs at high carbon inputs (Burgin and Hamilton 2007). As our ditch systems did not contain much aquatic plant biomass, it is unlikely that the NO3 uptake by plants and algae was influential in the ditches. Other possible sources of NH4 + in the water could be cation exchange of adsorbed NH4 + by Fe2+ (but this only occurs at very high Fe2+ concentrations: Loeb et al. 2007), and leaching through groundwater from surrounding, managed agricultural areas. NH4 + inhibits methanotrophy and therefore may reduce CH4 oxidation and increase its emission (Conrad and Rothfuss 1991), which may explain the positive correlation between the NH4 + and CH4 fluxes in our study. The positive correlation of NH4 + with CO2 emission is in line with the findings of other studies.

Our finding is that the two most significant predictors of CH4 fluxes were the PO4 3− concentration in the sediment of lakes and ditches and the Fe2+ concentration in the water of lakes and ditches. In anaerobic sediments, Fe3+ will be reduced to Fe2+. At the sediment–water interface some of the Fe2+ will be oxidized to Fe3+ (how much depends on the oxygen concentration just above the sediment) and some of this will be released into the water. Thus a high concentration of Fe2+ in the water is related to anaerobic conditions. Both Fe concentration and SO4 2− concentration correlate with PO4 3− availability at the sediment–water interface. The PO4 3− in sediments is bound to Fe3+ and when the Fe3+ is reduced to Fe2+, PO4 3− will be released to the water (e.g. Smolders et al. 2006; Smolders and Roelofs 1993).

Overall, a higher trophic status was positively correlated with summer emissions of CO2 and CH4, while the depth of the water and the pH were inversely correlated with CO2 emission. It is therefore likely that decreasing the inputs of organic matter and nutrients (for example, by changing the management of the surrounding areas) will reduce emissions and that this effect will be strongest in drainage ditches.

Much of the uncertainty in flux estimates is due to temporal variation. So, also diurnal, seasonal, annual and inter annual variability must be studied in more detail to get insight in climatic responses, extreme drought/rainfall events and the influence of management in the surrounding catchments. In this respect, there is a need for long-term, continuous measurements of emissions (e.g. by eddy covariance).

Conclusion

The current study focused on emissions from temperate, shallow lakes (n = 5) and drainage ditches (n = 14) in agricultural peat areas in the Netherlands. It was found that in general, both these types of waterbodies are important sources of CO2 and CH4. The ditches had significantly higher CO2 and CH4 fluxes than the lakes. Trophic status was an important indicator of the magnitude of fluxes. 87% of the variation in the summer fluxes of CH4 could be explained by PO4 3− in the sediment and Fe2+ concentration in the water, and 89% of the CO2 flux could be explained by water depth, EC and pH. Our results can be used to refine greenhouse gas emission inventories and to ascertain possible ways of reducing the release of CO2 and CH4 from water bodies to the atmosphere. Decreasing the nutrient loads and input of organic substrates to ditches and lakes will likely reduce summer emissions of CO2 and CH4 from these water bodies.

Notes

Acknowledgements

Many thanks to Roos Loeb and Joop Harmsen for their critical comments on an earlier version of this manuscript, and to Jan van Walsem, Frits Gilissen and John Beijer for their assistance in the field and in the laboratory. We are also grateful to the State Forestry Service for the use of their boats and to Arjan Hensen and Petra Kroon who carried out the eddy covariance measurements at L1. Joy Burrough advised on the English. This study was funded by Wageningen University, The Province of North Holland, CarboEurope-IP and the Klimaat voor Ruimte Project (BSIK).

Open Access

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.

References

  1. Bastviken D, Ejlertsson J, Tranvik L (2002) Measurement of methane oxidation in lakes: a comparison of methods. Environ Sci Tech 36:3354–3361CrossRefGoogle Scholar
  2. Bastviken D, Cole J, Pace M et al (2004) Methane emissions from lakes: dependence of lake characteristics two regional assessments and a global estimate. Glob Biogeochem Cycles 18:GB4009. doi:  10.1029/2004GB002238
  3. Borges AV, Delille B, Schiettecattle LS et al (2004) Gas transfer velocities of CO2 in three European estuaries (Randers Fjord, Scheldt, and Thames). Limnol Oceanogr 49(5):1630–1641Google Scholar
  4. Burgin AJ, Hamilton SK (2007) Have we overemphasized the role of denitrification in aquatic ecosystems? A review of nitrate removal pathways. Front Ecol Environ 5(2):89–96CrossRefGoogle Scholar
  5. Casper P (1992) Methane production in lakes of different trophic state. Arch Hydrobiol Beih Ergebn Limnol 37:149–154Google Scholar
  6. Casper P, Chan OC, Furtrado ALS et al (2003) Methane in an acidic bog lake: the influence of peat in the catchment on the biogeochemistry of methane. Aquat Sci 65:36–46. doi: 1015-1621/03/010036-11 CrossRefGoogle Scholar
  7. Cole JJ, Caraco NF, Kling GW et al (1994) Carbon-dioxide supersaturation in the surface waters of lakes. Science 265(5178):1568–1570CrossRefGoogle Scholar
  8. Conrad R, Rothfuss F (1991) Methane oxidation in the soil surface-layer of a flooded rice field and the effect of ammonium. Biol Fert Soils 12(1):28–32CrossRefGoogle Scholar
  9. De Haan H, van Liere L, Klapwijk SP et al (1993) The structure and function of fen lakes in relation to water table management in The Netherlands. Hydrobiologia 265:1–3. doi: 10.1007/BF00007266 Google Scholar
  10. Frei M, Razzak MA, Hossain MM et al (2006) Methane emissions and related physicochemical soil and water parameters in rice-fish systems in Bangladesh. Agric Ecosyst Environ 120:391–398. doi: 10/1016/j.agee CrossRefGoogle Scholar
  11. Guerin F, Abril G (2007) Significance of pelagic aerobic methane oxidation in the methane and carbon budget of a tropical reservoir. J Geophys Res Biogeosci 112(G3). Art. number GO3006Google Scholar
  12. Gulati R, van Donk E (2002) Lakes in the Netherlands their origin eutrophication and restoration: state-of-the-art review. Hydrobiologia 478:73–106CrossRefGoogle Scholar
  13. Hendriks DMD, van Huissteden J, Dolman AJ et al (2007) The full greenhouse gas balance of an abandoned peat meadow. Biogeosci Discuss 4:277–316CrossRefGoogle Scholar
  14. Huttunen JT, Väisänen T, Heikkinen M et al (2002) Exchange of CO2 CH4 and N2O between the atmosphere and two northern boreal ponds with catchments dominated by peatlands or forests. Plant Soil 242:137–146CrossRefGoogle Scholar
  15. Huttunen JT, Väisänen T, Helllsten SK et al (2006) Methane fluxes at the sediment-water interface in some boreal lakes and reservoirs. Boreal Environ Res 11:27–34. ISSN:1239-6095Google Scholar
  16. Inubushi K, Otake S, Furukawa Y et al (2005) Factors influencing methane emission from peat soils: comparison of tropical and temperate wetlands. Nutr Cycl Agroecosyst 71:93–99CrossRefGoogle Scholar
  17. Juutinen S, Rantakari M, Kortelaine P et al (2009) Methane dynamics in different boreal lake types. Biogeosciences 6:209–223CrossRefGoogle Scholar
  18. Kana TM (1994) Membrane inlet mass spectrometer for rapid high-precision determination of N2 O2 and Air in environmental water samples. Anal Chem 66:4166CrossRefGoogle Scholar
  19. Kankaala P, Taipale S, Nykanen H et al (2003) Oxidation, efflux, and isotopic fractionation of methane during autumnal turnover in a polyhumic, boreal lake. J Geophys Res Biogeosci 112:G2 G02003Google Scholar
  20. Kosten S, Roland F, Da Motta Marques DML, Van Nes EH, Mazzeo N, Sternberg LDSL, Scheffer M, Cole JJ (2010) Climate-dependent CO2 emissions from lakes. Glob Biogeochem Cycles. doi: 10.1029/2009GB003618
  21. Kroon PS, Hensen A, Zahniser MS et al (2007) Suitability of quantum cascade laser spectrometry for CH4 and N2O eddy covariance measurements. Biogeosciences 4:715–728CrossRefGoogle Scholar
  22. Kroon PS, Schrier-Uijl AP, Hensen A et al (in press) Annual balances of CH4 and N2O from a managed fen meadow using eddy covariance flux measurements. Eur J Soil SciGoogle Scholar
  23. Lamers PM, van Roozendaal ME, Roelofs GM (1998) Acidification of freshwater wetlands: combined effects of non-airborne sulfur pollution and desiccation. Water Air Soil Pollut 105:95–106CrossRefGoogle Scholar
  24. Le mer J, Roger P (2001) Production oxidation emission and consumption of methane by soils: a review. Eur J Soil Biol 37(1):25–50CrossRefGoogle Scholar
  25. Lloyd D, Scott RI (1983) Direct measurement of dissolved gases in microbiological systems using membrane inlet mass spectrometry. J Microbiol Methods 1:313–328CrossRefGoogle Scholar
  26. Loeb R, van Daalen E, Lamers LPM et al (2007) How soil characteristics and water quality influence the biogeochemical response to flooding in riverine wetlands. Biogeochemistry 85:289–302. doi: 10.1007/s1-533-007-9135-x CrossRefGoogle Scholar
  27. Minkkinen K, Laine J (2006) Vegetation heterogeneity and ditches create spatial variability in methane emissions from peatlands drained for forestry. Plant Soil 285:289–304CrossRefGoogle Scholar
  28. Rantakari M, Kortelainen P (2005) Interannual variation and climatic regulation of the CO2 emission from large boreal lakes. Glob Change Biol 11:1368–1380CrossRefGoogle Scholar
  29. Rantakari M, Kortelainen P (2008) Controls of organic and inorganic carbon in randomly selected Boreal lakes in varied catchments. Biogeochemistry 91(2–3):151–162CrossRefGoogle Scholar
  30. Repo ME, Huttunen JT, Naumov AV et al (2007) Release of CO2 and CH4 from small wetland lakes in western Siberia. Tellus 59:788–796. doi: 10.1111/j.1600-0889.2007.00301.x CrossRefGoogle Scholar
  31. Roulet NT, Moore TR (1995) The effect of forestry drainage practices on the emission of methane from northern peatlands. Can J For Res 25(3):491–499CrossRefGoogle Scholar
  32. Saarnio S, Winiwater W, Leitão J (2009) Methane release from wetlands and watercourses in Europe. Atmos Environ 43:1421–1429. doi: 10.1016/j.atmosenv.2008.04.007 CrossRefGoogle Scholar
  33. Scheid D, Stubner S, Conrad R (2003) Effects of nitrate- and sulfate-amendment on the methanogenic populations in rice root incubations. Fems Microbiol Ecol 43(3):309–315CrossRefGoogle Scholar
  34. Scholten JCM, Conrad R, Stams AJM (2002) Effect of 2-bromo-ethane sulfonate molybdate and chloroform on acetate consumption by methanogenic and sulfate-reducing populations in freshwater sediment. Fems Microbiol Ecol 32(1):35–42Google Scholar
  35. Schrier-Uijl AP, Kroon PS, Leffelaar PA et al (2009a) Methane emissions in two drained peat agro-ecosystems with high and low agricultural intensity. Plant Soil. doi: 10.1007/s11104-009-0180-12009 Google Scholar
  36. Schrier-Uijl AP, Kroon PS, Hensen A et al (2009b) Comparison of chamber and eddy covariance based CO2 and CH4 emission estimates in a heterogeneous grass ecosystem on peat. Agric For Meteorol. doi: 10.1016/j.agrformet.2009.11.007 Google Scholar
  37. Smolders AJP, Roelofs JGM (1993) Sulphate-mediated iron limitation and eutrophication in aquatic ecosystems. Aquat Bot 46:247–253CrossRefGoogle Scholar
  38. Smolders AJP, Lamers LPM, Lucassen ECHET et al (2006) Internal eutrophication: how it works and what to do about it—a review. Chem Ecol 22(2):93–111CrossRefGoogle Scholar
  39. St. Louis VL, Duchemin CAE, Rudd JWM et al (2000) Reservoir surfaces as sources of greenhouse gases to the atmosphere: a global estimate. Bioscience 50:766–775. doi: 10.1641/0006-3568(2000)050[0766:RSASOG CrossRefGoogle Scholar
  40. Stadmark J, Leonardson L (2005) Emissions of greenhouse gases from ponds constructed for nitrogen removal. Ecol Eng 25:542–551. doi: 10.1016/j.ecoleng.2005.07.004 CrossRefGoogle Scholar
  41. TOP10vector Product information (in Dutch). Topografische Dienst Kadaster Emmen the Netherlands. Available at http://www.tdn.nl/ (verified 6 Dec 2007)
  42. van Bodegom PM, Scholten JCM (2001) Microbial processes of CH4 production in a rice paddy soil: model and experimental validation. Geochemica et Cosmochimica Acta 65(13):2055–2066CrossRefGoogle Scholar
  43. van Huissteden J, Maximov TC, Dolman AJ (2005) High methane flux from an arctic floodplain (indigirka lowlands eastern siberia). J Geophys Res Biogeosci 110. Art number G02018Google Scholar
  44. Veenendaal EM, Kolle O, Leffelaar PA et al (2007) CO2 exchange and carbon balance in two grassland sites on eutrophic drained peat soils. Biogeosciences 4:1027–1040CrossRefGoogle Scholar
  45. Walter KM, Zimov SA, Chanton JP (2006) Methane bubbling from Siberian thaw lakes as a positive feedback to climate warming. Nature 443:71–75. doi: 10.1038/nature05040 CrossRefGoogle Scholar
  46. Walter KM, Smith LC, Chapin FS (2007) Methane bubbling from northern lakes: present and future contributions to the global methane budget. Phil Trans R Soc 365:1657–1676CrossRefGoogle Scholar
  47. Wang HJ, Lu JW, Wang WD (2006) Methane fluxes from the littoral zone of hyper eutrophic Taihu Lake China. J Geophys Res 111:D17109. doi: 10.1029/2005JD006864 CrossRefGoogle Scholar
  48. Whiting GJ, Chanton JP (2001) Greenhouse carbon balance of wetlands: methane emissions versus carbon sequestration. Tellus 53B:521–528. ISSN 0280-6509 2001Google Scholar

Copyright information

© The Author(s) 2010

Authors and Affiliations

  • A. P. Schrier-Uijl
    • 1
  • A. J. Veraart
    • 2
  • P. A. Leffelaar
    • 3
  • F. Berendse
    • 1
  • E. M. Veenendaal
    • 1
  1. 1.Nature Conservation and Plant Ecology GroupWageningen UniversityWageningenThe Netherlands
  2. 2.Department of Aquatic Ecology and Water Quality ManagementWageningen UniversityWageningenThe Netherlands
  3. 3.Plant Production Systems GroupWageningen UniversityWageningenThe Netherlands

Personalised recommendations