Atmospheric methane removal by boreal plants



[1] Several studies have proposed aerobic methane (CH4) emissions by plants. If confirmed, these findings would further increase the imbalance in the global CH4 budget which today underestimates CH4sinks. Oxidation by OH-radicals in the troposphere is the major identified sink followed by smaller contribution from stratospheric loss and oxidation by methano- and methylotrophic bacteria in soils. This study directly investigated CH4 exchange by plants in their natural environment. At a forest site in central Sweden, in situ branch chamber measurements were used to study plant ambient CH4 exchange by spruce (Picea abies), birch (Betula pubescens), rowan (Sorbus aucuparia) and pine (Pinus sylvestris). The results show a net uptake of CH4 by all the studied plants, which might be of importance for the methane budget.

1. Introduction

[2] The first report of CH4 emissions by green plants by Keppler et al. [2006] proposed global CH4 emissions by vegetation in the range of 62 to 236 Tg each year, which would comprise 10–30% of the total CH4 entering the atmosphere [Lowe, 2006]. Several subsequent studies supported the findings of aerobic CH4 emissions [e.g., Wang et al., 2008; McLeod et al., 2008; Vigano et al., 2008; Messenger et al., 2009], but the extrapolation of the results by Keppler using net primary production as a scalar has been criticized. A more conservative estimate was made using biome leaf biomass and photosynthesis as scaling factors, which resulted in emissions in the range of 10–60 Tg yr−1 [Kirschbaum et al., 2006]. A number of other extrapolations have been made [Houweling et al., 2006; Butenhoff and Khalil, 2007; Ferretti et al., 2007] all resulting in a lower range than that of Keppler et al. Ultraviolet (UV) radiation has been shown to drive an aerobic production of CH4 from plant tissue and pectin [McLeod et al., 2008; Vigano et al., 2008]. Estimates of global foliar CH4emissions from UV-irradiated pectin resulted in a value of 1 Tg yr−1 [Bloom et al., 2010]. UV driven demethylation of the methoxyl groups of pectin in the cell wall of the plant has been suggested as a potential source of CH4 [McLeod et al., 2008; Keppler et al., 2008; Messenger et al., 2009]. UV radiation is indicated to act on a photosensitizer, which generates reactive oxygen species (ROS), which in turn attack pectic polysaccharides so that CH4 formation occurs [Messenger et al., 2009]. Besides UV-radiation, physical injury of the plant [Z.-P. Wang et al., 2009] and other environmental stress factors might also stimulate ROS activity [Z.-P. Wang et al., 2009; Qaderi and Reid, 2011; Wishkerman et al., 2011]. At the same time it is well established that ROS can react with water to form OH [Logan et al., 1981], the principal atmospheric sink of CH4. The CH4 emissions from plants have also been suggested to stem from dissolved CH4 in the water drawn into the plant and subsequently released through diffusion [Dueck et al., 2007] or transpiration [Nisbet et al., 2009].

[3] The importance of the emissions and the mechanisms through which they are released are still discussed [Bruhn et al., 2012]. Some studies have not found evidence for substantial aerobic CH4 emissions [Dueck et al., 2007; Beerling et al., 2008; Kirschbaum and Walcroft, 2008]. There is also little evidence from in-situ studies on plant emission. Comparison of soil chamber measurements between plots with intact vegetation and removed vegetation led to increased production of CH4 in the intact plot at a tropical savanna [Sanhueza and Donoso, 2006] and at two alpine grasslands [Cao et al., 2008; S. Wang et al., 2009]. However Cao et al. also found in their study that alpine shrubs instead consumed CH4 and S. Wang et al. [2009] explained the discrepancy found in their study with differences in soil water content and soil temperature. There have also been a few studies on methane exchange by canopies. Mikkelsen et al. [2011]found indications of periodic canopy emissions using a gradient-based method whereas two studies using eddy covariance measurements didn't find any significant emission from the canopy [Bowling et al., 2009; Smeets et al., 2009].

[4] Sphagnum spp. mosses are the only plants so far reported to consume CH4 by symbiosis with partly endophytic methanotrophic bacteria. Carbon dioxide (CO2) produced from oxidation of CH4 is then fixed to the plant during photosynthesis [Raghoebarsing et al., 2005].

[5] Numerous plants have epiphytic bacterial associations, in this case pink-pigmented facultative methyltrophs (PPFMs) that can consume methanol emitted by plants [Holland and Polacco, 1994]. One strain of these bacteria, methylobacterium sp. BJ001T that was isolated from poplar tree tissue has been reported to be able to use CH4 as a sole source of carbon and energy [Van Aken et al., 2004]. This report was not considered definitive [Dedysh et al., 2004].

2. Method

[6] We directly measured CH4 exchange by boreal plants with the ambient atmosphere in a boreal forest (Table 1). Semi-continuous field measurements on branches of different plants were made to study direct CH4exchange by plants and its dependence on photosynthetically active radiation (PAR), ultraviolet radiation (UV-radiation), temperature and photosynthesis. Because of the strong correlation between different environmental parameters, e.g., radiation and temperature, under natural conditions, additional laboratory measurements were also made to study the controlling factors. The CH4 exchange measurements were made on spruce (Picea abies), birch (Betula pubescens), rowan (Sorbus aucuparia) and pine (Pinus sylvestris) in the 100-year-old Norunda forest stand in central Sweden at 60°5′ N, 17°29′ E. In 2009, the measurements took place in a forest stand that was thinned in 2008, whereas measurements in 2010 were in an untouched part of the forest. The measurements on birch in 2009 were made during the senescent period because environmental stresses had been posited to play an important role in affecting emissions [Z.-P. Wang et al., 2009; Qaderi and Reid, 2011; Wishkerman et al., 2011] (see Table 1 for details on measurements periods). Since the branch chamber measurements were integrated in a larger system with measurements of methane oxidation in the soil and methane gradients within the forest, the branches had to be selected within a limited area with the requirement of being reachable from the ground, but in a position that is reached by the sun at least part of the day. This explains the low PAR range for part of the measurement periods.

Table 1. Details on Measurement Periodsa
Plant TypeType of MeasurementMeasuring PeriodMeasuring FrequencyNumber of Recordings Available for Analysis (n)Plant Height (m)Shoot Leaf Area (m2)Temperature Range (°C)PAR Range (μmol m−2 s−1)GPP Range (μmol m−2 s−1) for PAR > 2UV Range (μmol m−2 s−1)
  • a

    Information on plant sample and ranges of temperature, PAR, GPP and UV-radiation. 90% of the data fall within the Temperature, PAR, GPP and UV-radiation ranges that are shown in the table.

Spruce1In situ13/8–15/9 20091/hour285∼200.0317−190−435−11−00−2.8
Birch1In situ16/9–19/10 20091/ hour361∼2.50.019−3−160−140−6−40−1.5
Spruce2In situ7 /7–28/7 20101/ hour399∼200.01713−300−117−7−20−1.1
Spruce3In situ29 /7–12/8 20101/ hour281∼10.02613−220−76−7−10−1
RowanIn situ27/8–10/9 20101/ hour279∼4.50.0085−160−80−13−20−0.8
Birch2In situ10/9–24/9 20101/ hour203∼10.0116−170−33−5−10−0.4
Pine1In situ29/9–4 /10 20101/ hour97∼0.40.0031−120−36−10−60−0.5
Spruce4Lab24 hours Feb 20114/ hour65∼0.50.0085−250−382−25−8-
Spruce5Lab24 hours Feb 20114/ hour61∼0.50.01475−250−358−17−4-
Spruce6Lab72 hours Feb 20114/ hour198∼0.50.01685−260−380−11−2-
Pine2Lab24 hours Feb 20114/ hour70∼0.30.0055−250−327−25−4-
Pine3Lab24 hours Feb 20114/ hour139∼0.30.00195−260−375−100−14-

[7] A temperature controlled, automated branch chamber was used in combination with a high precision off-axes ICOS lazer (LGR-Los Gatos Research) to determine changing mixing ratios of CH4 in the chamber headspace. The sides of the branch chamber exposed to the sun were quartz glass, which is transparent to UV radiation. The volume of the chamber was 0.0057 m3. The chamber sealed around the stems of the leaves or needles being studied so that the leaves remained intact throughout the measurements. Changes of headspace concentration were measured after closure of the chamber by recirculating the air through the gas analyzer during a period of 5 min every hour. During closure, the air inside the chamber was kept to within ±1°C of the ambient air, and a fan was used to mix the headspace air. The cooling of the air was provided with a peltier cooler controlled by a CR 1000 data logger (Campell Inc., Logan, USA). The measuring frequency was 0.1 Hz in 2009 and 1 Hz in 2010. Besides CH4, the analyzer measured CO2 in 2009 and CO2 and H2O in 2010. The flow rate through the analyzer was 1.2 l/min.

[8] Before the measurements started, the chamber was put in an oven for 48 hours at 70°C to prevent possible CH4 emissions from the chamber itself. The chamber was also run empty, but with equal procedures as the rest of the measurements, for a couple of weeks in 2009 to make sure there is no spurious CH4 flux. The mean of the CH4 exchange for this period was not significantly different from zero.

[9] In February 2011, laboratory measurements were made on horticultural specimens of spruce (Picea glauca conica) and pine (Pinus mugo var. pumilio). Samplings of the same species as measured in the field where not available at this time of the year. Exactly the same equipment was used as for the in situ measurements. The only difference was that flux measurements were performed four times per hour. The CH4 exchange was tested at three temperature intervals, 3–7°C, 13–17°C and 23–27°C and four different intervals of PAR, 0 μmol m−2 s−1, 50–150 μmol m−2 s−1, 250–350 μmol m−2 s−1 and 350–450 μmol m−2 s−1. The measured CH4concentrations had to be corrected for dilution effects by water vapor, but since the water vapor was not measured in 2009 it had to be calculated . We used the Ball-Berry equation, to calculate the stomatal conductance toCO2, gs (μmol m−2 s−1), as: math formula where g0 = 0.01 · 106 (μmol m−2 s−1) is the stomatal conductance [Collatz et al., 1991], m = 8.75 (dimensionless) is a fitting parameter [Xu and Baldocchi, 2003]. A (μmol m−2 s−1) is photosynthesis rate, h (dimensionless) is relative humidity and Cs (μmol CO2 (mol air)−1) is the mixing ratio of CO2 in air at the leaf surface. Cs is calculated as math formula [Liu et al., 2009], where Ca (μmol CO2 (mol air)−1)) is the ambient CO2 concentration in the air and gb = 0.075 m s−1 is the boundary layer conductance of the shoot, assuming 1 m s−1 wind speed in the chamber [Martin et al., 1999]. The transpiration rate E (μmol m−2 s−1) is calculated as: math formula, where math formula and gs (μmol m−2 s−1) is the stomatal conductance to water vapor, es (kPa) is the saturation vapor pressure at air temperature, Ta (°C), ea (kPa) is the ambient vapor concentration and Pa (kPa) is the air pressure. We then used the measured ambient water vapor mixing ratio in the nearby tower as the initial concentration at time for chamber closure and then we added transpired water to calculate the successively increasing water vapor concentration in the chamber. As the next step, we calculated the dry concentration of CH4 for each time step as math formula, where w (mmol mol−1) is the water vapor molar ratio. In 2010 w was measured and the correction could be done directly.

[10] The CH4 mixing ratio obtained by the analyzer and after dilution corrections, c (μmol mol−1), was then converted to C (μmol m−3) by C = c · P/(R · T), where P = 101325 Pa is the standard atmospheric air pressure, R = 8.314 J mol K−1 is the gas constant and T (K) is the chamber temperature.

[11] From the concentration data, a linear fit was made for a two-minute interval beginning immediately after chamber closure to retrievedC/dt. We calculated the r2 values for the fits of five different intervals, which were slightly offset to each other and selected the fit with the highest r2 value. All fluxes with an r2 value significant at 95% level where kept for further analyses. The CH4 flux (JCH4flux) was calculated as math formula, where V (m3) is the chamber volume and A (m2) is the projected leaf area of the enclosed shoot.

[12] Mean CH4 exchange and spearman correlations between CH4 exchange and PAR, temperature, UV and GPP were calculated. The whisker length of 1.5 in the box and whisker plot (see Figure 2) is suggested by McGill et al. [1978].

3. Results

[13] The results from the branch chamber measurements show a significant mean uptake of CH4 by all studied plants (Figure 1).

Figure 1.

CH4 exchange (μmol m−2 h−1) for 12 measuring periods. The data are expressed per unit (m2) of leaf area. The middle line of the box and whisker plot represents the median of all recordings including nighttime measurements. The edges of the box are the 25th and 75th percentiles, the whiskers, (black dotted lines) are the extreme values not considered outliers. Values larger than q3 + w (q3–q1) or smaller than q1–w (q3–q1) are considered outliers, where q1 and q3 are the 25th and 75th percentiles, respectively, and w = 1.5 is the whisker length. Negative values represent a flux from the atmosphere towards the branch. The median values are all significantly different from zero at 99% significance level.

[14] There was a consistent, small but significant negative correlation between CH4 exchange and PAR for six of the seven species measured in situ with increasing CH4 uptake for higher values of PAR (Figure 2 and Table 2). The measurements on Pine1 were not significantly correlated with PAR. The correlation coefficients for CH4exchange and UV-radiation were similar to those of PAR with increasing CH4uptake for higher values of UV-radiation (Table 2). Only three of the seven in situ measurements, Spruce1, Spruce2 and Birch1, were significantly and positively correlated with temperature (Table 2). The average CH4 uptake per unit of leaf area across all species and environmental conditions for the in situ measurements was 0.7 μmol m−2 h−1.

Figure 2.

The correlation of CH4 flux (μmol m−2 h−1) with PAR (μmol m−2 s−1) for the seven different shoots studied in situ. Only PAR values larger than 5 μmol m−2 s−1are included in the plots. The c-value is the correlation coefficient. * Indicates that the correlation is significant at 95% significance level.

Table 2. Correlation Between CH4 Exchange and PAR, CH4 Exchange and Temperature and CH4 Exchange and UV Radiation for in Situ Measurementsa
Plant TypenCorrelation Coefficient, CH4 and PARCorrelation Coefficient, CH4 and TemperatureCorrelation Coefficient, CH4 and UV-Radiation
  • a

    The correlation with PAR and UV is calculated for PAR-values larger than 5μmol m−2 s−1.

  • b

    Significant at 95% level.


[15] In the laboratory measurements, CH4 exchange was significantly (95%) correlated with PAR (range 0–450 μmol m−2 s−1) for all measurements with correlation coefficients of −0.23, −0.28 and −0.22 for temperature intervals 3–7, 13–17 and 23–27°C, respectively (Table 3). CH4uptake was positively correlated with temperature at 90% significance level with correlation coefficients 0.14 and 0.19 for PAR intervals -5-5 and 250–350 respectively, and negatively correlated with temperature with correlation coefficients −0.01 (p = 0.9), and −0.19 (p = 0.1) for PAR intervals 50–150 and 350–450 respectively (Table 4).

Table 3. Correlation Between CH4 Exchange and PAR at Fixed Temperature Intervals for All Species Measured in Laboratory
Temperature Range (°C)nCorrelation Coefficient, CH4 and PARp
Table 4. Correlation Between CH4 Exchange and Temperature at Fixed PAR Intervals for All Species Measured in Laboratory
PAR Range (μmol m−2 s−1)nCorrelation Coefficient, CH4 and Temperaturep

[16] The correlation of CH4 exchange with GPP was better for the laboratory measurements than for the in situ measurements. Of the in situ measurements, only Spruce1 and Birch1 showed a significant positive correlation between CH4 exchange and GPP with correlation coefficients of 0.57 (Figure 3) and 0.52 respectively (Table 5).

Figure 3.

Correlation between CH4 (μmol m−2 h−1) exchange and GPP (μmol m−2 s−1) for Spruce1. The correlation coefficient is 0.57. Notice that a negative GPP means uptake from the atmosphere.

Table 5. Correlation Between CH4 Exchange and GPP for All Plants Studieda
Plant Typen-ValueCorrelation Coefficient, CH4 and GPP
  • a

    GPP is calculated for PAR > 2.

  • b

    Significant at 95% level.


4. Discussion

[17] In contrast to earlier studies of CH4 exchange by plants, we find a net consumption by all plants studied both in situ and in the laboratory. The presence of endophytic or epiphytic bacteria with the ability to consume CH4 would be a possible explanation for this [Raghoebarsing et al., 2005; Van Aken et al., 2004]. The timescale of the atmospheric OH sink is too large for it to be an explanation.

[18] The strong correlation between PAR and temperature in situ makes it difficult to sort out which parameter has the strongest control. The laboratory measurements gave more distinct results: the uptake increased consistently with increasing PAR while there was no consistent control on the uptake by temperature (Tables 3 and 4). This indicates that the CH4 sink is located somewhere inside the leaves and that the diffusion rate is controlled by the stomatal conductance. Stomatal conductance increases practically linearly with light at the range of PAR encountered in this study. The temperature response found for a several PAR intervals with a decreasing uptake for increasing temperatures could also be explained by a response to vapor pressure deficit (VPD) with stomata responding in the opposite direction, i.e., closing when temperature (and VPD) increases.

[19] What speaks against a microbial process is the negative correlation between CH4 consumption and temperature found in the laboratory studies since bacterial activities are normally favored by increasing temperatures. However, if it is the diffusion rate through the stomata that is the limiting factor for bacterial consumption, this hypothesis could still be valid.

[20] The emissions of CH4 by plants that have been reported in several studies are thought to depend on environmental stresses that activate ROS. It is possible that this emission process occurs simultaneously with a consumption of CH4 by bacteria. High levels of ROS would then control the CH4 exchange leading to net emissions. Most plants in their natural environment however, might not experience such a high stress level. The absence of UV radiation in the laboratory might explain the higher CH4 uptake and the higher correlation of the uptake with GPP in this particular environment.

[21] Our observations also contain positive CH4 fluxes particularly at low GPP (or PAR), but this is mainly attributed to the uncertainty in the measurements. Under the condition of very low fluxes, the uncertainty in measured concentrations caused by signal noise, become significant.

[22] If we scale up our observed in situ uptake of CH4 by multiplying the mean value of 0.7 μmol m−2 h−1 with the leaf area index of the Norunda forest (4.8) we obtain a value of 3.4 μmol m−2 h−1. This is close to the mean CH4 soil oxidation rate in various forest soils based on 28 studies, which is 4.0 μmol m−2 h−1 [Jang et al., 2006], this indicates that the canopy might play an equally important role as the soil in the global context.

[23] Two recent studies give alternative explanations to the slow-down in the growth rate of atmospheric methane in the last decades. One of them indicates that it is due to a stabilization of fossil-fuel emissions [Aydin et al., 2011] whereas the other explains it by a decrease in microbial methane sources in the northern hemisphere [Kai et al., 2011]. Our results offer a third explanation: that an increasing amount of CH4 has been taken up by vegetation during the last decades as a consequence of increased greenness [Myneni et al., 1997], NPP [Nemani et al., 2003] and GPP [Chen et al., 2006] as observed by satellite remote sensing.


[24] Support for this work was provided by Formas and by the Linnaeus Centre LUCCI ( funded by the Swedish Research Council. We thank Anders Båth for field assistance.

[25] The Editor thanks the anonymous reviewer for assistance in evaluating this paper.