It is hard to avoid hearing about the greenhouse effect following the documentary ‘Inconvenient Truth’ and the Stern Report. Methane (CH4) is the second most potent greenhouse gas after carbon dioxide (CO2), and currently contributes 0.48 W m−2 to anthropogenic radiative forcing compared with 1.66 W m−2 from CO2 (Intergovernmental Panel on Climate Change (IPCC), 2007). Methane has traditionally been thought to originate from anaerobic decomposition in swamps and ruminants with additional release from fossil fuel deposits. However, Keppler et al. (2006) measured methane emission from plants and plant products under aerobic conditions. This controversial finding sparked discussion and reassessment of the global methane budget. The claim that plants emit methane under aerobic conditions has now been tested by growing plants in an atmosphere containing 13CO2 and looking for 13C-labelled methane. This was the objective of the study by Dueck et al. reported in New Phytologist (2007).
‘Compiling a global methane budget is a complex task that is inherently uncertain.’
Methane in the Earth's atmosphere has risen dramatically from a preindustrial concentration of 715 ppb to 1760 ppb in 2000. Unlike CO2, the majority of methane emitted each year is destroyed, with the remaining 4% adding to the atmospheric concentration. Although there are 200 CO2 molecules for every CH4 molecule in the atmosphere, the greenhouse effect of CO2 is only 3.5 times that of methane. Consequently, there is justifiable concern over the size of anthropogenic methane emissions and whether they can be reduced. A global network of stations measures the concentration of many atmospheric trace gases and the isotopic composition of some (http://www.esrl.noaa.gov/gmd/dv/). The growth over the last 21 yr in globally averaged atmospheric methane and CO2 concentrations is shown in Fig. 1 (courtesy of Ed Dlugokencky, NOAA, Earth System Research Laboratory). The annual increase in atmospheric methane has slowed recently. This has been attributed to decreasing anthropogenic emissions from fossil fuels, but is thought to be only a temporary pause (Bousquet et al., 2006).
Compiling a global methane budget is a complex task that is inherently uncertain. However, there are increasing amounts of information from satellite imagery and isotopic data to constrain the system. An example budget is shown in Fig. 2 that corresponds most closely to that in the IPCC third assessment report (Lelieveld et al., 1998; Houghton et al., 2001). About 40% of emissions are attributed to natural sources. These are predominantly anaerobic wetlands, with the remainder from termites and the ocean. The major anthropogenic sources are ruminant animals and their waste, release from the mining and use of fossil fuel, anaerobic production from rubbish in landfill, burning of biomass and paddy rice production.
Aerobic methane emission
In 2006, a remarkable observation was published by Keppler et al. (2006). They incubated detached plant material (both fresh and dried) in air and observed an increase in methane concentration. Methane in gas samples was separated by gas chromatography and measured by isotope ratio mass spectrometry. Much greater rates of methane emission were observed with intact plants enclosed in a chamber. The report of aerobic methane production from plant material (Keppler et al., 2006) has proved controversial for several reasons. Firstly, it proposed a global methane source of between 62 and 236 Tg yr−1, representing 10–40% of the accepted annual source. Secondly, although the 13C isotopic signature was consistent with it being derived from methoxyl groups in lignin and pectin, no mechanism has been put forward for the aerobic reaction. Thirdly, the implication that terrestrial vegetation produces methane under aerobic conditions could impact on the benefits of using plants to sequester CO2 from the atmosphere.
A novel source of methane of this magnitude caused a reassessment of methane budgets and fluxes. However, it is crucial that the observations be validated by independent means. This was the objective of the study by Dueck et al. reported in New Phytologist (2007). Given the atmospheric background concentration of around 2000 ppb CH4 and the need to resolve small changes, Dueck et al. approached the problem by growing plants in an atmosphere in which 99% of the CO2 contained 13C. Consequently, any methane emitted from these plants would be isotopically distinct from background methane. They measured 13C-methane with photoacoustic spectroscopy which had a detection limit of 3 ppb. Assuming the average methane emission rate in the light from intact plants measured by Keppler et al. (2006) of 374 ng g−1 h−1, then the methane concentration in air leaving the gas exchange chambers should have increased by 65 ppb (for 10 g of plant with a flow rate of 90 l h−1). However, this was not what was found.
‘Of course, when scaling emission rates to the globe, it is woody vegetation that dominates the calculation.’
Dueck et al. grew six species of plants, three of which had been measured by Keppler et al. (2006). When measured in continuous flow gas exchange chambers, the difference in methane concentration between control chambers and those containing a 13C-labelled plant was below the detection limit. Although the difference was not significant, this corresponded to an emission rate of 21 ng g−1 h−1, which is similar to the rate measured by Keppler et al. (2006) with detached plant material in the light. Keppler et al. (2006) observed an increase in methane emission rates as temperature increased or following exposure to sunlight compared with material kept in darkness. However, in the case of Dueck et al., increasing the temperature and light intensity during the measurement made no difference.
Given the detection limit faced by Dueck et al., a second experiment was performed using all the plants in the isotope labelling growth chamber. After 9 wk of growth in 13CO2, the chambers were flushed with ambient air scrubbed of CO2 and 550 ppm 13CO2 was added once again. The air was then repeatedly sampled at 2-d intervals over the subsequent 6 d to look for any accumulation of 13C-labelled methane. Once again, the methane concentration change was below the detection limit. This corresponded to an emission rate of between –0.9 and 0.4 ng g−1 h−1. These two experiments do not corroborate the formation of methane under aerobic conditions reported by Keppler et al. (2006).
Because of the need to grow 13C-labelled plants within 9 wk, it was not possible for Dueck et al. to use woody species in their study. Keppler et al. (2006) included Picea abies in their intact measurements, which had an emission rate in the light of 81–86 ng g−1 h−1, representing 20% of the average rate for intact plants. Of course, when scaling emission rates to the globe, it is woody vegetation that dominates the calculation. Therefore, further work is needed using woody plants and especially tropical species to resolve this issue.
Spatial and isotopic constraints
The spatial map of atmospheric methane concentration is now updated every 6 d from satellite imagery using an instrument with the acronym SCIAMACHY (Frankenberg et al., 2005). A large aerobic methane emission from tropical forests was considered when synthesizing this satellite data (Bergamaschi et al., 2007). It could be accommodated by reducing other sources and/or increasing sinks.
Another constraint can be applied by considering the 13C composition of methane. Methane escaping from coal and natural gas was found to have a value of –40‰ whereas methane from ruminants and wetlands had a value of –60‰ and methane produced by burning biomass was –24‰ (Quay et al., 1999). Methane collected under aerobic conditions had values between –48 and –60‰ for C3 and –45 to –47‰ for C4 plants, consistent with methoxyl carbon rather than the bulk carbon in biomass, which was around –25 and –15‰ for C3 and C4 plants, respectively (Keppler et al., 2006). These values were combined by Ferretti et al. (2007) with other information such as fire and C3:C4 composition, to derive an estimate of aerobic methane production that was only 20% of that calculated by Keppler et al. (2006).
Value of carbon sequestration
Although Keppler et al. (2006) did not specifically mention it, in the accompanying commentary, Lowe (2006) raised the spectre that new forests planted to sequester carbon under the Kyoto Protocol may actually increase greenhouse warming as a result of their aerobic methane production. This suggestion created a storm in the media and was challenged by Keppler & Rockmann (2007) as well as by others. Kirschbaum et al. (2006) pointed out that aerobic methane would only offset the benefit from sequestering carbon by a few per cent. Both Kirschbaum et al. (2006) and Parsons et al. (2006) also criticized the scaling-up methodology originally used by Keppler et al. (2006), arriving at global emissions between 15 and 30% of the original estimate.
Dueck et al. have been unable to confirm the original finding of aerobic methane emission by Keppler et al. (2006). Using 13C-labelled plant material to unequivocally identify methane from the plants, they failed to measure any significant emission. By contrast, Keppler et al. (2006) measured methane emission into either air containing normal background methane concentrations or methane-free air, and the methane had a characteristic 13C signature. Which result is more credible? Given the failure to verify the original claim, further experimental work is needed. It seems unlikely that others will attempt this, because detecting aerobic methane emission requires specialist equipment and getting a negative result published is difficult. However, this controversial claim has reverberated around the world and, before we adjust the zero offset knob on our instruments, we should recall the prelude to the discovery of the ozone hole (Farman et al., 1985). Satellite data show greater methane concentrations in equatorial regions of South America and Africa than are currently predicted by models (Bergamaschi et al., 2007). While it is possible to change models, it would be best if the changes were based on solid experimental data.