Wetlands and the global carbon cycle: what might the simulated past tell us about the future?


Wetlands play an important role in the global carbon (C) cycle. Flooded or saturated conditions in these ecosystems limit the availability of oxygen to soil microbes and decomposition of organic matter proceeds slowly under the water-logged conditions (Megonigal et al., 2004). As a result of slow rates of anaerobic decomposition, wetlands have accumulated c. 500 Pg of C in their soils globally (approximately one-third of the terrestrial soil C), with the majority of this C stored in peatland soils, which are characterized by > 40 cm of surface organic matter (Bridgham et al., 2006). Anaerobic soil conditions also result in the production of methane (CH4) by methanogenic microbes, and wetlands are responsible for 15% to 40% of current global CH4 emissions (Denman et al., 2007). Given that CH4 is a potent glasshouse gas with 25-times the global warming potential of CO2 (Forster et al., 2007), changes in wetland CH4 emissions can have important implications for the global climate. Indeed, modeling approaches have linked wetland CH4 fluxes to past changes in climate (e.g. Loulergue et al., 2011), and a great deal of current research is exploring how wetland CH4 dynamics will respond to future global changes, including increases in atmospheric [CO2] (summarized in van Groenigen et al., 2011). Recent work by Boardman et al., in this issue of New Phytologist (pp. 898–911), adds an exciting new perspective to this important research area by experimentally demonstrating for the first time that lower atmospheric [CO2] during the Last Glacial Maximum (LGM) could suppress CH4 flux from some wetlands. Their results help provide a potential explanation for the lower atmospheric [CH4] observed during glacial maxima; moreover, these findings have implications for our understanding of current and future wetland CH4 dynamics.

‘… the response of wetland CH4 dynamics to global change is not straightforward …’

In contrast to most terrestrial ecosystems where decomposition of organic matter can be carried out by a single microorganism using oxygen as a terminal electron acceptor for respiration, decomposition in anaerobic wetland environments relies on a number of complementary and competing microbial processes (Megonigal et al., 2004; Fig. 1). As in upland ecosystems, biopolymers (e.g. plant detritus) are initially hydrolyzed by exoenzymes to simpler monomers. In wetlands, however, these monomers are subsequently broken down through a series of sequential fermentation reactions that ultimately generate low molecular weight alcohols, fatty acids (including acetate), H2 and CO2. Methanogens are able to use these fermentation end products – primarily acetate, H2 and CO2– to produce CH4. However, the C sources that fuel methanogens can also be used by a number of competing microbial processes that use a variety of terminal electron acceptors (TEAs), including: nitrate ions (NO3), manganese(III, IV), iron(III), oxidized humic substances, and sulfate ions (SO42−) to complete respiration. These alternative TEAs are more energetically favorable than methanogenesis and can suppress the production of CH4 in many wetland ecosystems. Following methanogenesis, CH4 can leave the system through a combination of ebullition (bubble formation), diffusion across the water–air interface and flux through emergent vegetation. CH4 lost via diffusion is subject to oxidation to CO2 by methanotrophic bacteria using oxygen (or perhaps other TEAs), while CH4 lost via ebullition and through plants largely bypasses this methanotrophic activity. Thus, the controls on CH4 emission from wetlands are complex and frequently mediated by non-methanogenic microbial processes.

Figure 1.

Methane (CH4) flux in wetland environments is controlled by complex interactions between a number of complementary and competing microbial processes. Ultimately carbon flows from complex, plant-derived organic matter to simple carbon molecules used by methanogens to produce CH4 which can be emitted from the wetland to the atmosphere (solid lines). Elevated atmospheric [CO2] can impact CH4 emissions through (A) changes in carbon availability, (B) changes in oxygen availability and/or (C) changes in the pathways of CH4 flux to the atmosphere (dashed lines).

Elevated atmospheric [CO2] can impact wetland CH4 emissions through a number of different pathways that are likely to be mediated by the plant community response to elevated [CO2] (Fig. 1). The production of CH4 is ultimately limited by C availability for methanogens, and this C supply is regulated by plants on a number of different time scales (Fig. 1, pathway A). Increased plant productivity under elevated [CO2] is likely to increase labile C availability in the short term via the exudation of recent photosynthate (as low molecular weight C compounds including acetate) into the volume of soil dominated by roots (i.e. the rhizosphere). Over the longer term, changes in plant community composition, productivity, or belowground allocation can also influence C availability through changes in the quality of soil organic matter (Keller et al., 2004). In addition to supplying labile C, wetland plant roots also supply oxygen to an otherwise anaerobic soil volume and this root oxygen loss (ROL) can influence CH4 flux in multiple ways (Fig. 1, pathway B). Increased ROL can result in a decreased net CH4 flux by stimulating CH4 oxidation or by regenerating more energetically favorable TEAs that suppress CH4 production (Laanbroek, 2010). Finally, elevated [CO2] has the potential to alter the pathway of CH4 loss from wetlands through changes in vegetation community structure (Fig. 1, pathway C). For example, a change in stem density or an increase in cover by nonaerenchymous woody species could change the importance of CH4 flux through plant intercellular air spaces, alter the fraction of CH4 subjected to methanotrophy, and impact net CH4 emissions.

Despite the development of a number of process-based wetland CH4 biogeochemistry models in recent years (e.g. Tang et al., 2010), our ability to incorporate wetland CH4 dynamics into accurate models to predict future CH4 emissions is still limited by gaps in our understanding of the mechanistic controls of wetland CH4 cycling. These challenges are particularly pronounced in peatland ecosystems, including the bogs and fens studied by Boardman et al., which are characterized by dramatic differences in CH4 cycling for reasons that are not well understood. These differences are likely to be mediated by a complex set of mechanistic controls on CH4 production, including: differences in vegetation communities (e.g. Hines et al., 2008); differences in methanogenic communities (e.g. Galand et al., 2010); and differences in the relative importance of competing microbial processes such as sulfate and humic reduction (Vile et al., 2003; Lipson et al., 2010). This lack of mechanistic understanding limits our ability to project how CH4 dynamics in these globally important ecosystems will respond to global changes, including changes in atmospheric [CO2]. For example, while CH4 emissions from fen mesocosms were suppressed by simulated LGM [CO2], CH4 emissions from bog mesocosms were not influenced by this decline in atmospheric [CO2] (Boardman et al.). However, the mechanism for this difference remains speculative. Even within the fen mesocosms, which responded to a simulated LGM [CO2], Boardman et al. were unable to conclusively demonstrate a mechanism for the observed suppression in CH4 emission. They initially hypothesized that decreases in CH4 emissions would be coupled to decreased methanogenic substrate (e.g. acetate) availability as a result of decreased photosynthate allocation to the rhizosphere under LGM [CO2] (Fig. 1, pathway A). However, acetate concentrations in the fen mesocosms were higher in the simulated LGM treatment in both years of the experiment (Boardman et al.), demonstrating that the response of wetland CH4 dynamics to global change is not straightforward and may involve more than simple direct effects on the vegetation community (e.g. White et al., 2008).

While the work of Boardman et al. highlights the challenges of understanding the controls of wetland CH4 dynamics, it also clearly demonstrates that changes in atmospheric [CO2] can impact CH4 emissions. Specifically, CH4 emissions from fen mesocosms were suppressed by 29% in the LGM [CO2] treatment suggesting that current rates of CH4 emission from this fen are higher than they were in the past. Using data provided by Boardman et al. and assuming a linear response of CH4 stimulation over the 200 ppm [CO2] range used in their experiment, the increase in [CO2] from c. 280 to 390 ppm which has occurred since the industrial revolution (Forster et al., 2007) would have resulted in a stimulation of CH4 flux by c. 20% in this fen ecosystem. While admittedly a first approximation, this increase in CH4 flux, which may have already occurred over the last century, is comparable to the 13% increase in CH4 emissions in wetlands in response to simulated future elevated [CO2] reported in a recent meta-analysis (van Groenigen et al., 2011). Interestingly the largest suppression of fen CH4 emissions occurred in warm summer months (Boardman et al.), and this interaction between temperature and elevated [CO2] serves to reiterate the importance of considering interactions between multiple global change factors as controls of ecosystem processes (Norby & Luo, 2004). This study thus serves as an important reminder that anthropogenic activities have likely already impacted ecosystem processes and will continue to do so in the future. Anthropogenic impacts on wetland CH4 emissions, and the potential consequences of these impacts for the global climate, justify the continued exploration of the underlying mechanistic controls of wetland CH4 dynamics and how these controls will respond to multi-factor global change in the future.