Carbonyl sulfide: a new tool for understanding the response of the land biosphere to climate change


The global climate is changing mostly because of the anthropogenic emission of greenhouse gases, especially CO2, into the atmosphere. The rate of atmospheric CO2 growth in the future will depend on the balance between emissions and sink strengths. The importance of land photosynthesis in dampening the CO2 increase in the atmosphere is now well recognized and both climate and global carbon (C) models incorporate photosynthesis–climate feedbacks. However, we still lack robust tools for partitioning different component fluxes (e.g. photosynthesis, respiration, decomposition) and for assessing the effects of climate change and atmospheric CO2 increase on the C-sequestration potential of the biosphere. The paper by Stimler and coworkers in this issue of New Phytologist (pp. 869–878) shows that carbonyl sulfide (COS) represents a useful tracer of gross photosynthesis. This offers the perspective of an additional independent tool to study the terrestrial C cycle and to investigate ecosystem responses to global change.

‘A peculiarity in comparing the two fluxes … COS uptake in leaves is unidirectional and no COS re-emission has been detected, whereas the CO2 flux is bidirectional …’

About three-quarters of the anthropogenic emission of CO2 is caused by the burning of fossil fuel, with a small contribution from cement production, while the remaining emissions are attributable to deforestation and land-use change. At present, the rate of increase of [CO2] in the atmosphere is about half that of CO2 emissions, because some of the CO2 emitted is dissolved in the oceans, some is taken up by land vegetation and a certain amount of ‘missing C’ may be ascribed to an underestimation of belowground allocation (Burgermeister, 2007). The sink capacity of the land biosphere is determined by the balance between photosynthetic CO2 fixation and release of CO2 by respiration and decomposition (plus land-use change), with a great sink potential represented by the soil. Under present conditions, the terrestrial biosphere is acting as a C sink, because photosynthesis exceeds the release of CO2 by respiration and decomposition, although massive amounts of C are being released as a result of deforestation and other land-use change. Provided that the reduction of emissions is imperative in the future, terrestrial vegetation may be regarded as a major C sink and might help in mitigating the increase in atmospheric [CO2]. However, it is uncertain whether land vegetation will act as a sink in the future, because many environmental and anthropogenic factors influence the uptake or release of CO2 by the ocean and land biosphere. These factors include naturally induced and human-induced climate changes, fertilizing effects by CO2 and nitrogen depositions, changes in ocean currents, fires and human activities. Separation of individual fluxes is therefore of primary importance for understanding how the C cycle is going to be influenced by human activities and to model future greenhouse gas concentrations and future global climate.

Land and ocean uptake can be separated by different atmospheric measurements of concentrations (CO2 and O2) and isotopic abundances (13CO2 and CO18O). In particular, measurements of O2 and CO2 concentrations allow such partitioning because terrestrial photosynthetic CO2 uptake involves the release of O2 in a fixed stochiometry, whereas dissolution of CO2 in the ocean (nonbiological ocean uptake) has no effect on atmospheric [O2]. Stable isotopes of C and O are also used for partitioning CO2 sinks. The use of δ13C for this purpose is based on the relatively large fractionation occurring during photosynthesis in C3 land plants (Farquhar et al., 1982; Brugnoli & Farquhar, 2000). This leads to organic C being significantly depleted in 13C compared with atmospheric CO2 which, consequently, is enriched in 13C. By contrast, oceanic uptake involves relatively small fractionation effects. Therefore, changes in the δ13C of atmospheric CO2 is indicative of the extent to which variations are dominated by biospheric or oceanic uptake. Process-based models of oceanic and terrestrial C cycling have been developed, compared and tested against in situ measurements.

Measurements of δ13C allow the net CO2 flux to be separated into photosynthetic and respiratory component fluxes, and autotrophic and heterotrophic respiration to be distinguished at different scales of complexities, from the ecosystem to the global scale (Ciais et al., 1995, 1997; Fung et al., 1997).

At the ecosystem level, measurements of the net ecosystem exchange (NEE) are obtained routinely by micrometeorological measurements, such as the eddy covariance approach (Baldocchi et al., 1988) using fast-response sonic anemometers and CO2 sensors installed on towers above the ecosystems, to measure the exchange of CO2 and water vapour between a land patch and the atmosphere. Many measurement sites are now active as part of the FLUXNET network ( However, eddy covariance only provides NEE without information about the component fluxes. The combination of isotopic and eddy covariance measurements (e.g. Bowling et al., 2001) allows partitioning of the NEE into photosynthetic and respiratory component fluxes, provided that reliable assessment of canopy photosynthetic discrimination, the 13C isoflux (i.e. an approximation of the vertical isotopic flux) and the isotopic ratio of ecosystem respiration are obtained. Assessment of these parameters requires isotopic measurements of atmospheric CO2, analysis of δ13C of plant components and the estimate of δ13C of respired CO2 by Keeling plots (Keeling, 1958).

While at an early stage, measurements of atmospheric CO2 were limited to 13C, more recently the 18O : 16O ratio has started to be routinely measured. The 18O composition of atmospheric CO2 reflects the equilibration of CO2 and water in leaves, a process strongly dependent on the catalysis by the enzyme carbonic anhydrase (CA). After dissolution, CO2 exchanges virtually all oxygen atoms with water and, because water in leaves is highly enriched in 18O as a result of evaporative fractionation (Barbour, 2007), the CO2 released into the atmosphere has contrasting 18O signatures depending on the source (leaves, trunk, root and soil respiration). Hence, combining δ13C and δ18O analyses and concentration measurements allows the NEE to be separated into photosynthetic and respiratory fluxes and to identify the autotrophic and heterotrophic components (Yakir & Wang, 1996; Bowling et al., 2001; Yakir, 2003; Scartazza et al., 2004; Knohl & Buchmann, 2005).

The work by Stimler et al. introduces a new tracer, COS, that can be used as an additional tool to assess gross CO2 uptake by land vegetation. Carbonyl sulfide is the most abundant reduced sulfur trace gas in the troposphere, it has a long lifetime and it can reach the stratosphere where it is involved in oxidation reactions forming stratospheric aerosols. Total global COS sources and sinks are approximately balanced.

The role of vegetation as a major sink for tropospheric COS has been studied for > 20 yr and it is now well recognized, but, until recently, uncertainties of COS measurements and in the estimate of the sink strength were still quite large. A close relationship between COS uptake and photosynthetic CO2 uptake has been previously demonstrated (Sandoval-Soto et al., 2005). The deposition velocity for COS has been reported to be somewhat higher than that for CO2 with the ratio ranging from 1 to 10 (Sandoval-Soto et al., 2005; Stimler et al.). A peculiarity in comparing the two fluxes under normal atmospheric gas compositions is that COS uptake in leaves is unidirectional and no COS re-emission has been detected, whereas the CO2 flux is bidirectional (i.e. the net uptake measured is the result of gross uptake and respiratory release). This is explained by the fact that in leaves COS undergoes hydration catalysed by CA, which leads to the virtually irreversible formation of hydrogen sulfide (H2S).

The work by Stimler et al. shows that COS exchange in leaves is linked to gross CO2 uptake and it is under stomatal control; stomata seem to respond to COS, probably via an H2S effect. In addition, Stimler et al. demonstrate that there is no respiration-like COS emission and there is no apparent cross-interaction or toxicity between CO2 and COS, although the hydration of both is mediated by CA. The common pathway through CA is also the reason for the reported relationship between COS uptake and isotopic discrimination against 18O in CO2. The observed stomatal response to COS and CA specific activity on COS need to be investigated further.

Taken together, the results reported by Stimler et al. are very relevant because they provide a new tracer for photosynthetic primary productivity in the study of the response of terrestrial ecosystems to climate change. Carbonyl sulfide uptake can also be used to study the physiological ecology of plants, to investigate relevant photosynthetic features, such as mesophyll conductance, CA and 18O discrimination. The availability of this new tracer, in addition to 13C and 18O discrimination, can help increase the accuracy of different flux estimates obtained by each individual tracer, improving global C models and help in the study of terrestrial ecosystem responses to global change.