[1] Carbonyl sulfide (COS) is an atmospheric trace gas that participates in some key reactions of the carbon cycle and thus holds great promise for studies of carbon cycle processes. Global monitoring networks and atmospheric sampling programs provide concurrent data on COS and CO 2 concentrations in the free troposphere and atmospheric boundary layer over vegetated areas. Here we present a modeling framework for interpreting these data and illustrate what COS measurements might tell us about carbon cycle processes. We implemented mechanistic and empirical descriptions of leaf and soil COS uptake into a global carbon cycle model (SiB 3) to obtain new estimates of the COS land flux. We then introduced these revised boundary conditions to an atmospheric transport model (Parameterized Chemical Transport Model) to simulate the variations in the concentration of COS and CO 2 in the global atmosphere. To balance the threefold increase in the global vegetation sink relative to the previous baseline estimate, we propose a new ocean COS source. Using a simple inversion approach, we optimized the latitudinal distribution of this ocean source and found that it is concentrated in the tropics. The new model is capable of reproducing the seasonal variation in atmospheric concentration at most background atmospheric sites. The model also reproduces the observed large vertical gradients in COS between the boundary layer and free troposphere. Using a simulation experiment, we demonstrate that comparing drawdown of CO 2 with COS could provide additional constraints on differential responses of photosynthesis and respiration to environmental forcing. The separation of these two distinct processes is essential to understand the carbon cycle components for improved prediction of future responses of the terrestrial biosphere to changing environmental conditions.

Summary• Carbonyl sulfide (COS) exchange in C 3 leaves is linked to that of CO 2 , providing a basis for the use of COS as a powerful tracer of gross CO 2 fluxes between plants and the atmosphere, a critical element in understanding the response of the land biosphere to global change.• Here, we carried out controlled leaf-scale gas-exchange measurements of COS and CO 2 in representative C 3 plants under a range of light intensities, relative humidities and temperatures, CO 2 and COS concentrations, and following abscisic acid treatments.• No 'respiration-like' emission of COS or detectable compensation point, and no cross-inhibition effects between COS and CO 2 were observed. The mean ratio of COS to CO 2 assimilation flux rates, A s ⁄ A c , was c. 1.4 pmol lmol )1 and the leaf relative uptake (assimilation normalized to ambient concentrations, (A s ⁄ A c ) (C a c ⁄ C a s )) was 1.6-1.7 across species and conditions, with significant deviations under certain conditions. Stomatal conductance was enhanced by increasing COS, which was possibly mediated by hydrogen sulfide (H 2 S) produced from COS hydrolysis, and a correlation was observed between A s and leaf discrimination against C 18 OO.• The results provide systematic and quantitative information necessary for the use of COS in photosynthesis and carbon-cycle research on the physiological to global scales.

Carbonyl sulfide (COS) and C 18 OO exchange by leaves provide potentially powerful tracers of biosphere-atmosphere CO 2 exchange, and both are assumed to depend on carbonic anhydrase (CA) activity and conductance along the diffusive pathway in leaves. We investigated these links using C 3 and C 4 plants, hypothesizing that the rates of COS and C 18OO exchange by leaves respond in parallel to environmental and biological drivers. Using CA-deficient antisense lines of C 4 and C 3 plants, COS uptake was essentially eliminated and discrimination against C The seasonal cycling of CO 2 concentration measured in the background atmosphere is often taken as evidence of the breathing of Earth and corresponds to about 10 Pg carbon. This is less than 10% of what we estimate to be the total cycling of CO 2 due to respiration and photosynthesis of the terrestrial biosphere (Beer et al., 2010). The discrepancy stems from the fact that most of the cycling CO 2 mixes in the atmosphere near the surface, and we only see the net sum of respiration and photosynthesis, which are often more or less balanced over the seasonal cycle. Modeling and measurements of CO 2 exchange by ecosystems over diurnal cycles have been used to deconvolve the primary biological processes (e.g. Desai et al., 2008). However, our confidence in these approaches disintegrates as we move to larger scales because of the increase in the internal mixing of CO 2 fluxes. Other approaches are needed to quantify the basic physiological processes at these scales.