A kinetic analysis of leaf uptake of COS and its relation to transpiration, photosynthesis and carbon isotope fractionation

. Carbonyl sulﬁde (COS) is an atmospheric trace gas that holds great promise for studies of terrestrial carbon and water exchange. In leaves, COS follows the same pathway as CO 2 during photosynthesis. Both gases are taken up in enzyme reactions, making COS and CO 2 uptake closely coupled at the leaf scale. The biological background of leaf COS uptake is a hydrolysis reaction catalyzed by the enzyme carbonic anhydrase. Based on this, we derive and test a simple kinetic model of leaf COS uptake, and relate COS to CO 2 and water ﬂuxes at the leaf scale.


Introduction
Carbonyl sulfide (COS) is a source of stratospheric sulfate aerosols and plays an important role in stratospheric ozone chemistry (Crutzen, 1976;Andreae and Crutzen, 1997).Globally, the main source of COS is the ocean, and uptake by leaves and soil are the main sinks (Kettle et al., 2002;Kesselmeier et al., 1999;van Diest et al., 2008).The seasonality of COS in the northern extratropical atmosphere is dominated by COS uptake by terrestrial vegetation (Montzka et al., 2007;Suntharalingam et al., 2008).This is because COS is taken up in enzyme reactions in leaves, similar to CO 2 .But in contrast to CO 2 , there is no concurrent COS release from terrestrial ecosystems.Monitoring atmospheric COS concentration can thus provide valuable information on terrestrial gross carbon fluxes that cannot be obtained by measuring CO 2 alone.
Based on the close coupling of leaf COS and CO 2 uptake, the global COS sink by vegetation has been often derived from estimates of CO 2 fluxes (Chin and Davis, 1993;Kesselmeier et al., 1993;Andreae and Crutzen, 1997;Watts, 2000;Kettle et al., 2002;Sandoval-Soto et al., 2005).Recently, it has been proposed to invert this approach, with the aim to obtain estimates of global terrestrial gross CO 2 uptake from atmospheric COS measurements (Montzka et al., 2007;Campbell et al., 2008;Suntharalingam et al., 2008).COS fluxes have been converted into CO 2 fluxes, or vice versa, using the ratio of COS to CO 2 deposition velocities (uptake rates divided by ambient concentration) observed during chamber experiments (Sandoval-Soto et al., 2005).The ratio of deposition velocities is identical to the normalized uptake ratio (see Sect. 5).Most ratios are >1, with typical values between 2 and 3.This has been interpreted as a preferential U. Seibt et al.: Analysis of leaf COS uptake reaction for COS over CO 2 by carbonic anhydrase (CA), the enzyme responsible for the irreversible reaction of COS in leaves (Kesselmeier and Merk, 1993;Protoschill-Krebs et al., 1996).
Here, we introduce a process-oriented description of COS uptake at the leaf level, analogous to that for leaf CO 2 and water fluxes.To test the equation, we compare predicted COS uptake to data obtained in field and laboratory chambers (Kuhn et al., 1999;Sandoval-Soto et al., 2005).The new description directly relates COS uptake at the leaf level to stomatal conductance.As a consequence, the normalized uptake ratio for leaf COS and CO 2 fluxes can provide an estimate of C i /C a , the ratio of intercellular to atmospheric CO 2 , an important plant gas exchange parameter that cannot be measured directly.

Expressing leaf exchange of COS in analogy to CO and water vapour
Based on Fick's law of diffusion, leaf fluxes can be described as the product of conductance and concentration gradient (Penman and Schofield, 1951;Cowan, 1977).For example, the transpiration flux, F w (mmolm −2 s −1 ), is calculated from: where w a and w i (mmolmol −1 ) are the vapour mole fractions of ambient air and intercellular spaces, respectively, and g w (molm −2 s −1 ) is the leaf conductance to water vapour diffusion.For ecosystem studies, g w is composed of leaf boundary layer (g bw ) and stomatal conductance (g sw ): g w = (1/g sw + 1/g bw ) −1 .In chamber studies, fans often provide enough ventilation to render the boundary layer component negligible, and g w ≈ g sw .Leaf conductance is usually obtained from measured values of transpiration (F w ), vapour mole fraction at measured air temperature and relative humidity (w a ), and assuming saturated air at measured leaf temperature (w i ).
Similarly, photosynthetic CO 2 uptake, F c (µmol m −2 s −1 ), can be calculated from: where C a,c and C i,c (µmolmol −1 ) are the mole fractions of CO 2 in ambient air and intercellular spaces, respectively, and g sc (molm −2 s −1 ) is the stomatal conductance to CO 2 (again, for well ventilated chamber measurements).In contrast to water vapour, it is impossible to directly estimate C i,c , and hence to obtain g sc from measurements of F c and C a,c .Thus, g sc is usually derived from g sw based on the known relationship between CO 2 and water vapour conductances: g sc = g sw /R w−c , where R w−c ≈ 1.6 (Lide, 2008) is the ratio of water vapour and CO 2 diffusivities.
In analogy to the above, the equation for leaf COS uptake, F COS (pmolm −2 s −1 ), can be written as: where C aCOS and C mCOS (pmolmol −1 ) are the mole fractions of COS in ambient air and at the reaction sites with the enzyme carbonic anhydrase (CA), respectively.The above combines two diffusion steps: from C aCOS to C iCOS in the intercellular spaces (equivalent to C i,c in Eq. 2), and from C iCOS to C mCOS within the mesophyll cells.In other words, we consider a different diffusion endpoint for COS (C a to C m ) than for CO 2 (C a to C i ).Accordingly, the leaf conductance to COS (g tCOS ) includes both stomatal conductance (g sCOS ) and internal conductance (g iCOS : to account for the transfer of COS into the mesophyll cells (and the biochemical reaction rates): The spatial distribution of CA in leaf mesophyll cells is not known.We expect a location "upstream" of Rubisco, the enzyme responsible for the reaction with CO 2 in the chloroplasts.This is because CA enhances the solution of CO 2 and thus its supply to Rubisco.Hence, we hypothesize that CA is located directly adjacent to the intercellular spaces, so that g iCOS should be much larger than g sCOS .Carbonic anhydrase is a very efficient catalyst for the reaction COS+H 2 O→H 2 S+CO 2 (Protoschill-Krebs et al., 1996;Notni et al., 2007).COS release has not been observed even at low ambient COS (Sandoval-Soto et al., 2005).Thus, we assume that C mCOS C aCOS , so that Eq. (3) can be written to a good approximation as: The value of g iCOS is not known, but we can obtain a reasonable approximation for g tCOS by assuming a constant internal conductance in g tCOS = (1/g sCOS + 1/g iCOS ) −1 .Alternatively, for a fixed stomatal conductance or mean values such as large scale estimates, we can also use a constant ratio of stomatal to internal conductance in g tCOS = g sCOS (1 + g sCOS /g iCOS ) −1 .Similar to CO 2 fluxes, COS fluxes can now be derived from C aCOS and g sw based on the relationship between COS and water vapour conductances: g sCOS = g sw /R w−COS .The ratio R w−COS has not been determined experimentally yet.In the following, we provide theoretical and empirically derived estimates of R w−COS .

Ratios of diffusivities of COS, CO 2 and water vapour
The relationships between stomatal conductances for two different gases correspond to their diffusivity ratios: R w−c = D a,w /D a,c , R w−COS =D a,w /D aCOS , and R c−COS =D a,c /D aCOS , where D a,w , D a,c , and D aCOS are the diffusion coefficients of water vapour, CO 2 and COS in air.
Table 1.Theoretical values of diffusivities D a,g (cm 2 s −1 ) of water vapour, CO 2 and COS in air (parameters taken from Bird et al., 2007).Empirical estimates of D a,w and D a,c from Massman (1998).All estimates were derived for a temperature of 23C, the average air temperature during the chamber experiments.For plants at typical field conditions, the effect of temperature on the ratios of diffusivities is probably small.The diffusion coefficient D a,g (cm 2 s −1 ) of a gas (subscript g) in ambient air (subscript a) can be calculated from Chapman-Enskog theory (Bird et al., 2007): where R (8.31 JK −1 mol −1 ) is the gas constant, T (K) is temperature, M (gmol −1 ) is the molar mass of the gas, N A (6.022 10 23 mol −1 ) is the Avogadro constant, p (atm) is ambient pressure, σ a,g = (σ a +σ g )/2 ( Å) is the collision diameter of the molecules, and a,g is the dimensionless collision integral for diffusion.This approach was chosen because it provides molecular parameters for all three gases (Bird et al., 2007).
Theoretical values of diffusion coefficients for water vapour, CO 2 and COS in air are summarized in Table 1.The theoretical diffusivity ratio for CO 2 and water is 1.66, somewhat higher than the empirical estimate of 1.58 (Massman, 1998).Based on the difference between theoretical and empirical estimate, as well as the analysis of uncertainties presented in Massman (1998), theoretically predicted diffusivity ratios are probably within 10% of the empirical values.Thus, the resulting diffusivity ratios for COS are R w−COS =2.0±0.2 with respect to water vapour, and R c−COS =1.2±0.1 with respect to CO 2 .

Testing the leaf COS equation with chamber observations
To evaluate the relationships developed in the previous section, we calculate leaf COS uptake using Eq. ( 4) with the theoretical estimate of R w−COS , and compare the predicted to observed COS fluxes.We use concurrent data on transpiration, COS and CO 2 fluxes from leaf-scale chamber measurements (Kuhn et al., 1999;Sandoval-Soto et al., 2005).
Briefly, experiments were carried out with an automated system consisting of two dynamic (flow-through) chambers, one with enclosed tree branch, and an empty reference chamber.COS mole fractions were quantified by consecutive sampling in both chambers, and COS uptake determined from their differences (flux= conc•flowrate/leaf area).Transpiration and net CO 2 fluxes were determined similarly from differential measurements of sample vs reference chamber air.
We then obtained g sw values from transpiration rates (Eq.1).
In this context, we note that the appropriate values for use in Eqs. ( 1), ( 2) and ( 4) are w a , C a,c and C aCOS in the air surrounding the observed leaves or plants, i.e. in the sample chamber.Often, they first need to be calculated from reference chamber values and flux rates.
One data set was obtained on branches of Fagus sylvatica (European beech) in laboratory chambers (Sandoval-Soto et al., 2005).COS mole fractions were quantified using an automated setup (von Hobe et al., 2000) with an analytical precision of about 8 pmol mol −1 , plus any uncertainties in the chamber system (Sandoval-Soto et al., 2005).The average COS mole fractions in the reference and sample chamber were 450 ppt and 340 ppt, respectively.Rates of leaf COS uptake were predicted from Eq. (4) using g sw and C aCOS data, and based on R w−COS =2.The predicted fluxes are in good agreement with the observed COS uptake rates over three days of measurements (Fig. 1).As the enzyme reaction of CA with COS is light independent (Protoschill-Krebs et al., 1996), COS uptake can continue in the dark as long as stomata remain open.However, the uncertainties in observed COS fluxes were often larger than the fluxes in the dark (i.e.no significant differences between reference and sample mole fractions).
The second data set was obtained on branches of Quercus agrifolia (Coast Live oak) in field chambers (Kuhn et al., 1999).COS mole fractions were quantified using cryogenic trapping and a gas chromatograph system, with an overall precision of about 25 pmol mol −1 (Kuhn et al., 1999).The predicted COS fluxes are reasonable compared to the measured COS uptake, and mostly within the limits of uncertainties of the measurements (Fig. 2).We only show the results for times where the observed COS fluxes were larger than their uncertainties.
Using Eqs. ( 1) and (4), and observed chamber mole fractions and flux rates, we also obtained first empirical estimates of R w−COS , the ratio of conductances of COS and water vapour.We assumed a constant g iCOS of 0.  4), assuming R w−COS =2 and constant internal conductance, g iCOS =0.2), and observed uptake rates from measurements on Fagus sylvatica in laboratory chambers (original data from Sandoval-Soto et al., 2005).For comparison, we also show predicted rates where internal conductance is assumed to be a negligible limitation to COS diffusion (g iCOS =20).Shaded areas indicate measurements in the dark.
relating it to the mean g sCOS for day-time measurements by a factor of ten (g sCOS /g iCOS ≈0.1).As a test, we also used g iCOS of 20 molm −2 s −1 (g sCOS /g iCOS ≈0.001), i.e. effectively neglecting the internal part of the diffusion pathway.
For the re-analysed laboratory data on Fagus sylvatica, the mean ratio was R w−COS =2.0±0.3 (or 2.2 for g iCOS =20).For the re-analysed field data on Quercus agrifolia, the mean ratio was R w−COS =2.2±0.8 (or 2.4 for g iCOS =20).Both values are close to the theoretical estimate of 2, but the uncertainties are still large due to the high uncertainties in the measurements of water vapour and particularly COS fluxes.
5 The normalized uptake ratio of COS to CO 2 as a proxy for C i /C a COS fluxes can be directly compared to those of CO 2 by expressing them as uptake rates normalized by the ambient mole fraction of each gas, v COS and v CO 2 (molm −2 s −1 ).The normalized uptake ratio (v COS /v CO 2 ) is identical to the ratio of deposition velocities, but instead of using ambient concentrations as in deposition velocities (e.g.(Sandoval-Soto et al., 2005)), normalized uptake rates are defined relative to ambient mole fractions: The ratio v COS /v CO 2 was also used in a recent analysis of atmospheric COS data (Campbell et al., 2008) to calculate global COS uptake by vegetation from modelled leaf CO 2 fluxes: , where F c corresponds to gross primary production (GPP, see Eq. ( 1) in Campbell et al., 2008).4), assuming R w−COS =2 and constant internal conductance, g iCOS =0.2), and observed COS uptake from measurements on Quercus agrifolia in field chambers (original data from Kuhn et al., 1999).For comparison, we also show predicted rates where internal conductance is assumed to be a negligible limitation to COS diffusion (g iCOS =20).
Using Eqs. ( 2) and ( 4), we can rewrite the above definitions as: Thus, the normalized uptake ratio of COS to CO 2 , v COS /v CO 2 , can be written as: using g COS = (1/g sCOS + 1/g iCOS ) −1 , and R c−COS = g sc /g sCOS for the ratio of stomatal conductances.We write C i /C a (instead of C i,c /C a,c ) from hereon for simplicity.The C i /C a ratio is an important plant parameter that cannot be measured directly.It is an expression for the balance between the CO 2 supply (limited by stomatal conductance) and demand (limited by light and the efficiency of enzyme reactions).Based on Eq. ( 8) and the theoretical estimate of the stomatal conductance ratio (R c−COS =1.2±0.1), it should be possible to determine C i /C a from observed fluxes and mole fractions of COS and CO 2 : For the estimate of internal COS conductance (g sCOS /g iCOS =0.1) that is consistent with the CO 2 and water data from the laboratory experiments, Eq. ( 9) can be simplified to: The relationship between v COS /v CO 2 and C i /C a ratios is illustrated in Fig. 3 9) and R c−COS =1.2.We assume that leaf internal conductance is much higher than stomatal conductance (g sCOS /g iCOS =0.1 or 0.2).Thus, neglecting this component (i.e.assigning a value of 0.001) does not make a large difference compared to including it.It is also possible that g iCOS is a more limiting part of the diffusional pathway (0.5), with a narrower range of v COS /v CO 2 corresponding to typical C i /C a ratios.
3 correspond to C i /C a from 0.63 to 0.75, typical mean C i /C a ratios of C 3 plants under field conditions.In contrast, v COS /v CO 2 ratios below 1 or above 6 may indicate that processes other than photosynthetic uptake are contributing to the observed COS and/or CO 2 fluxes (for example, 7.8±1.3for ecosystem scale uptake, recalculated from Xu et al., 2002).Neglecting internal resistance (g sCOS /g iCOS =0.001) has little effect on the relationship between v COS /v CO 2 and C i /C a .On the other hand, if g iCOS is a much stronger component of the diffusion pathway (g sCOS /g iCOS =0.5), we would expect slightly lower v COS /v CO 2 of 1.5 to 2 for typical C i /C a values.Ratios of v COS /v CO 2 and corresponding C i /C a for published data are listed in Table 2. Several values are above the typical C i /C a range, probably because relative humidity and stomatal conductances are often higher in cuvettes compared to ambient conditions.Most v COS /v CO 2 ratios were calculated from COS and CO 2 mole fractions in atmospheric air, but the conditions relevant for gas exchange calculations are those of the chamber air surrounding the leaves.Recalculating with chamber air mole fractions usually leads to higher v COS /v CO 2 .For example, for the laboratory data on Fagus sylvatica presented in Fig. 1 (Sandoval-Soto et al., 2005), we obtained a mean v COS /v CO 2 of 2.4 based on air in the sample cuvette, substantially higher than the v COS /v CO 2 of 2.0 calculated from reference air.In addition, air supplied to chambers can have a very different composition than typical atmospheric air, particularly during laboratory experiments.
We assume that the enzyme CA is very efficient at converting the COS reaching the sites of reaction (Protoschill-Krebs et al., 1996), so that leaf COS uptake is primarily limited by diffusion.Biochemical studies are needed to resolve the extent to which CA activity may limit the overall rate.We implicitly include the CA reaction with the physical diffusion limitation in g iCOS .As a consequence, even if there is an enzymatic preference of CA for COS over CO 2 , it is not expressed in v COS /v CO 2 values.Instead, the higher normalized uptake of COS compared to CO 2 results from the higher reaction efficiency of CA, relevant for COS, compared to Rubisco, relevant for CO 2 , associated with back diffusion of non-assimilated CO 2 (see Montzka et al., 2007).This is expressed in the negligible internal concentration of COS compared to CO 2 (Eqs. 2 and 4), and the coupling of v COS /v CO 2 to C i /C a (Eq.9).
Also listed in Table 2 are the v COS /v CO 2 values for a range of ecosystem types used to estimate global COS uptake by vegetation (Campbell et al., 2008).The global mean v COS /v CO 2 is 2.2 which corresponds to a C i /C a of 0.66, a typical ratio for C 3 vegetation.Most ecosystem specific values are close to the global mean, except for one lower value of 0.44 for C i /C a of boreal forests.The relationships for v COS /v CO 2 developed here should also hold for C 4 plants, where CA is distributed throughout the mesophyll cells (Burnell and Hatch, 1988), but there is not enough data available for evaluation of Eq. ( 4).
Because of the direct coupling to stomatal conductance, measurements of COS fluxes could provide a valuable alternative to determine C i /C a at high humidity, when water fluxes may be too small to be reliably determined.As an illustration, we calculated C i /C a for Fagus sylvatica from stomatal conductance to COS (g sCOS ) and water vapour (g sw ).Over three days, we found C i /C a =0.67±0.07 from g sCOS , and C i /C a =0.69±0.01 from g sw (original data from Sandoval-Soto et al., 2005).The mean values are in good agreement, but the COS based estimate has larger uncertainties than the g sw based estimate.

Estimating the normalized uptake ratio of COS to CO 2 from carbon isotopes
The relationships described above (Eq.8) can also be used to obtain v COS /v CO 2 ratios from independent estimates of C i /C a .One possibility is to utilize the coupling of C i /C a and photosynthetic 13 C discrimination to determine v COS /v CO 2 ratios.Indeed, the use of COS to investigate CO 2 gas exchange is quite similar to the more traditional use of isotopic tracers in this context.In its simplest form, 13 C discrimination ( , ‰) during photosynthesis can be written as (Farquhar et al., 1982;Farquhar and Richards, 1984) www.biogeosciences.net/7/333/2010/Biogeosciences, 7, 333-341, 2010 Table 2. Normalized uptake ratios of COS to CO 2 (v COS /v CO 2 ) and their corresponding C i /C a ratios (based on Eq. ( 9), g sCOS /g iCOS =0.1) for published data (Sandoval-Soto et al., 2005), and ecosystem types used to estimate global COS uptake by vegetation (Campbell et al., 2008).Note that for published data from enclosure studies (except those recalculated from original data, indicated by *), v COS /v CO 2 was calculated from reference (or atmospheric) COS and CO 2 mole fractions.As the mole fractions of the air surrounding the leaves (i.e.chamber air) should be used instead, their actual v COS /v CO 2 values could be 10 to 15% higher than listed in the where a is the fractionation during CO 2 diffusion through the stomata (4.4‰, Craig, 1953), and b is the weighted fractionation during internal transfer of CO 2 and fixation by Rubisco and PEPc.Using Eq. (10) to substitute for C i /C a in Eq. ( 8) yields: This relationship has the advantage that measurements of the δ 13 C of leaf samples, ecosystem exchange, or the results of atmospheric inversion studies, can be used to derive v COS /v CO 2 ratios independent of CO 2 flux measurements.Here, we apply this approach to obtain v COS /v CO 2 for a global range of biomes (data from Table 2 in Lloyd and Farquhar (1994)).To be consistent with the original calculations, we derive C i /C a from their equation (Eq. 4 of Lloyd and Farquhar (1994), b=27.5, and including photorespiration) instead of Eq. ( 10).Most of the resulting v COS /v CO 2 ratios (Eq.8) fall between 2 and 3 (Table 3, for the best guess g sCOS /g iCOS =0.1).The lowest ratios are derived for dry biomes such as semi-desert (1.7), whereas tropical forests tend to have the highest ratios, >3 (Table 3), reflecting the increase in stomatal conductance with decreasing evaporative demand.
We calculate a GPP weighted global mean v COS /v CO 2 ratio of 2.8±0.3 (for C 3 plants, Eqs. ( 8) and (11), possible variations in g sCOS /g iCOS are included in the uncertainty estimate), larger or in the upper range of previous estimates (Sandoval-Soto et al., 2005;Montzka et al., 2007).With a global GPP estimate of 109.3Pg C yr −1 for 2001-2003 (Zhao et al., 2005), we estimate a global vegetation sink of COS in the order of 900±100 Gg S yr −1 .This is in the lower range of 730-1500 Gg S yr −1 (Sandoval-Soto et al., 2005) and 1200±300 Gg S yr −1 (Xu et al., 2002), but much higher than the estimate of 490 Gg S yr −1 (Suntharalingam et al., 2008).

Conclusions
We have developed a simple model of leaf COS uptake, analogous to the equations for leaf CO 2 and water fluxes.We describe these equations as process-oriented (not processbased) because they all require estimates of stomatal conductance, for which we do not yet have a fully mechanistic understanding.Leaf COS uptake predicted from the new equation was in good agreement with data from field and laboratory chambers (Kuhn et al., 1999;Sandoval-Soto et al., 2005), although with large uncertainties.(Lloyd and Farquhar, 1994), assuming that internal conductance is a negligible (g sCOS /g iCOS =0.001), intermediate (0.1, 0.2), or large (0.5) limitation in the diffusional pathway of COS.As a consequence of the close coupling of leaf COS and CO 2 uptake, the normalized uptake ratio of COS and CO 2 can be used to provide estimates of C i /C a , the ratio of intercellular to atmospheric CO 2 , an important plant gas exchange parameter that cannot be measured directly.In addition, COS and 13 C discrimination can be combined to obtain independent estimates of photosynthesis (GPP).The new process-oriented description provides a framework for understanding COS fluxes that should improve the usefulness of atmospheric COS and CO 2 measurements to obtain estimates of gross photosynthesis at regional to global scales.

Biome
To develop COS into a reliable proxy for C i /C a and GPP, we now need to reduce the experimental uncertainties, characterize concurrent terrestrial fluxes such as soil COS uptake, and evaluate the assumptions and parameters of the new COS model (Eq.4) under a wide range of environmental conditions.With better analytical precision, COS could also become a valuable addition to measurements at the ecosystem scale, particularly when water measurements are difficult or impossible due to high humidity, for example in tropical ecosystems.Thus, COS has the potential to provide new constraints on stomatal and canopy conductance, a fundamental -and difficult to measure -plant process that regulates the exchange of carbon, water and energy of the terrestrial biosphere.
U. Seibt et al.: Analysis of leaf COS uptake Notni et al., 2007), and rewrite Eq. ( 2) as: = g iCOS C iCOS Using Eqs.(A2) and (A4), the normalized uptake rates of CO 2 and COS are: and the normalized uptake ratio of COS to CO 2 can be written as: where the stomatal conductances are related by a constant factor (R c−COS = g sc /g sCOS ≈1.2).As an example, using an internal conductance of 0.2 mol m −2 s −1 for both CO 2 and COS, we would obtain for an average CO 2 stomatal conductance of g sc =0.1 mol m −2 s −1 from Eq. (A7): Accounting for internal conductance, 13 C discrimination during photosynthesis can be written as (Farquhar et al., 1982):  (Lloyd and Farquhar, 1994), this yields: and can be combined with Eq. ( 7) to: Additional processes such as boundary layer conductance and photorespiration can be included similarly in the above equations.

Fig. 1 .
Fig. 1.Leaf COS uptake predicted from transpiration data (Eq.(4), assuming R w−COS =2 and constant internal conductance, g iCOS =0.2), and observed uptake rates from measurements on Fagus sylvatica in laboratory chambers (original data fromSandoval- Soto et al., 2005).For comparison, we also show predicted rates where internal conductance is assumed to be a negligible limitation to COS diffusion (g iCOS =20).Shaded areas indicate measurements in the dark.

Fig. 2 .
Fig. 2.Leaf COS uptake predicted from transpiration data (Eq.(4), assuming R w−COS =2 and constant internal conductance, g iCOS =0.2), and observed COS uptake from measurements on Quercus agrifolia in field chambers (original data fromKuhn et al., 1999).For comparison, we also show predicted rates where internal conductance is assumed to be a negligible limitation to COS diffusion (g iCOS =20).
a m are the fractionations during stomatal diffusion and internal transfer of CO 2 , and b is the weighted fractionation during fixation by Rubisco and PEPc.With the approximation C i,c − C c,c = 0.1C a,c . Values of v COS /v CO 2 between 2 and Ratios of intercellular to ambient CO 2 mole fraction, C i /C a , can be obtained from the normalized uptake ratios of COS to CO 2 , v COS /v CO 2 , based on Eq. ( table.

Table 3 .
Normalized uptake ratios of COS to CO 2 (v COS /v CO 2 ) and C i /C a obtained from estimates of 13 C discrimination ( ) during photosynthesis