Response to the comments on “ Stomatal control of leaf fluxes of carbonyl sulfide and CO 2 in a Typha freshwater marsh ”

P1L3–4 I think, here, you could bemore specificwith respect to what we have learned so far: ‘LRU is known to increase under low light’. Revised. See P1L4. P1L15–17 reduce the emphasis on the role of stomatal control. Since stomatal conductance data have been added to the revised manuscript, the emphasis on the role of stomatal control is appropriate. Introduction This section is interesting and very clearly written. P2L2–10 maybe consider shortening this section, these concepts have already been amply discussed in the literature. We have shortened this paragraph by 25%. See P2L2–L8. P2L2 ‘COS has been shown to be a unique tracer’. Changed to “Carbonyl sulfide (COS) is a unique tracer for . . .”. See P2L2. P2L8 ‘The approach to estimate photosynthesis from COS fluxes’ This sentence has been removed for conciseness.


Introduction
Carbonyl sulfide (COS) has been shown as a unique tracer for land photosynthesis (also known as gross primary productivity, GPP) at regional to global scales (e.g., Montzka et al., 2007;Campbell et al., 2008;Berry et al., 2013;Hilton et al., 2017;Campbell et al., 2017). Globally, the largest sinks of COS are uptake by leaves and soils, and the largest sources are ocean emissions, followed by additional emissions from anthropogenic activities (Montzka et al., 2007;Berry et al., 2013;Launois 5 et al., 2015;Campbell et al., 2015). Since vegetation uptake dominates the COS exchange in land ecosystems (Berry et al., 2013), concurrent measurements of COS and CO 2 fluxes can be used to separate photosynthesis and respiration from the net carbon flux (Asaf et al., 2013;Billesbach et al., 2014). The COS tracer approach to photosynthesis is based on the coupling of leaf COS and CO 2 uptake (Sandoval-Soto et al., 2005;Seibt et al., 2010;Stimler et al., 2010Stimler et al., , 2011Wohlfahrt et al., 2012).
Understanding the quantitative relationship that ties together leaf COS and CO 2 fluxes is key to obtaining accurate estimates 10 of photosynthesis from COS measurements.
In leaves, COS and CO 2 follow the same stomatal diffusional pathway and similar hydrolytic reactions catalyzed by carbonic anhydrase (CA), with the main difference being that the hydrolysis goes reversibly for CO 2 but one-way for COS (Protoschill-Krebs et al., 1996;Notni et al., 2007). The reaction of COS with CA yields H 2 S and CO 2 (Schenk et al., 2004;Notni et al., 2007), without any observed COS (re)emission from leaves (Stimler et al., 2010). In contrast, CO 2 hydration is subject to 15 chemical equilibrium that depends on its diffusional supply versus its demand from fixation, leading to retrodiffusion to the atmosphere. CA-mediated hydrolysis therefore serves as the sink reaction of COS in leaves, but not of CO 2 .
The COS hydrolysis via CA has been shown to be light independent (Goldan et al., 1988;Protoschill-Krebs et al., 1996).
Since this reaction is also highly efficient (Ogawa et al., 2013), the COS uptake rate should be mostly controlled by the sequence of conductances along the diffusional pathway into leaves, i.e., substrate limited rather than enzyme limited (Goldan 20 et al., 1988;Sandoval-Soto et al., 2005;Seibt et al., 2010;Stimler et al., 2010). Leaf COS uptake should therefore respond to environmental variables that regulate stomatal conductance, including photosynthetically active radiation (PAR), because of the feedback from photosynthesis to stomatal conductance (e.g., Ball, 1988;Collatz et al., 1991). Thus, light regulates leaf COS uptake even though COS hydrolysis itself does not depend on light.
In contrast to the CO 2 flux that turns to emission at night, COS uptake may continue if stomata are not fully closed (Stimler 25 et al., 2010). To understand the relationship between daily integrated COS and CO 2 fluxes for regional flux inversion (e.g., Hilton et al., 2015), nighttime COS uptake needs to be constrained . Nighttime COS uptake has been found in a wheat field , a boreal pine forest (Kooijmans et al., 2017), and temperate forests (Berkelhammer et al., 2014;Commane et al., 2015;Wehr et al., 2017). However, most field studies based their findings upon indirect evidence of nighttime ecosystem COS uptake, with only one study reporting some direct leaf observations of nighttime uptake 30 (Berkelhammer et al., 2014).
The quantitative relationship between leaf COS uptake and photosynthesis required for COS-based photosynthesis estimatesfrom canopy to regional scales (e.g., Asaf et al., 2013;Campbell et al., 2017)-is commonly expressed in one parameter: leaf relative uptake (LRU). LRU is the ratio of leaf COS : CO 2 fluxes normalized by their respective ambient concentrations (Sandoval-Soto et al., 2005;Campbell et al., 2008). A mean LRU value of 1.6 has been reported for a wide range of species from leaf scale measurements in the laboratory (Stimler et al., 2010(Stimler et al., , 2011(Stimler et al., , 2012 and the field (Berkelhammer et al., 2014). But in the field, lower LRU values have also been observed, e.g., 1.3 in a wheat field  and 1.2 in a temperate forest (Commane et al., 2015), both estimated from ecosystem scale measurements.
For ecosystem and larger scale applications, a constant LRU of 1.6 has been assumed (e.g., Asaf et al., 2013;Hilton et al., 5 2015) despite the known dependence of LRU on PAR. LRU is found to decrease with light in both laboratory and field observations (Stimler et al., 2010(Stimler et al., , 2011Maseyk et al., 2014;Commane et al., 2015). Leaf level measurements in the laboratory show that LRU is stable at PAR above ca. 500 µmol m −2 s −1 , but increases sharply with decreasing PAR (Stimler et al., 2010(Stimler et al., , 2011. The stable LRU region is consistent with that of light-saturated photosynthesis and maximal stomatal conductance, and therefore low variations in COS and CO 2 fluxes (Stimler et al., 2011). At low light, the rate at which LRU increases differs 10 among species, with some showing a sharp increase to LRU values of ca. 9, while others show a more gradual or only slight increase. This LRU behavior results from the diverging responses of COS and CO 2 uptake in low light: CO 2 assimilation that is also controlled by light decreases more rapidly than COS uptake that is only controlled by stomatal conductance. Using a light dependent LRU instead of a constant value is therefore necessary for COS-based photosynthesis estimates. But in the field, the LRU-PAR relationship has only been approximated with ecosystem fluxes Commane et al., 2015), not 15 directly determined from leaf fluxes. For COS-based canopy photosynthesis estimates, we need direct knowledge of how LRU responds to PAR and other possible drivers in the field. Applications for longer timescales would further need daily integrated LRU values. This study is motivated by two research questions: 1) How does light control instantaneous and daily integrated LRU values?
2) How do stomatal responses to environmental variables regulate leaf COS uptake in the field? We report leaf COS and CO 2 20 fluxes measured in a Typha latifolia freshwater marsh during the peak growing season of June and July 2013. We then examine how environmental variables control fluxes and LRU through stomatal mechanisms, and discuss the implications for COSbased photosynthesis estimates. 25 We measured leaf fluxes of COS, CO 2 , and water from 31 May to 6 July 2013 (day of year 151-187) at the San Joaquin Freshwater Marsh (SJFM, 33 • 39 44.4 N, 117 • 51 6.1 W). The SJFM is located near the campus of the University of California at Irvine, at 3 m above sea level and 8 km northeast of the Pacific Ocean (Goulden et al., 2007). The SJFM is part of the University of California's Natural Reserve System. The site history and management have been described in Goulden et al. (2007). Briefly, the SJFM is a mature freshwater marsh, the remnant of once a 2100 ha wetland along the San Diego Creek. Since the 1960s, 30 the SJFM has been managed by flooding the area annually to a depth of approximately 1 m from December/January to March.

Site description
The standing water recedes by evapotranspiration and subsurface drainage and eventually disappears by midsummer (Goulden et al., 2007). A flux tower (5 m high) is located on a floating wooden platform near the northeastern edge of the SJFM. The platform is surrounded by dense vegetation dominated by Typha latifolia (broadleaf cattail). In contrast to most species in a mediterranean climate that grow in the rainy winter or early spring, the growing season of the marsh plants is summer due to the standing water.

Experimental setup
Leaf fluxes of COS, CO 2 , and H 2 O were measured with a flow-through (dynamic) chamber. The cylindrical chamber (18 cm 5 diameter, 38 cm height, 10.3 L volume) consisted of PFA Teflon film stretched between two aluminum rings connected by rods.
The PFA film was laid inside the structure such that only the Teflon was in contact with the sampled air. The chamber enclosed the upper sections of six tall cattail leaves with an average width of 1.5 cm. The leaves extended above and below the chamber.
The total leaf area in the chamber was estimated as 409.5 cm 2 . Skirts of Teflon film were wrapped around the leaves to provide a seal at both ends of the chamber. On one end, a high-speed axial fan (D344T, Micronel) was installed to provide ventilation 10 to keep the chamber at ambient conditions (i.e., within 1 p.p.m.v. of ambient CO 2 , tested at the start of the campaign). During measurement periods, the fan was turned off and its opening served as the inlet to allow airflow through the chamber. A second, smaller fan (F62, Micronel), attached to a stainless steel rod and placed inside the chamber, ran continuously to mix the air within the chamber. (ultrahigh purity) for a one-minute background correction every hour. Data from the QCL analyzer were recorded at 10 Hz and stored on the QCL hard drive. The RMS noise (1 σ) at 10 Hz was 11-18 parts per trillion in volume (p.p.t.v.) for COS during chamber measurements.
The leaf chamber was measured once per hour. We monitored chamber air concentrations during the five-minute measurement periods (i.e., while the ventilation fan was off), as well as the ambient air for one minute before and after these periods 25 (i.e., while the fan was running). Leaf fluxes were calculated from the transient changes with respect to the interpolated inlet (ambient) concentrations (Fig. 1). The apparent fluxes from blank chambers were characterized and were found to be negligible. Environmental data were obtained from various sensors including photosynthetically active radiation (PAR) (SQ-215, Apogee Instruments), ambient air temperature and humidity (HMP45AC, Vaisala), and chamber air and leaf temperature (type T thermocouples, PFA coated), and were stored at 10 s intervals on a datalogger (CR1000, Campbell Scientific). The datalogger 30 also controlled the operation of the high-speed ventilation fan.

Calculation of leaf fluxes
A mass balance equation is formulated for the gas species being measured (COS, CO 2 , or H 2 O), ] are the chamber volume and leaf area, respectively, and F [mol m −2 s −1 ] is 5 the flux rate to be calculated. Solving the mass balance equation with the initial condition C(t = 0) = C a , we obtain The flux rate F is then solved from the slope of the regressionŷ ∼ (1 −x). The standard error of the estimated F is also obtained from the regression. The flux calculation method described above does not require a steady state to be reached in the chamber.
A typical example of the chamber measurement period for COS with the fitted curve of concentration changes is shown in

Data quality control
All leaf flux and meteorological data have been quality checked and filtered. Conspicuously unrealistic data points in the meteorological data have been removed. For the flux data, we used several independent criteria to filter out bad measurements.
First, measurement periods with serious misfit of the shape of concentration changes during chamber closure or with strong drift in the ambient concentrations were discarded. Second, flux estimates associated with large RMSEs between fitted and 20 observed concentrations were also filtered out. Then, outliers in flux data were detected using the well-established Tukey's interquartile range method (Wilks, 2011). In addition, strongly positive CO 2 fluxes during the day and strongly negative CO 2 fluxes at night were also removed. Only the data points that passed all these filtering procedures were kept in the final data for analysis.

25
Leaf COS : CO 2 relative uptake ratio (LRU) is defined as the ratio of COS and CO 2 fluxes (F COS and F CO 2 ) normalized by their respective concentrations ( χ COS and χ CO 2 ), LRU is a dimensionless quantity. We confine our LRU analysis to occasions where both COS and CO 2 fluxes are negative (i.e., showing net uptake). Hence, LRU is only calculated during the daytime and is always positive.

Fitting light response curves for leaf COS and CO 2 fluxes and LRU
We used the LOWESS (locally weighted scatterplot smoothing) regression method to obtain smooth light response curves for COS flux, CO 2 flux, and LRU (see Fig. 5). The LOWESS regression method is a nonparametric method that does not require 5 any a priori known relationship between the predictor (here, PAR) and the response variables (COS flux, CO 2 flux, and LRU).
At each point in the range of the predictor, a low-degree polynomial is fitted to all the neighboring points to estimate the least squares response, weighted by the distances between the neighboring points and the current point (Cleveland et al., 1992). The calculation was performed with the Python statsmodels package (Seabold and Perktold, 2010).

10
During the campaign period in June 2013 covering the peak growing season of Typha latifolia, meteorological conditions changed little except for a few cloudy days (day of year 159, 160, and 181, Fig. 2d), and the diurnal patterns of leaf COS, CO 2 , and H 2 O fluxes therefore also remained similar (Fig. 2). The diurnal patterns of leaf fluxes and related variables are visualized with hourly binned medians and quartiles (Fig. 3). During the day, leaf uptake of COS and CO 2 showed similar patterns ( Fig. 3a, b), with uptake peaks in the morning and afternoon separated by a midday depression around local noon 15 (13:00). The midday depression was up to 36% for COS (5.5 pmol m −2 s −1 at 14 h versus 8.5 pmol m −2 s −1 at 11 h) and 40% for CO 2 (3.7 µmol m −2 s −1 at 13 h versus 6.1 µmol m −2 s −1 at 17 h), respectively. The morning peaks coincided for the two fluxes at around 11:00, whereas the afternoon peak occurred slightly later for COS (18:00) than for CO 2 (17:00). The afternoon peak of CO 2 flux was slightly stronger than its morning peak (Fig. 3b, c), probably because the chamber received more light in the afternoon than in the morning (Fig. 3e) due to a wider gap in the canopy to the west of the chamber than to other 20 directions. Leaf transpiration showed a decline at 11:00 ( Fig. 3c), but with an earlier afternoon peak (16:00) that coincided with the maximum vapor deficit (Fig. 3f). Contrary to COS and CO 2 fluxes, the diurnal pattern of water flux was strongly asymmetric due to the high vapor deficit in the afternoon (Fig. 3f), although the midday depression in stomatal conductance was roughly symmetric as indicated by COS uptake.
In contrast to daytime fluxes, nighttime fluxes of COS and CO 2 showed diverging patterns. At night, CO 2 was emitted 25 from leaf respiration (Fig. 3b), whereas COS uptake continued (Fig. 3a). Both fluxes had significantly smaller magnitudes than during the day, with CO 2 emissions of around 1 µmol m −2 s −1 , and COS uptake of around 2-3 pmol m −2 s −1 . Note that although COS emissions were occasionally observed at night (Fig. 2a), they were likely caused by the measurement uncertainty from high flow rates (∼6 s.l.m.), and the hourly medians indeed showed a robust pattern of nighttime COS uptake (Fig. 3a).
When averaged over the whole campaign, nighttime COS uptake was 23% of the total daily COS uptake by leaves. Nighttime 30 transpiration was minimal (Fig. 3c) as the vapor deficit was close to zero at night (Fig. 3f). Leaf relative uptake (LRU), the ratio of COS to CO 2 uptake normalized by their respective concentrations in the chamber, showed an asymmetric U-shape diurnal pattern (Fig. 3d). The LRU had highest values of 2-3 (medians binned by the hour) near dawn or dusk, with a gradual decrease throughout the morning and early afternoon, and had minima around 0.9 at 15:00 coinciding with the dip in COS uptake (Fig. 3d). LRU was stable in the late afternoon until an abrupt increase at 19:00 before sunset.

5
Overall, COS flux was well correlated with CO 2 flux, with an r 2 of 0.49 (Fig. 4a), reaffirming the shared stomatal control on both fluxes. The correlation between COS and water fluxes was lower: r 2 = 0.32 (Fig. 4b), and showed a wide spread during the day due to the asymmetric diurnal pattern of water fluxes (Fig. 3c). At night, COS fluxes showed larger variability than water fluxes as the vapor deficit was small (Fig. 3f).
The diurnal pattern of LRU (Fig. 3d) was consistent with the LRU response to PAR (Fig. 5c). The LRU values decreased 10 with increasing PAR (Fig. 5c) to around 1.0 at PAR above around 500-600 µmol m −2 s −1 . Surprisingly, the lowest LRU values during the day did not occur at the time of the highest PAR (Fig. 3d), but rather at the time of the highest vapor deficit (Fig. 3f) and moderately strong PAR (1000-1400 µmol m −2 s −1 ) due to the stronger stomatal limitation on fluxes as a response to the high evaporative demand. Leaf fluxes of COS and CO 2 showed similar light responses, increasing with PAR until they become light saturated, and decreasing at high light and high evaporative demand (Fig. 5a, b). However, the similarity in fluxes is not due to a common light response of the biochemical reactions that consume COS and CO 2 in leaves, since COS hydrolysis is light independent.
Instead, underlying the similar diurnal patterns and light responses (Figs. 3a, b, 5a, b) is the shared response of leaf COS 20 and CO 2 uptake to stomatal conductance, which increases with light because of the feedback between stomatal conductance and photosynthesis (Cowan, 1978;Farquhar and Sharkey, 1982;Ball, 1988;Collatz et al., 1991). At high light, when CO 2 assimilation is light saturated, leaf COS and CO 2 uptake is controlled by stomatal conductance in a similar way: both decline as stomatal conductance is reduced in response to high evaporative demand (Fig. 3a, b). At low light, COS and CO 2 diffusions are both reduced by low stomatal conductance, but CO 2 assimilation is additionally reduced by low light, causing a stronger 25 decrease in CO 2 uptake than COS uptake (Fig. 5a, b).
The most striking feature in the diurnal patterns of leaf COS and CO 2 uptake was the concurrent midday depression in the early afternoon (Fig. 3a-c), also affecting the light response curves of fluxes (Fig. 5a, b). From the smoothed light response trends (Fig. 5a, b), we found that COS uptake reached the maximum of 7.5 pmol m −2 s −1 at PAR = 493 µmol m −2 s −1 and decreased to 4.7 pmol m −2 s −1 at PAR = 1800 µmol m −2 s −1 (the typical PAR level at local noon), whereas CO 2 uptake reached 30 the maximum of 5.3 µmol m −2 s −1 at PAR = 740 µmol m −2 s −1 and decreased to 3.7 pmol m −2 s −1 at PAR = 1800 µmol m −2 s −1 .
The respective 37% and 31% reductions in COS and CO 2 uptake at typical midday light (1800 µmol m −2 s −1 ) with respect to their peak uptake indicate that stomatal conductance exerted a stronger control on COS uptake than CO 2 uptake (see sect. 4.3).
This behavior was driven by the stomatal response to high vapor deficit that always coincided with high PAR (Fig. 2d, e).
The reduction of stomatal conductance under high vapor deficit is a well-documented behavior that serves to curb excessive loss of water and optimize water use against carbon gain (Tenhunen et al., 1984;Ball, 1988;Collatz et al., 1991;Leuning, 1995).
Previously, the midday depression in plant COS uptake has been inferred from canopy scale measurements in a Mediterranean 5 pine forest in the winter (Asaf et al., 2013) and in a temperate forest in the summer (Commane et al., 2015), but has not been investigated directly at the leaf level. The current study, to our knowledge, offers the first field observations of the influence of midday depression on COS uptake at the leaf scale and reaffirms stomatal conductance as the dominant control of COS uptake.

COS uptake is an indicator of nocturnal stomatal conductance
The coupling between leaf COS and CO 2 fluxes breaks down at night because leaves produce CO 2 due to respiration, whereas 10 COS uptake may continue if stomata are not fully closed. At this site, nocturnal uptake contributed 23% of the total daily leaf COS uptake. This fraction is comparable to those reported from a wheat field (29 ± 5%, Maseyk et al., 2014), an alpine temperate forest (25-30%, Berkelhammer et al., 2014), a boreal pine forest (17%, Kooijmans et al., 2017), and a New England mixed forest (< 20% after subtracting soil uptake, Commane et al., 2015;Wehr et al., 2017). Collectively, these studies indicate that nocturnal uptake is typically 17-30% of the total canopy COS budget, a fraction that is too large to ignore in ecosystem and 15 regional COS budget studies. Understanding nocturnal COS uptake will therefore be necessary for COS-based photosynthesis estimates at daily and longer timescales.
For the T. latifolia leaves here, we obtained a mean value of 5 ± 1 mmol m −2 s −1 for the nocturnal stomatal conductance to COS (g s,COS ) if internal conductance (g i,COS ), the combination of mesophyll conductance and biochemical reaction coefficient, is ignored (g s,COS g i,COS ). This translates to 10±2 mmol m −2 s −1 for the stomatal conductance to water (g s ), after accounting 20 for the different diffusivities of water and COS in the air with a ratio of 2.0 (Seibt et al., 2010). The nocturnal g s,COS is at the lower end of values reported for other ecosystems, ranging from 1.6 mmol m −2 s −1 for a New England mixed forest (Wehr et al., 2017) to 5-20 mmol m −2 s −1 for a Scots pine forest (Kooijmans et al., 2017), 11.5 mmol m −2 s −1 for a wheat field , and 13-20 and 22-66 mmol m −2 s −1 for pine and poplar trees, respectively, in an alpine temperate forest (Berkelhammer et al., 2014).

25
Although these observations span a wide range of values across plant species and ecosystem types, the fraction of nocturnal uptake in the daily canopy COS budget lies in a much narrower range of 17-30%. This convergence indicates that nocturnal values may be directly coupled to daytime stomatal conductance. Hence, it may be beneficial for large scale applications to relate nocturnal stomatal conductance to daytime observable parameters, e.g., 5.5% of the light saturated value for a wheat field  or 2.5% of the daytime maximum value in a New England mixed forest (Wehr et al., 2017).

30
In land biosphere models, nocturnal stomatal conductance has been typically parameterized with a small fixed value regardless of plant type, for example, 10 mmol m −2 s −1 in the Community Land Model v4.5 (Oleson et al., 2013). This fixed-value parameterization may introduce biases in the nighttime COS fluxes and long-term COS budget in regional simulations, which in turn propagate into the COS-based photosynthesis estimates. For better estimates of nighttime COS fluxes and transpiration, open, whereas water fluxes become very small as the ambient air typically gets close to saturation at night. We expect COS measurements to be particularly beneficial in tropical rainforests and other environments that experience high humidity.

The environmental determinants of leaf relative uptake (LRU)
5 Leaf COS to CO 2 relative uptake (LRU) is an important parameter that links plant COS uptake with GPP. Observations at leaf and ecosystem scales show that LRU is primarily controlled by light, following an asymptotically decreasing trend with increasing PAR (Fig. 5c; Stimler et al., 2010Stimler et al., , 2011Maseyk et al., 2014;Commane et al., 2015). Such a pattern originates from the differential responses of COS and CO 2 uptake to light, because unlike photosynthesis, COS uptake responds only indirectly to light through changes in stomatal conductance (Stimler et al., 2011). Using the nonparametric LOWESS fit without assuming 10 an a priori relationship between LRU and PAR, we found an LRU-PAR relationship (Fig. 5c) similar to the decaying power law (LRU = a · PAR −b ) reported by Maseyk et al. (2014). Based on this and previous studies, the light response of LRU may be generalized empirically with a decaying power law fit (Stimler et al., 2010(Stimler et al., , 2011Maseyk et al., 2014;Commane et al., 2015).
We identified vapor deficit as secondary environmental driver of LRU, resulting from the differential effects of low humidity induced stomatal closure on COS and CO 2 fluxes (Fig. 5a, b; see also sect. 4.1). High vapor deficit tends to reduce LRU 15 values in mid-afternoon, when LRU is expected to reach light-saturated values according to the LRU-PAR relationship. This is because stomatal conductance is a more dominant component in the diffusional pathway for COS than for CO 2 . Using the resistance analog (the inverse of conductance, i.e., r s = g −1 s ), we can combine all sub-stomatal terms (mesophyll and chloroplast wall conductances and biochemical reaction coefficient) into a single internal resistance term (r i,COS or r i,CO 2 ). Because of the strong affinity of β-CA for COS (Ogawa et al., 2013), COS is more readily consumed at the CA active site than CO 2 is at 20 the carboxylation site of RuBisCO (Stimler et al., 2010;Berry et al., 2013), leading to a much smaller contribution of internal resistance to the COS diffusional pathway, For example, based on rough estimates of light-saturated values of stomatal conductance (g s,H 2 O = 80 mmol m −2 s −1 ) and COS and CO 2 fluxes (Fig. 5), for a relative decrease in stomatal conductance (g s,COS and g s,CO 2 ) of 50% at high vapor deficit, the 25 total resistance of COS uptake increases by 37% whereas that of CO 2 uptake only increases by 28%. Thus, when CO 2 uptake is light saturated, a decrease in stomatal conductance due to high vapor deficit will reduce COS uptake more than CO 2 uptake, and result in a lower LRU (7% for the examples above, from 1.07 to 1.0).
Previous laboratory studies have not found any significant response of LRU to relative humidity (Stimler et al., 2010(Stimler et al., , 2011, but it is possible that the vapor deficit in the experiments was not strong enough to initiate partial stomatal closure. At our site, The asymptotic LRU value at high light (PAR > 600 µmol m −2 s −1 ) at our site was around 1.0 (Fig. 5c). This value is lower than the mean LRU of 1.61 ± 0.26 from laboratory measurements across a range of species (Stimler et al., 2012), which has been used as a representative LRU in regional GPP inversion studies from COS measurements (e.g., Hilton et al., 2015). The low asymptotic LRU value reported here is, however, similar to values seen in some grasses and shrub species (Stimler et al., 2012). Lower LRU values have also been reported from field studies, for example, 1.3 in a wheat field  5 and 1.2 in a mixed temperate forest at high PAR (Commane et al., 2015). The discrepancy between LRU values measured under laboratory and field conditions may come from variations in environmental drivers, for example, vapor deficit, or plant water status that regulates stomatal responses. The LRU responses to environmental conditions can also differ by plant species (Stimler et al., 2012). In ecosystem or regional scale applications, LRU values that are diagnosed from process-based models (Berry et al., 2013;Hilton et al., 2015) may be preferable to an assumed value of 1.6. 10 4.4 Daily integrated leaf relative uptake ratio and its implications for regional flux estimates Beyond the ecosystem scale, daily values of LRU can be useful for large scale COS applications. The daily (24 h) mean LRU at this site showed large day-to-day variations (1.4-3.6) and also had large uncertainty due to the variability and measurement uncertainty in nighttime CO 2 fluxes (Fig. 6). In contrast, the daytime mean LRU, averaged over the day length of 14 hours, did not show strong variability (1.0-1.8) and had an average value of 1.2 across the campaign. The daytime mean LRU was 15 consistently lower than the daily (24 h) mean LRU, since the latter includes nocturnal COS uptake and CO 2 emissions.
We found a good correlation between daytime mean LRU and daytime mean PAR (r = −0.525; Fig. 6b), similar to Maseyk et al. (2014). This indicates that the LRU-PAR relationship is preserved at the daily timescale, supporting the use of COS as a photosynthetic tracer at large scales where measurements are often made at daily or longer intervals. On overcast days, the daytime mean LRU values were higher than on clear days (Fig. 6a), as expected from the light response of LRU. We expect 20 the relationship between daytime means of LRU and PAR to be useful for calculating daytime mean LRU empirically from meteorological conditions for GPP estimates. Since the use of COS as a GPP tracer in an inverse modeling framework requires the uncertainty in LRU to be smaller than that in the a priori GPP estimates , future studies should be dedicated to understanding LRU variability in the field for accurate COS-based GPP estimates.

25
From direct field observations at the leaf scale, our study has shown that leaf COS and CO 2 fluxes share broadly similar diurnal patterns driven by the common stomatal responses to light and vapor deficit. In the early morning and late afternoon, the increase of COS uptake with light is caused by increasing stomatal conductance, since the COS reaction with CA is light independent. Around midday, vapor deficit becomes a limiting factor of stomatal conductance and drives the midday depression in COS and CO 2 uptake. 30 We have identified three distinct physiological regimes that control LRU variability over the course of a day: 1. In the early morning when both PAR and vapor deficit are low, biochemical reactions of CO 2 are light limited. As a result, leaf CO 2 uptake is more restricted than COS uptake that is only stomatal conductance limited, causing LRU to be high and to decrease with PAR.
2. Around midday and in the early afternoon when both PAR and vapor deficit are high, midday depression occurs and both CO 2 and COS diffusion processes are limited by the low stomatal conductance. Since COS uptake is more sensitive to 5 stomatal conductance, vapor deficit becomes the key driver of LRU at this time of the day.
3. In the late afternoon when PAR declines but vapor deficit is still quite high, stomatal conductance is more limited by vapor deficit compared to the morning. This causes COS uptake and LRU to be lower than in the corresponding morning time with the same PAR, and leads to the asymmetric diurnal pattern of LRU.
We have validated the previously reported light dependence of LRU directly at the leaf scale in field conditions. At high 10 light, LRU converges to 1.0, much lower than the typical value of 1.6 reported from laboratory conditions. In addition, we identified vapor deficit as a secondary but non-negligible effect on LRU when it begins to limit stomatal conductance. The LRU-PAR relationship also holds between the daytime mean LRU and PAR values. The coupling between leaf COS and CO 2 fluxes during the peak growing season of the Typha latifolia vegetation lends strong support to the use of COS as a quantitative tracer for canopy photosynthesis.