Carbonyl sulfide (COS) is used as a tracer of
Carbonyl sulfide (COS) is a sulfur-containing analogue of
For COS application as a tracer of ecosystem
Soil COS exchange has often been measured by incubations in the lab (e.g.,
Bunk et al., 2017; Kesselmeier et al., 1999; Liu et al., 2010; Van Diest and
Kesselmeier, 2008), by static or dynamic chambers in the field (e.g.,
Berkelhammer et al., 2014; Kitz et al., 2017; Sun et al., 2018; Yi et al.,
2007; Mseyk et al., 2014) and using models (e.g., Ogée et al., 2016; Sun
et al., 2015; Whelan et al., 2016). In spite of these efforts, more field
measurements of soil COS exchange are clearly needed as a basis for
elucidating underlying mechanisms, as well as obtaining a better quantitative
record of the possible range of soil COS fluxes under natural conditions.
Note also that previous studies have focused on agricultural soils (Maseyk et
al., 2014), wetlands (Whelan et al., 2013), boreal forest soils (Sun et al.,
2018) and grasslands (Kitz et al., 2017), but several ecosystems are
understudied, such as in those the Mediterranean. Finally, soil profile
measurements will also be useful for the validation of soil models of COS
exchange (Sun et al., 2015). The objective of this study was to apply dynamic
chambers measurements, constrained by simultaneous soil gradient method, to
assess the spatial and temporal variations in soil COS and
The study was conducted in an orchard in Rehovot, Israel
(31
We used the commercially available quantum cascade laser (QCL) system
(Aerodyne Research, Billerica, MA) with a tunable laser absorption spectrometer
(model: QC-TILDAS-CS) to measure COS,
A custom-made stainless-steel cylindrical chamber of 177 cm
Gas exchange rates,
Four campaigns of soil concentration profile measurements were carried out during 1–2 March, 20–26 April, 10 May and 22–28 June 2016. The trace gas at five soil depths of 0, 2.5, 5.0, 10 and 20 cm was sampled at each of the three micro-sites: BR, BT and UT.
Four individual Decabon tubes were inserted at adjacent but different points
into the soil (to avoid communication between tubes during sampling) to the
different depths indicated above and connected directly to the QCL positioned
close by the mobile lab. At least 1 day after insertion and ensuring
sealing between tubing and soil, soil air was sampled with flow rate of
80 mL min
Assuming that at the selected measurement sites, soil trace gas is only
transported by diffusion, soil COS and
The Penman (1940) function was used to describe the soil diffusion
coefficient (
Soil COS fluxes showed significant heterogeneity at both the spatial
(micro-sites) and temporal (seasonal) scale (Fig. 1). Overall, the hourly soil
COS flux varied from
Spatial variability of soil COS flux at three sites: between
trees
On the diurnal timescale, soil COS flux were generally higher in the
afternoon (peaking around 15:00–16:00), declining at night and in the early
morning (Fig. 1). On the seasonal timescale, soil COS fluxes showed both
changes in rates and shifts from net uptake to net emission, with the site
hierarchy differing in the different seasons (Fig. 1). At the UT site where
only COS uptake was observed, the highest rates were observed in winter and
peak summer (December and August) with diurnal mean rates of nearly
During the hot summer (August 2015 and July 2016), differences in micro-site
soil water content (
Mean values of soil COS and
Relationship of soil COS flux and soil moisture. Each data point
represents the diurnal average (
The response of soil COS fluxes to soil temperature varied among the three
measurement sites (Fig. 3). The BT and BR sites showed a near-linear response
with a shift from uptake to emission around 25
Soil COS flux as a function of temperature and its linear
regression line. Each data point represents the diurnal average (
Pearson product–moment correlation analysis results showed that hourly soil
COS flux was significantly related to soil moisture and temperature (at the
0.001 level), and the soil moisture had stronger environmental controls on
the soil COS flux (
A comprehensive assessment of the effects of soil moisture (
The average soil concentration gradient of COS and
Mean COS and
Soil COS and
As noted above, the profile data generally exhibited the steepest gradient in the top few centimeters of the soil, indicating that the dominating COS sink (and
likely also the
Estimates of soil COS and
Soil was always a source of
The relationships between soil COS and
The observed soil–atmosphere COS exchange rates observed in this study (both
mean and range; Fig. 1, Table 1) are consistent with values reported in a
range of other ecosystems (
The observed range in the soil–atmosphere exchange fluxes reflected
significant heterogeneity on both the spatial and the temporal scales. The
spatial-scale heterogeneity clearly reflected the contrasting micro-site
conditions with lower temperatures and higher moisture under the trees (UT
sites) compared with the higher temperatures and lower moisture in exposed
soil between rows (BR sites), with intermediate, partially shaded, conditions
between trees (BT sites). Indeed, a large fraction of the variations in the
COS flux (
Temporal variations were observed both on the daily and seasonal timescales.
Diurnal changes were, however, minor compared to the changes from winter to
summer at all micro-sites. Shifts from uptake to emission were observed
essentially only on the seasonal timescale (Fig. 1). This likely reflected
the dominance of soil moisture in relation to the COS flux rates. This is because
COS uptake is thought to be related to carbonic anhydrase (CA) activity in soil (Kesselmeier et al., 1999), which could be via microorganisms (Piazzetta et al., 2015), such as bacteria (Kamezaki et al., 2016; Kato et al., 2008) or fungi (Bunk et al., 2017; Li et al., 2010; Masaki et al., 2016). CA activity is also influenced by soil moisture (Davidson and Janssens, 2006; Seibt et al., 2006), although soil moisture can also directly influence soil gas diffusion rates (Ogée et al., 2016; Sun et al., 2015). The effect of CA on COS exchange can also be related to root distribution and the effects of CA activity within plant roots (Seibt et al., 2006; Viktor and Cramer, 2005; Whelan and Rhew, 2015). This could influence the spatial variations and soil moisture effects on COS exchange in this study as most of the roots were distributed around the restricted trees' drip irrigation zone at UT sites and were sparse in the dryer areas, such as the BR and BT sites (unquantified observations).
At least part of the variations in soil COS fluxes could also reflect the
differential effects of environmental conditions on COS uptake and production
process (Ogée et al., 2016). Solubility in soil water (with a COS
solubility of 0.8 mL mL
We use SRU values to assess the relative importance of the soil COS flux
compared with the canopy and indicate shifts from conservative links between
processes influencing COS and
The differential effects of changing environmental conditions on production and
uptake processes were reflected in the relatively large spatial and temporal
heterogeneity observed in the soil COS exchange at our site. However, the
contrasting effects of production and emission may explain both the sharp
increase in SRU values at high temperatures, as the effects of production
counteract uptake (Fig. 6b), and the much lower sensitivity to temperature of
COS flux compared to that of
The application of COS as a tracer for canopy
Complementing our chamber measurements with soil profile measurements of COS
and
Note that the gradient method based on the Fick's diffusion law has its own limitations (Kowalski and Sánchezcañete, 2010; Sánchez-Cañete et al., 2017; Bekele et al., 2007). However, it is a simple, low-cost approach and can help diagnose the magnitude of soil fluxes, which can also help in identifying belowground processes and their locations.
Our detailed analysis of the spatial and temporal variations in soil–atmosphere exchange of COS provided new information on a key uncertainty in the application of ecosystem COS flux to assess productivity. Furthermore, we provide validation of the surface chamber measurements that are generally in use, by the additional gradient approach. Our results show that both micro-sites and seasonal variations in COS fluxes were related to soil moisture, temperature and the distance from the tree (likely reflecting root distribution), but we suggest that soil moisture is the predominant environmental control over soil COS exchanges at our site. A simple algorithm was sufficient to forecast most of the variations in soil COS flux, supporting its incorporated into ecosystem-scale applications, as we recently demonstrated in a parallel study at the same site (Yang et al., 2018).
Clearly, uncertainties are still associated with soil processes involving COS, the differential effects of soil moisture, temperature and communities of microorganisms, and they are likely to contribute to both the spatial and temporal variations in soil net COS exchange and require further research.
The data used in this study are archived and available from the corresponding author upon request (dan.yakir@weizmann.ac.il).
DY designed the study; FY, RQ, FT, RS and DY performed the experiments. FY and FT analyzed the data. DY and FY wrote the paper with discussions and contributions to interpretations of the results from all coauthors.
The authors declare that they have no conflict of interest.
We are grateful to Omri Garini, Madi Amer and Boaz Ninyo-Setter for their help. This work was supported by the Minerva foundation, a joint NSFC-ISF grant 2579/16, the Israel Science Foundation (ISF 1976/17), the German Research Foundation (DFG) as part of the CliFF Project and the JNF-KKL. Fulin Yang is supported by the National Natural Science Foundation of China (41775105) and the Natural Science Foundation of Gansu Province (17JR5RA341).
This paper was edited by Janne Rinne and reviewed by Mary E. Whelan and one anonymous referee.