Global gridded anthropogenic emissions inventory of carbonyl sulfide

Atmospheric Environment 183 (2018) 11–19

Contents lists available at ScienceDirect

Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv

Global gridded anthropogenic emissions inventory of carbonyl sulfide a,∗

b

c

d

e

T e

Andrew Zumkehr , Tim W. Hilton , Mary Whelan , Steve Smith , Le Kuai , John Worden , J. Elliott Campbellb,∗∗ a

Sierra Nevada Research Institute, University of California, Merced, Merced, CA, 95340, USA Department of Environmental Studies, University of California, Santa Cruz, Santa Cruz, CA, 95064, USA Joint Global Change Research Institute, PNNL, College Park, MD, 20740, USA d Department of Atmospheric and Oceanic Sciences, University of California, Los Angeles, Los Angeles, USA e JPL Earth Sciences, California Institute of Technology, Pasadena, CA, 91125, USA b c

A R T I C LE I N FO

A B S T R A C T

Keywords: Carbonyl sulfide Anthropogenic Global Atmospheric sources

Atmospheric carbonyl sulfide (COS or OCS) is the most abundant sulfur containing gas in the troposphere and is an atmospheric tracer for the carbon cycle. Gridded inventories of global anthropogenic COS are used for interpreting global COS measurements. However, previous gridded anthropogenic data are a climatological estimate based on input data that is over three decades old and are not representative of current conditions. Here we develop a new gridded data set of global anthropogenic COS sources that includes more source sectors than previously available and uses the most current emissions factors and industry activity data as input. Additionally, the inventory is provided as annually varying estimates from years 1980–2012 and employs a source specific spatial scaling procedure. We estimate a global source in year 2012 of 406 Gg S y−1 (range of 223–586 Gg S y−1), which is highly concentrated in China and is twice as large as the previous gridded inventory. Our large upward revision in the bottom-up estimate of the source is consistent with a recent top-down estimate based on air-monitoring and Antarctic firn data. Furthermore, our inventory time trends, including a decline in the 1990's and growth after the year 2000, are qualitatively consistent with trends in atmospheric data. Finally, similarities between the spatial distribution in this inventory and remote sensing data suggest that the anthropogenic source could potentially play a role in explaining a missing source in the global COS budget.

1. Introduction

(Campbell et al., 2017; Hilton et al., 2017). While other ecosystem sources and sinks have been observed (Commane et al., 2015; Whelan et al., 2016), the COS plant sink is dominant at regional scales (Campbell et al., 2008). Existing observations of COS concentrations, from NOAA airborne and tower monitoring networks, as well as remote sensing platforms, can be utilized to drive the regional COS tracer approach (Carmichael, 2003; Montzka et al., 2007; Kuai et al., 2014, 2015; Glatthor et al., 2015; Wang et al., 2016). The major source of atmospheric COS is derived from oceans (Berry et al., 2013; Kuai et al., 2014; Launois et al., 2014; Glatthor et al., 2015). The spatial separation of this source from the dominant sink of COS by terrestrial vegetation allows for continental regions to be well suited for the COS tracer approach. However, other continental sources and sinks of COS exist at certain locations and times. These other fluxes must be quantified for the COS tracer approach to be applied with confidence. Of these other fluxes that must be quantified, the largest continental source is emissions from anthropogenic activities (Campbell et al., 2015).

Measurement based estimates of regional-scale carbon fluxes are needed to improve our understanding of carbon-climate feedbacks (Cox et al., 2000; Friedlingstein et al., 2006; Field et al., 2007). While measurements of CO2 over terrestrial vegetation are useful for quantifying the net land-atmosphere exchange of carbon (Gurney et al., 2002), they are not useful for partitioning the underlying photosynthesis (gross primary production = GPP) and respiration fluxes at regional scales. An emerging approach for solving this problem, is the use of regional and global COS observations to estimate the underlying GPP sink (Montzka et al., 2007; Campbell et al., 2008, 2017; Hilton et al., 2017]. The dominant COS sink at regional scales is uptake by terrestrial vegetation in a process that is closely related to GPP (Sandoval-Soto et al., 2005; Campbell et al., 2008; Berry et al., 2013). The COS plant sink is largely controlled by stomatal conductance which in turn is closely related to GPP, particularly for scales of analysis that are integrated in both space (regional to global) and time (monthly to annual) ∗

Corresponding author. Corresponding author. E-mail addresses: [email protected] (A. Zumkehr), [email protected] (J.E. Campbell).

∗∗

https://doi.org/10.1016/j.atmosenv.2018.03.063 Received 29 August 2017; Received in revised form 28 March 2018; Accepted 31 March 2018 Available online 31 March 2018 1352-2310/ © 2018 Elsevier Ltd. All rights reserved.

Atmospheric Environment 183 (2018) 11–19

A. Zumkehr et al.

note that the most recent synthesis of source estimates and emission factors for biomass burning is reported by Campbell et al. (2015). Furthermore, soil emissions from managed lands are not included here as they are typically modeled separately using soil biogeochemistry models that account for concurrent soil sinks and sources (Whelan et al., 2016; Hilton et al., 2017). We estimate emissions for the years 1980 through 2012 due to the limited availability of concurrent data for COS sources. Data past 2012 exists for some sources, but these data were not incorporated into our result for simplicity in presenting the dataset. For each anthropogenic source, four parameters are required to estimate the magnitude of the source and then to distribute the source in space and time: industry activity, emissions factor, sub-country spatial scaling, and temporal scaling. Often the industry activity data set also provides the information for country-level spatial and temporal scaling. Sub-country spatial scaling is performed by distributing country-level estimates to a grid based on a gridded proxy flux. The IPCC code associated with the gridded proxy flux is used as the criteria for identifying the flux that most closely describes the specific anthropogenic estimate that needs to be distributed. Table 1 summarizes the input data used in the creation of each of the anthropogenic sources of COS calculated in this study. The following sections describe the specific methods and data used to create the estimates of direct or indirect emissions of anthropogenic COS for each source sector.

The current gridded anthropogenic COS inventory suffers from multiple limitations that weaken its suitability for use in the COS tracer approach. This approach involves the use of gridded inventories as input to 3-D atmospheric transport models to constrain the influence of anthropogenic activities on COS observations (Berry et al., 2013). The only currently available global gridded anthropogenic inventory is from Kettle et al. (2002). However, recent work has demonstrated that the Kettle inventory underestimates the magnitude of this source, has a biased spatial distribution, and ignores large inter-annual variation in the source (Campbell et al., 2015). These limitations of the Kettle inventory are due to multiple factors including the use of industry activity data from over three decades ago, integration of a small subset of the available emission factor data, and spatial scaling by global SO2 emissions rather than industry activity data. More recent inventories of global anthropogenic COS sources have corrected for many of these limitations (Campbell et al., 2015; Lee and Brimblecombe, 2016), but have not been extrapolated to spatially-explicit data sets that are the required input for atmospheric transport models. Additionally, highresolution regional gridded data have been developed (Blake et al., 2004; Zumkehr et al., 2017), but these inventories do not provide temporally-explicit estimates and they are limited to the U.S. and Asia. In addition to using gridded anthropogenic inventories to infer regional GPP, these inventories may also help resolve a missing source in the global COS budget. Comparisons of global atmospheric transport simulations and COS observations have identified a large missing source of atmospheric COS that is 230–800 Gg S yr−1, and has generally been attributed to a missing tropical oceanic source (Suntharalingam et al., 2008; Berry et al., 2013; Glatthor et al., 2015; Kuai et al., 2015; Lennartz et al., 2016). However, ocean cruise data show under-saturation of ocean surface waters with respect to COS and low global annual direct emissions of COS from oceans, suggesting that the oceans alone may not be able to account for the missing COS source (Lennartz et al., 2016). While additional ocean cruise data are needed, revisiting the global simulation studies with updated anthropogenic inventories would be useful for testing the hypothesis that anthropogenic activities could contribute to the missing COS source. Given the limitations of the Kettle gridded inventory and the recent advances in understanding of the anthropogenic source (Campbell et al., 2015; Lee and Brimblecombe, 2016), we have developed a new global gridded inventory of the primary emission sectors. The inventory has a 0.1° spatial resolution and uses the most current emission factors and industry activity data as input. The inventory is provided as annually varying estimates from years 1980–2012 and employs a source specific spatial scaling procedure.

2.1. Coal Coal combustion is traditionally considered the largest single direct industrial source of COS (Watts, 2000; Blake et al., 2008). We consider two classes of coal consumption in this study: industrial coal consumption and residential coal consumption for cooking and heating. Residential coal has not been included in previous global inventories, but recent work suggests potentially large emissions of COS from residential coal consumption in China (Du et al., 2016). 2.2. Industrial coal Country-level industry activity data of industrial coal consumption is readily available (Smith et al., 2011). While stack and plume observations for COS emission factors from coal power plants in the United States have been made (Khalil and Rasmussen, 1984; Blake et al., 2008), varying combustion efficiencies, emissions controls, and the sulfur content of coal in other countries suggests that the U.S. emission factors may not be appropriate to apply globally without modification. To address this limitation, previous work has used the ratio of SO2 emissions to coal consumption for each country to scale emissions factors (Campbell et al., 2015). This correction for estimating emissions factors is adopted in this study. For sub-country spatial scaling, industrial SO2 emissions are used to proportionally distribute the estimated country-level COS estimates (Joint Research Centre, 2011). Uncertainty for our OCS source estimate from industrial coal combustion is derived from the uncertainty in the emissions factor ( ± 18%).

2. Methods The inventory developed here includes direct and indirect anthropogenic sources. Indirect sources of COS result from anthropogenic carbon disulfide (CS2) emissions which are rapidly oxidized to COS in the atmosphere (Watts, 2000; Wang et al., 2001; Kettle et al., 2002). The atmospheric oxidation of CS2 to COS has been estimated to have a molar conversion rate of 87% (Barnes et al., 1994). Source estimates reported in this study are the sum of direct and indirect sources. Uncertainty estimates are based on the range of emission factor data within each sector as noted below. For sectors with only a single reported emission factor, we apply an uncertainty of 50% as in Watts (2000). The anthropogenic sources of COS considered in this study are: rayon production, aluminum smelting, carbon black production, industrial and residential coal consumption, agricultural chemicals, pulp & paper industries, industrial solvent applications, titanium dioxide production and tire wear (Watts, 2000; Campbell et al., 2015; Du et al., 2016; Lee and Brimblecombe, 2016). Industrial coal consumption includes manufacturing and electricity production. We did not include biomass burning (e.g. open burning, agriculture waste, biofuel) but we

2.3. Residential coal Emission factors for residential coal consumption have been made in laboratory coal stove experiments and household air-sampling in China (Du et al., 2016). The observed emission factors were 0.57 ± 0.10 and 1.43 ± 0.32 g COS emitted per kg of coal for the laboratory and farmhouse studies, respectively (approximately 50 times larger than power plant emission factors), which encompass the range of uncertainty in our estimate for COS from this source (Du et al., 2016). Here, an average of the laboratory and farmhouse emission factors is assumed for the baseline scenario and the range of the two emission factors is the estimated uncertainty. Country-level residential 12

Global estimates of carbon black production for select years k,l,m n,o,p,q; global production sharel,m Country-level productiont Country-level productiont,u Country-level cellulosic fiber productionw Country-level cellulosic fiber productionw

Carbon Black

Chemical industry NH3 f Manufacturing industry CO2f Industrial N2Of Industrial N2Of

14.7 g COS kg−1 titanium dioxide producedr Emissions scaled from previous work by industry activity datav 0.25 g CS2 g−1 yarnx 0.12 g CS2 g−1 staple (year ≤ 2000)x; 0.07 g CS2 g−1 staple (year > 2000)y 1.17 kg rubber y−1 vehicle−1, 43% COS, 57% CS2ab; 1.6% rubber released as sulfurac

b

Transportation CO2f

Chemical industry NH3g

Energy industry and waste incinerator SO2f Energy for buildings SO2f Industrial N2Of

Industrial CO2

10 g OCS kg−1 carbon black produced; 30 g CS2 kg−1 carbon black produced; 99% removal from developed countriesr

0.47–1.75 g COS kg−1 coal consumedj solvent CS2: 80% is emitted to atmosphere,b

0.6 kg COS/Mg Al for prebake; 1.2 for Söderberge 4.7–6.8 μmol COS mol CO2 −1 h; Scaled by SO2 emissions

(Blagoev and Funada, 2011). (Chin and Davis, 1993). c (Ramankutty et al., 2008). d (U.S. Geological Survey, 2012). e (Kimmerle and Noel, 1997). f (Joint Research Centre, 2011). g (Smith et al., 2011). h (Blake et al., 2008). i (United Nations Statistics Division, 2016a). j (Du et al., 2016). k (IARC - World Health Organization, 1984). l (IARC - World Health Organization, 2010). m (Büchel et al., 2000). n (Crump, 2000). o (Gandhi, 2005). p (ICBA, 2016a). q (Lee and Brimblecombe, 2016). r (Blake et al., 2004). s (U.S. Geological Survey, 2015). t (Food and Agriculture Organization, 2010). u (United Nations Statistics Division, 2016b). v (EPA Office of Air Quality Planning and Standards and EPA Office of Air and Radiation, 2012). w (Fiber Economics Bureau, 2014). x (U.S. Environmental Protection Agency, 2003). y (Dodd et al., 2013). z (Dargay et al., 2007). aa (OICA, 2016). ab (Pos and Berresheim, 1993). ac (Susa and Haydary, 2013). ad (The World Bank, 2016).

a

Tires

Cars in use (years ≥ 2004)z; Cars 1000 population−1 (yeas < 2004)z,aa

Country-level residential coal consumptioni Country-level consumption of CS2 (CEH other category)a

Residential Coal Industrial Solvents

Titanium Dioxide Pulp & Paper Rayon Yarn Rayon Staple

Country-level coal consumption

g

Industrial Coal

Aluminum Smelting

Agriculture landc

Proportion of CS2 emitted to atmosphere: 80%b

Country-level consumption of CS2 (Chemical Economics Handbook (CEH) agricultural chemicals category)a Country-level primary aluminum productiond

Agricultural Chemicals f

Sub-Country Spatial Scaling

Emissions Factor

Industry Activity

Source

Table 1 Input data for the anthropogenic sources of COS considered in this study.

productions productiont,u cellulosic fiber productionw cellulosic fiber productionw Yearly cars in use (years ≥ 2004)aa; Product of population by vehicle ownership (yeas < 2004)z,aa,ad

Yearly Yearly Yearly Yearly

Yearly residential coal consumptioni Yearly consumption of CS2 and average growth rates of CS2 consumptiona Linear interpolation of industry activity data

Yearly consumption of CS2 and average growth rates of CS2 consumptiona Yearly country-level aluminum primary productiond Yearly country level-coal SO2 emissionsg

Temporal Scaling

A. Zumkehr et al.

Atmospheric Environment 183 (2018) 11–19

13

Atmospheric Environment 183 (2018) 11–19

A. Zumkehr et al.

2.7. Pulp and paper industry

coal consumption was obtained from the United Nations Statistics Division Energy Statistics Database (United Nations Statistics Division, 2016a). Similar to the industrial coal source described in the previous section, residential SO2 is used as the spatial proxy for sub-country scaling (Joint Research Centre, 2011).

The pulp and paper industry is another potentially important source of anthropogenic COS that has not been included in gridded estimates of global anthropogenic COS emissions in the past (Kettle et al., 2002). Yearly, country-level industry activity data is available from the United Nations statistics database (UN Stats) up to the year 2007 and after 2007 from a separate UN industry capacity report (Food and Agriculture Organization, 2010; United Nations Statistics Division, 2016b). Sub-country spatial scaling is performed using CO2 from combustion in the manufacturing industry because the IPCC code describing this section of the EDGAR v4.2 data set includes paper production (Joint Research Centre, 2011). We estimate emission factors of 1.7 g Mg−1 and 0.06 g Mg−1 for CS2 and COS based on U.S. emission estimates of 90 Mg y−1 and 3 Mg y−1 for CS2 and COS and 51.6 million tonnes of pulp and paper production (EPA Office of Air Quality Planning and Standards and EPA Office of Air and Radiation, 2012). We are not aware of uncertainty estimates for the emission factor data.

2.4. Aluminum Aluminum smelting is a significant source of atmospheric COS (Harnisch et al., 1995; Watts, 2000). There are two main methods used for aluminum smelting: Söderberg and prebake (Kimmerle and Noel, 1997). Annual, country-level primary production of aluminum is obtained from USGS reports (U.S. Geological Survey, 2012). To obtain the COS emitted per mass of aluminum produced, the emissions factors of 0.6 and 1.2 kg COS Mg Al−1 are used for prebake and Söderberg, respectively (Kimmerle and Noel, 1997), resulting in a 33% uncertainty range for this source of COS. Prebake is a more efficient smelting technique, resulting in fewer emissions of many chemical species, including COS. It is estimated that in the year 2012, 90% of the global aluminum production was produced through the prebake smelting process (Campbell et al., 2015). We use this ratio for the year 2012 and we assume that the lower rates of global production in earlier years are due to lower rates of prebake smelting to simulate the gradual adoption of the new prebake technology over time. The COS source estimated for each country is proportionally distributed over EDGAR v4.2 industrial CO2 emissions for the appropriate year (Joint Research Centre, 2011).

2.8. Rayon production Emissions of CS2 during rayon production are thought to be the largest anthropogenic COS source (Campbell et al., 2015). Rayon production can be divided into two categories, yarn and staple, each having different emissions factors, spatial distributions, and magnitudes of production. Annual, country-level production is regularly reported in the industry journal Fiber Organon (Fiber Economics Bureau, 2014). Because N20 is also produced during rayon production, we use industrial N2O estimates for sub-country scaling of both rayon yarn and staple (Joint Research Centre, 2011; Gullingsrud, 2017). Yarn rayon has an emissions factor of 0.25 g CS2 per gram of yarn produced, while staple rayon production has an emissions factor of 0.12 g CS2 per gram of staple material produced (U.S. Environmental Protection Agency, 2003; Campbell et al., 2015). In recent years, improved technology for the production of staple rayon has reduced the emissions factor to 0.07 g CS2 per gram of staple produced; however, the yarn emissions factor remains unchanged (Dodd et al., 2013; Campbell et al., 2015). Emissions factor uncertainty has not been reported. Here, the newer emissions factor for staple rayon production is used for added production capacity occurring after the year 2000 and the older emissions factor is applied for all other capacity. Drops in capacity are subtracted from production using the older emissions factor because it is more likely that factories with older technology would be taken offline in times of reduced demand because they have higher CS2 input costs relative to the newer factories. Likewise, if the production capacity is replenished after a drop in demand, the replenished capacity is added back with the older emissions factor to simulate older factories that were not in use during times of lower rayon demand and are brought back online after the market recovered.

2.5. Agricultural chemicals and industrial solvents Agricultural chemicals and industrial solvents derived from CS2 are an indirect source of COS. These agricultural chemicals are commonly used as insecticides and fumigants (Hinds, 1902). A component of CS2 used to form agricultural chemicals is released to the atmosphere and then oxidized to COS (Chin and Davis, 1993; Wang et al., 2001). Approximately 80% of the CS2 used as industrial solvents and applied to croplands is emitted to the atmosphere and no uncertainty on this emission factor has been reported (Chin and Davis, 1993). We convert our country-level emissions into a sub-country gridded map using maps of croplands as a spatial proxy (Ramankutty and Foley, 1999) for agricultural chemicals. Annual, country-level data for CS2 agricultural chemical and industrial use is available in the Chemical Economics Handbook (CEH) report (Blagoev and Funada, 2011).

2.6. Pigment industry There are two pigments that show potential for being significant sources of anthropogenic COS: carbon black and titanium dioxide (TiO2) (Blake et al., 2004; Lee and Brimblecombe, 2016). Global, annual carbon black production estimates were obtained from multiple studies (IARC - World Health Organization, 1984; Büchel et al., 2000; Crump, 2000; Gandhi, 2005; ICBA, 2016b). Years with no reported global carbon black production were approximated through linear interpolation. Country-level production data were obtained for TiO2 from the U.S. Geological Survey (2015) and carbon black production was obtained for years 2000 (Büchel et al., 2000) and 2005 (IARC - World Health Organization, 2010). Emissions factors for carbon black are 10 g COS kg−1 and 30 g CS2 kg−1 (U.S. Environmental Protection Agency, 2003). The emissions factor for TiO2 is 14.7 g COS kg−1 (no CS2 emissions) (Blake et al., 2004). Uncertainty estimates for these emission factors have not been reported. A reduction in the emission factor of 99% is applied to developed countries due to emissions controls (Blake et al., 2004). Subcountry level scaling is based on industry NH3 emissions (Joint Research Centre, 2011).

2.9. Tires Automotive tire wear also emits COS and CS2 (Pos and Berresheim, 1993; Lee and Brimblecombe, 2016). Data for the approximate number of cars in use is available after the year 2004 (OICA, 2016) and for the years prior to 2004, country-level ownership rates (cars per 1000 population) are available. The latter is multiplied by yearly population data (thousands in population) to estimate total cars in use for earlier years (Dargay et al., 2007; The World Bank, 2016). Transportation CO2 emissions from EDGAR v4.2 are used as the sub-country spatial proxy (Joint Research Centre, 2011). An emissions rate per vehicle per year is calculated by assuming that the average automobile deposits 1.17 kg of tire rubber to the road each year and that the sulfur within the rubber (accounting for 1.6% of the rubber in the tires) is eventually released as either COS or CS2 (57% as CS2, 43% as COS) (Chin and Davis, 1993; Pos and Berresheim, 1993; 14

Atmospheric Environment 183 (2018) 11–19

A. Zumkehr et al.

Susa and Haydary, 2013). Emission factor uncertainty is not available. 3. Results and discussion Total anthropogenic sources (indirect and direct) for the gridded inventories from Kettle and our inventory show large differences in both spatial distribution and total magnitude (Fig. 1). The anthropogenic source from the Kettle inventory is relatively evenly distributed between Asia, North America, and Europe (Fig. 1a). However, our inventory shows a high concentration of emissions in Asia. We find that 45% of the global source is emitted by China and the remaining source is relatively evenly distributed between India, North America, and Europe. The anthropogenic source for our inventory relative to the Kettle inventory is 7 and 10 times larger in China and India, respectively. While earlier years in our inventory have lower emissions in Asia, there are no years in our inventory that resemble the dramatically lower Asian emissions suggested by the Kettle inventory. The large differences between the spatial distribution in these two inventories is primarily due to the use of industry activity data. The spatial distribution in the Kettle inventory is entirely based on industrial SO2 emissions from the 1980's. However, the spatial distribution in our inventory is based on annual, country-level industry activity data for each COS source sector. This spatial distribution of anthropogenic sources is supportive of the most critical applications of the COS tracer for inferring GPP. Low anthropogenic activity in the tropics is significant because this region is a hot spot for GPP uncertainty and a recently observed depression in atmospheric COS over the Amazon is thought to be closely related to GPP (Beer et al., 2010; Glatthor et al., 2015). The relatively low anthropogenic source in North America is supportive of GPP inference in this region where air-monitoring sites are concentrated and current COS applications have shown similar results to solar induce fluorescence (Hilton et al., 2017). Our inventory shows long-term trends that are not accounted for in the climatological Kettle inventory (Fig. 2). The most prominent trends are a decline in the global source in the 1990's, followed by growth with a temporary decline during the global economic recession in 2008 and 2009. While previous work noted a decline in the rayon source in the

Fig. 2. Time series of global anthropogenic COS sources, including direct and indirect emissions. The dashed black line represents the climatological estimate from the Kettle inventory. The pulp & paper source is omitted from this figure due to the insignificance of the estimated source.

1990's (Campbell et al., 2015), here we also find that residential coal declined during the 1990's. The decreasing residential coal trend was not identified previously because emission factors for residential coal combustion were only recently measured and they have not been used in previous global inventories (Du et al., 2016). During the 1990's the residential coal and rayon sources decreased by 32 and 40 Gg S y−1, respectively. However, it should be noted that the estimated decline in COS emissions from residential coal is based entirely on a decline in residential coal consumption and not on a change in the emission factor. Future research will be needed to consider the potential for a change in the emission factor, particularly for anthracite coal that may have been dominant before the 1990's (Du et al., 2016). The subsequent growth in the anthropogenic COS source in the 2000's is primarily due to increases in the rayon staple and industrial coal sectors of 11 and 36 Gg S y−1, respectively. These long-term trends are the result of different regional changes (Fig. 3). The 1990's decline is closely associated with the decline in the European source. The 1990's decline in the residential coal source in China is offset by the growth in other source sectors in China during the 1990's, such as rayon. The subsequent growth in the 2000's is driven by rapid increases in the Chinese rayon industry. The long-term trends in the anthropogenic inventories are

Fig. 1. Comparison of the global gridded anthropogenic COS flux from the climatological Kettle inventory and this study for year 2012. These source inventories include direct emissions from COS and indirect sources from anthropogenic CS2 emissions.

Fig. 3. Time series of the total anthropogenic source of COS (Gg S y−1) from the highest contributing regions.

15

Atmospheric Environment 183 (2018) 11–19

A. Zumkehr et al.

primarily due to the addition of the domestic coal and pigment industry sectors. Our estimate is 30% smaller than the Lee and Brimblecombe inventory, primarily due to differences in source estimates for the pulp and paper sector. Lee and Brimblecombe estimated the pulp and paper sector using published point observations of furnace concentrations which required assumptions about global furnace volume and the relationship between point furnace concentrations and stack emissions. We used a more direct approach using U.S. EPA stack emissions data and global pulp and paper production activity. For each source sector, Fig. 5 displays the countries that are the top contributors for the year 2012. China dominates, while the U.S. and India are responsible for the next largest sources. Russia, Germany and Canada are also significant contributors to global anthropogenic COS sources, though these countries' emissions are declining as is the case with many developed countries. Fig. 6 shows the regional anthropogenic COS sources by sector for the year 2012. In all regions, except for the U.S. and Canada, the rayon sources (yarn and staple) are dominant. In the U.S., the rayon industry went into decline, leaving pigment industries to dominate the anthropogenic COS emissions in this region. While we estimate China to have several large sources, it is specifically the rayon (yarn and staple) and coal (industry and residential) sources that make China by far the dominant global contributor of anthropogenic COS. India dominates the agricultural chemicals source and Europe has a slight lead over China for the aluminum source. While this study proposes that the sources presented here are a more contemporary and complete picture of anthropogenic COS sources, uncertainties persist. The emissions factors for the various components of the anthropogenic source of COS are the largest source of uncertainty for the magnitudes of emissions presented in this study and thus, emissions factors research presents a valuable opportunity for future work. Uncertainty estimates for indirect sources from industrial CS2 emissions are also caused by the estimates of the atmospheric oxidation rate which is assumed to have an uncertainty of 50% (Chin and Davis, 1993; Launois et al., 2015). These uncertainty ranges are reflected in Figs. 4 and 6. This study suggests that anthropogenic sources of COS could account for a portion of the missing source of COS in the global budget. While top-down studies of remote sensing data have been used to provide evidence for a missing source in the oceans, we report some qualitative consistency in the remote sensing data with the large anthropogenic source presented here. High concentrations of COS near Asia have been observed by the NASA Tropospheric Emissions Spectrometer (TES) (Fig. 7a) (Kuai et al., 2015). Similarly, the upward revision of COS anthropogenic sources is largely driven by growth in emission in China (Fig. 7b). Future work could apply an atmospheric transport model to test whether these inventories and quantitatively consistent with the remote sensing trends.

Fig. 4. A comparison of annual global anthropogenic sources of COS (Gg S y−1) from the Kettle inventory and from this study for the year 2012. Fluxes reporting zero for Kettle on the right-hand side of the red dotted line were not considered in the Kettle inventory. When referring to this study, the coal category includes combined residential and industrial sources; pigments include carbon black and titanium dioxide and rayon includes yarn and staple. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

qualitatively consistent with observed atmospheric COS trends. The decline in the 1990's source is qualitatively consistent with the decline observed in Antarctic firn and remote sensing data during this period (Montzka et al., 2004; Campbell et al., 2017). The increasing source after the year 2000 and temporary decline during the global economic recession is also consistent with atmospheric data. Although a previous report of NOAA air-monitoring data for years 2000 through 2005 showed no trend (Montzka et al., 2007), an updated presentation of the NOAA data from the same sites but extended to include the years 2000 through 2015 shows a generally increasing trend and a temporary decline during 2009 (Campbell et al., 2017). Similar trends have also been observed in remote sensing data (Toon et al., 2018; Kremser et al., 2015; Lejeune et al., 2017). Our baseline total emission estimate is 406 Gg S y−1 (lower-estimate: 223 Gg S y−1, upper-estimate: 586 Gg S y−1) which is roughly double the Kettle estimate of 180 Gg S y−1 (Fig. 4). The upward revision is due to the inclusion of additional source sectors, the use of more recent anthropogenic activity data, and in some sectors the use of more representative emission factor data. While both studies accounted for coal, our estimate is 6 times larger due to the inclusion of residential coal, the use of current coal consumption levels, and the integration of plume study data showing that the Kettle emission factor for industrial coal was underestimated by a factor of 2.5 (Blake et al., 2008; Campbell et al., 2015; Du et al., 2016). The rayon estimates for both studies are surprisingly similar, but only for the most recent years of our inventory. In the 1990's when the rayon industry was in decline, our rayon source estimate was considerably smaller than the Kettle estimate. While our inventory includes several sectors that were not accounted for in the Kettle inventory, the most important is the pigment industry with an estimated source of 76 Gg S y−1. Our bottom-up estimate of the anthropogenic source is consistent with the top-down estimates of Campbell et al. (2017). In the top down analysis, atmospheric COS data were used to estimate the global COS sources and sinks using a constrained optimization approach. The constraint for the anthropogenic source was the previously reported bottom-up range of 150–360 Gg S y. The top-down optimal value was found to be at the high end of the bottom-up range (Table S3 in Campbell et al., 2017), which is supportive of the upward revision presented in our bottom-up inventory. We also compare our source estimates to the non-gridded inventories from Campbell et al. (2015) and Lee and Brimblecombe (2016). Our estimates are 60% larger than the Campbell inventory,

4. Conclusions Previous work extrapolated global estimates of COS emissions over SO2 grids to provide a climatological inventory that became the standard for gridded anthropogenic COS sources (Kettle et al., 2002). However, this inventory was created over a decade ago, incorporates industry activity data from over three decades ago and uses outdated emissions factors. Recent work suggests that this inventory is not representative of current conditions (Campbell et al., 2015). Here we present an updated inventory, providing emission estimates with a higher spatial resolution, more sophisticated spatial scaling, including more component sources and more years of historical estimates, while using the most current industry activity data and emissions factors as input. While uncertainty still remains, the level of uncertainty is likely within the functional range for regional-scale modeling efforts.

16

Atmospheric Environment 183 (2018) 11–19

A. Zumkehr et al.

Fig. 5. Fig. 5. The share of global anthropogenic COS sources from the highest contributing countries by source for 2012. “Rest of World” data includes the contribution of all countries excluding the five countries listed in each legend. 17

Atmospheric Environment 183 (2018) 11–19

A. Zumkehr et al.

Blake, N.J., Streets, D.G., Woo, J.-H., Simpson, I.J., Green, J., 2004. Carbonyl sulfide and carbon disulfide: large-scale distributions over the western Pacific and emissions from Asia during TRACE-P. J. Geophys. Res. 109 (D15), D15S05. http://dx.doi.org/10. 1029/2003JD004259. Blake, N.J., Campbell, J.E., Vay, S.A., Fuelberg, H.E., Huey, L.G., Sachse, G., Meinardi, S., Beyersdorf, A., Baker, A., Barletta, B., Midyett, J., Doezema, L., Kamboures, M., McAdams, J., Novak, B., Rowland, F.S., Blake, D.R., 2008. Carbonyl sulfide (OCS): large-scale distributions over North America during INTEX-NA and relationship to CO2. J. Geophys. Res. 113 (D9). http://dx.doi.org/10.1029/2007JD009163. D09S90. Büchel, K.H., Moretto, H.-H., Werner, D., Woditsch, P., 2000. Industrial Inorganic Chemistry, 2 edition. Wiley-VCH, Weinheim, Germany. Campbell, J.E., Carmichael, G.R., Chai, T., Mena-Carrasco, M., Tang, Y., Blake, D.R., Blake, N.J., Vay, S.A., Collatz, G.J., Baker, I., Berry, J.A., Montzka, S.A., Sweeney, C., Schnoor, J.L., Stanier, C.O., 2008. Photosynthetic control of atmospheric carbonyl sulfide during the growing season. Science (80-. ). 322 (5904), 1085–1088. http://dx. doi.org/10.1126/science.1164015. Campbell, J.E., Whelan, M.E., Seibt, U., Smith, S.J., Berry, J.A., Hilton, T.W., 2015. Atmospheric carbonyl sulfide sources from anthropogenic activity: implications for carbon cycle constraints. Geophys. Res. Lett. 42 (8), 3004–3010. http://dx.doi.org/ 10.1002/2015GL063445. Campbell, J.E., Berry, J.A., Seibt, U., Smith, S.J., Montzka, S.A., Launois, T., Belviso, S., Bopp, L., Laine, M., 2017. Large historical growth in global terrestrial gross primary production. Nature 544 (7648), 84–87. http://dx.doi.org/10.1038/nature22030. Carmichael, G.R., 2003. Evaluating regional emission estimates using the TRACE-P observations. J. Geophys. Res. 108 (D21), 8810. http://dx.doi.org/10.1029/ 2002JD003116. Chin, M., Davis, D.D., 1993. Global sources and sinks of OCS and CS2 and their distributions. Global Biogeochem. Cycles 7 (2), 321–337. http://dx.doi.org/10.1029/ 93GB00568. Commane, R., Meredith, L.K., Baker, I.T., Berry, J.A., Munger, J.W., Montzka, S.A., Templer, P.H., Juice, S.M., Zahniser, M.S., Wofsy, S.C., 2015. Seasonal fluxes of carbonyl sulfide in a midlatitude forest. Proc. Natl. Acad. Sci. U.S.A. 112 (46), 14162–14167. http://dx.doi.org/10.1073/pnas.1504131112. Cox, P.M., Betts, R.A., Jones, C.D., Spall, S.A., Totterdell, I.J., 2000. Acceleration of global warming due to carbon-cycle feedbacks in a coupled climate model. Nature 408 (6809), 184–187. Crump, E.L., 2000. Economic Impact Analysis for the Proposed Carbon Black Manufacturing NESHAP. Dargay, J., Gately, D., Sommer, M., 2007. Vehicle ownership and income growth, worldwide: 1960-2030. Energy J. 28 (4), 143–170. http://dx.doi.org/10.2307/ 41323125. Dodd, N., Cordella, M., Wolf, O., Waidløw, J., Stibold, M., Hansen, E., 2013. Revision of the European Ecolabel and Green Public Procurement Criteria for Textile Products, Sevilla, Spain. Du, Q., Zhang, C., Mu, Y., Cheng, Y., Zhang, Y., Liu, C., Song, M., Tian, D., Liu, P., Liu, J., Xue, C., Ye, C., 2016. An important missing source of atmospheric carbonyl sulfide: domestic coal combustion. Geophys. Res. Lett. 43 (16), 8720–8727. http://dx.doi. org/10.1002/2016GL070075. EPA Office of Air Quality Planning and Standards, and EPA Office of Air and Radiation, 2012. Residual Risk Assessment for the Pulp & Paper Source Category. Fiber Economics Bureau, 2014. Fiber Organon 94–116. Available from: http://www. fibereconomics.com/feb3c.htm. Field, C.B., Lobell, D.B., Peters, H.A., Chiariello, N.R., 2007. Feedbacks of terrestrial ecosystems to climate change. Annu. Rev. Environ. Resour. 32 (1), 1–29. http://dx. doi.org/10.1146/annurev.energy.32.053006.141119. Food and Agriculture Organization, 2010. Pulp and Paper Capacities, Rome. ISSN 0255–7665. Friedlingstein, P., Cox, P., Betts, R., Bopp, L., von Bloh, W., Brovkin, V., Cadule, P., Doney, S., Eby, M., Fung, I., Bala, G., John, J., Jones, C., Joos, F., Kato, T., Kawamiya, M., Knorr, W., Lindsay, K., Matthews, H.D., Raddatz, T., Rayner, P., Reick, C., Roeckner, E., Schnitzler, K.G., Schnur, R., Strassmann, K., Weaver, A.J., Yoshikawa, C., Zeng, N., 2006. Climate-carbon cycle feedback analysis: results from the C4MIP model intercomparison. J. Clim. 19 (14), 3337–3353. http://dx.doi.org/10.1175/ JCLI3800.1. Gandhi, B., 2005. Reassessment of One Exemption from the Requirement of a Tolerance for Carbon Black, Washington D.C.Reassessment of One Exemption from the Requirement of a Tolerance for Carbon Black, Washington D.C. Glatthor, N., Höpfner, M., Baker, I.T., Berry, J., Campbell, J.E., Kawa, S.R., Krysztofiak, G., Leyser, A., Sinnhuber, B.-M., Stiller, G.P., Stinecipher, J., von Clarmann, T., 2015. Tropical sources and sinks of carbonyl sulfide observed from space. Geophys. Res. Lett. 42 (22), 10,082–10,090. http://dx.doi.org/10.1002/2015GL066293. Gullingsrud, A., 2017. Fashion Fibers + Studio Access. Fairchild Books. Gurney, K.R., Law, R.M., Denning, A.S., Rayner, P.J., Baker, D., Bousquet, P., Bruhwiler, L., Chen, Y., Ciais, P., Fan, S., Fung, I.Y., Gloor, M., Heimann, M., Higuchi, K., John, J., Maki, T., Maksyutov, S., Masarie, K., Peylin, P., Prather, M., Pak, B.C., Randerson, J., Sarmiento, J., Taguchi, S., Takahashi, T., Yuen, C., 2002. Towards robust regional estimates of CO2 sources and sinks using atmospheric transport models. Nature 415 (6872), 626–630. http://dx.doi.org/10.1038/415626a. Harnisch, J., Borchers, R., Fabian, P., Kourtidis, K., 1995. Aluminium production as a source of atmospheric carbonyl sulfide (COS). Environ. Sci. Pollut. Res. 2 (3), 161–162. http://dx.doi.org/10.1007/BF02987529. Hilton, T.W., Whelan, M.E., Zumkehr, A., Kulkarni, S., Berry, J.A., Baker, I.T., Montzka, S.A., Sweeney, C., Miller, B.R., Campbell, J.E., 2017. Peak growing season gross uptake of carbon in North America is largest in the Midwest USA. Nat. Clim. Change (May). http://dx.doi.org/10.1038/NCLIMATE3272.

Fig. 6. Individual sources of anthropogenic COS (Gg S y-1) by region. The regions of China, Europe, U.S. & Canada and India are shown here because they are dominant source regions of atmospheric COS from anthropogenic activity. The pulp & paper source is omitted from this figure due to the insignificance of the estimated source.

Fig. 7. Free troposphere COS (ppt) for Tropospheric Emissions Spectrometer measurements in June 2006 (a) Portion of the figure obtained from: Kuai et al. (2015), and total estimated anthropogenic COS emissions for 2006 from this study (b).

References Barnes, I., Becker, K.H., Patroescu, I., 1994. The tropospheric oxidation of dimethyl sulfide: a new source of carbonyl sulfide. Geophys. Res. Lett. 21 (22), 2389–2392. http://dx.doi.org/10.1029/94GL02499. Beer, C., Reichstein, M., Tomelleri, E., Ciais, P., Jung, M., Carvalhais, N., Rodenbeck, C., Arain, M.A., Baldocchi, D., Bonan, G.B., Bondeau, A., Cescatti, A., Lasslop, G., Lindroth, A., Lomas, M., Luyssaert, S., Margolis, H., Oleson, K.W., Roupsard, O., Veenendaal, E., Viovy, N., Williams, C., Woodward, F.I., Papale, D., 2010. Terrestrial gross carbon dioxide uptake: global distribution and covariation with climate. Science 329 (5993), 834–838. http://dx.doi.org/10.1126/science.1184984. Berry, J., Wolf, A., Campbell, J.E., Baker, I., Blake, N., Blake, D., Denning, A.S., Kawa, S.R., Montzka, S.A., Seibt, U., Stimler, K., Yakir, D., Zhu, Z., 2013. A coupled model of the global cycles of carbonyl sulfide and CO2 : a possible new window on the carbon cycle. J. Geophys. Res. Biogeosci. 118 (2), 842–852. http://dx.doi.org/10.1002/jgrg. 20068. Blagoev, M., Funada, C., 2011. Marketing Research Report: Carbon Disulfide, CEH, (625.5000C). pp. 1–65. Available from: https://www.ihs.com/products/chemicaleconomics-handbooks.html.

18

Atmospheric Environment 183 (2018) 11–19

A. Zumkehr et al.

org/10.1029/2006JD007665. D09302. OICA, 2016. Vehicles in Use, OICA. Available from: http://www.oica.net/category/ vehicles-in-use/, Accessed date: 9 September 2016. Pos, W.H., Berresheim, H., 1993. Automotive tire wear as a source for atmospheric OCS and CS2. Geophys. Res. Lett. 20 (9), 815–817. http://dx.doi.org/10.1029/ 93GL00972. Ramankutty, N., Foley, J.A., 1999. Estimating historical changes in land cover: North american croplands from 1850 to 1992. Global Ecol. Biogeogr. 8 (5), 381–396. Ramankutty, N., Evan, A.T., Monfreda, C., Foley, J.A., 2008. Farming the planet: 1. Geographic distribution of global agricultural lands in the year 2000. Global Biogeochem. Cycles 22, 19. Sandoval-Soto, L., Stanimirov, M., von Hobe, M., Schmitt, V., Valdes, J., Wild, A., Kesselmeier, J., 2005. Global uptake of carbonyl sulfide (COS) by terrestrial vegetation: estimates corrected by deposition velocities normalized to the uptake of carbon dioxide (CO2). Biogeosciences 2 (2), 125–132. Smith, S.J., Van Aardenne, J., Klimont, Z., Andres, R.J., Volke, A., Arias, S.D., 2011. Anthropogenic Sulfur Dioxide Emissions : 1850 – 2005. pp. 1101–1116. http://dx. doi.org/10.5194/acp-11-1101-2011. Suntharalingam, P., Kettle, A.J., Montzka, S.M., Jacob, D.J., 2008. Global 3-D model analysis of the seasonal cycle of atmospheric carbonyl sulfide: implications for terrestrial vegetation uptake. Geophys. Res. Lett. 35 (19), L19801. http://dx.doi.org/10. 1029/2008GL034332. Susa, D., Haydary, J., 2013. Sulphur distribution in the products of waste tire pyrolysis. Chem. Pap. 67 (12), 1521–1526. http://dx.doi.org/10.2478/s11696-012-0294-4. The World Bank, 2016. Population, Total, World Bank Open Data. Available from: http://data.worldbank.org/, Accessed date: 9 September 2016. Toon, Geoffrey C., Jean-Francois, L. Blavier, Sung, Keeyoon, 2018. Atmospheric carbonyl sulfide (OCS) measured remotely by FTIR solar absorption spectrometry. Atmos. Chem. Phys. 18.3, 1923–1944. U.S. Environmental Protection Agency, 2003. Compilation of Air Pollutant Emission Factors, Stationary Point and Area Sources, AP-42. (Washington, D.C.). U.S. Geological Survey, 2012. USGS Minerals Information: Aluminum. Available from: http://minerals.usgs.gov/minerals/pubs/commodity/aluminum/. U.S. Geological Survey, 2015. Mineral Commodity Summaries: Titanium Dioxide. Available from: https://minerals.usgs.gov/minerals/pubs/commodity/titanium/. United Nations Statistics Division, 2016a. Energy Statistics Database. Available from: http://unstats.un.org/unsd/energy/edbase.htm, Accessed date: 1 January 2017. United Nations Statistics Division, 2016b. Industry Statistics Database. Available from: http://unstats.un.org/unsd/industry//ics_intro.asp, Accessed date: 1 January 2016. Wang, L., Zhang, F., Chen, J., 2001. Carbonyl sulfide derived from catalytic oxidation of carbon disulfide over atmospheric particles. Environ. Sci. Technol. 35 (12), 2543–2547. Wang, Y., Deutscher, N.M., Palm, M., Warneke, T., Notholt, J., Baker, I., Berry, J., Suntharalingam, P., Jones, N., Mahieu, E., Lejeune, B., Hannigan, J., Conway, S., Mendonca, J., Strong, K., Campbell, J.E., Wolf, A., Kremser, S., 2016. Towards understanding the variability in biospheric CO2 fluxes: using FTIR spectrometry and a chemical transport model to investigate the sources and sinks of carbonyl sulfide and its link to CO2. Atmos. Chem. Phys. 16 (4), 2123–2138. http://dx.doi.org/10.5194/ acp-16-2123-2016. Watts, S.F., 2000. The mass budgets of carbonyl sulfide, dimethyl sulfide, carbon disulfide and hydrogen sulfide. Atmos. Environ. 34 (5), 761–779. http://dx.doi.org/10.1016/ S1352-2310(99)00342-8. Whelan, M.E., Hilton, T.W., Berry, J.A., Berkelhammer, M., Desai, A.R., Campbell, J.E., 2016. Carbonyl sulfide exchange in soils for better estimates of ecosystem carbon uptake. Atmos. Chem. Phys. 16 (6), 3711–3726. http://dx.doi.org/10.5194/acp-163711-2016. Zumkehr, A., Hilton, T.W., Whelan, M., Smith, S., Campbell, J.E., 2017. Gridded anthropogenic emissions inventory and atmospheric transport of carbonyl sulfide in the U.S. J. Geophys. Res. Atmos. http://dx.doi.org/10.1002/2016JD025550.

Hinds, W.E., 1902. Carbon bisulphid as an insecticide. In: Farmers' Bulletin, vol. 145. U.S. Department of Agriculture, Washington D. C., pp. 28. IARC - World Health Organization, 1984. Evaluation of the carcinogenic risk of chemicals to humans. IARC Monogr. 33. IARC - World Health Organization, 2010. Some aromatic amines, organic dyes, and related exposures. IARC Monogr. Eval. Carcinog. Risks Hum. 93, 9–38. http://dx.doi. org/10.1007/s10350-006-0552-z. ICBA, 2016a. Carbon black User's guide. Carbon Black User's Guide 36 Int. Carbon Black Assoc. Available from: http://www.carbon-black.org/images/docs/2016-ICBACarbon-Black-User-Guide.pdf. ICBA, 2016b. What Is Carbon Black? Available from: http://www.carbon-black.org/ index.php/what-is-carbon-black, Accessed date: 17 January 2017. Joint Research Centre, 2011. Global Emissions EDGAR v4.2. Available from: http:// edgar.jrc.ec.europa.eu. Kettle, A.J., Kuhn, U., von Hobe, M., Kesselmeier, J., Andreae, M.O., 2002. Global budget of atmospheric carbonyl sulfide: temporal and spatial variations of the dominant sources and sinks. J. Geophys. Res. 107, 4658. http://dx.doi.org/10.1029/ 2002JD002187. Khalil, M.A.K., Rasmussen, R.A., 1984. Global sources, lifetimes and mass balances of carbonyl sulfide (OCS) and carbon disulfide (CS2) in the earth's atmosphere. Atmos. Environ. 18 (9), 1805–1813. http://dx.doi.org/10.1016/0004-6981(84)90356-1. Kimmerle, F., Noel, L., 1997. COS, CS2 and SO2 emissions from prebake Hall Heroult cells. Arvida Reasearch Dev. Cent. 153–158. Kremser, Stefanie, et al., 2015. Positive trends in Southern Hemisphere carbonyl sulfide. Geophy. Res. Lett. 42.21, 9473–9480. Kuai, L., Worden, J., Kulawik, S.S., Montzka, S.A., Liu, J., 2014. Characterization of Aura TES carbonyl sulfide retrievals over ocean. Atmos. Meas. Tech. 7 (1), 163–172. http://dx.doi.org/10.5194/amt-7-163-2014. Kuai, L., Worden, J.R., Campbell, J.E., Kulawik, S.S., Li, K., Lee, M., Weidner, R.J., Montzka, S.A., Moore, F.L., Berry, J.A., Baker, I., Denning, A.S., Bian, H., Bowman, K.W., Liu, J., Yung, Y.L., 2015. Estimate of carbonyl sulfide tropical oceanic surface fluxes using Aura Tropospheric Emission Spectrometer observations. J. Geophys. Res. Atmos. 120 (20), 11,012–11,023. http://dx.doi.org/10.1002/2015JD023493. Launois, T., Peylin, P., Belviso, S., Poulter, B., 2014. A new model of the global biogeochemical cycle of carbonyl sulfide – Part 2: use of OCS to constrain gross primary productivity of current vegetation models. Atmos. Chem. Phys. Discuss. 14 (20), 27663–27729. http://dx.doi.org/10.5194/acpd-14-27663-2014. Launois, T., Belviso, S., Bopp, L., Fichot, C.G., Peylin, P., 2015. A new model for the global biogeochemical cycle of carbonyl sulfide – Part 1: assessment of direct marine emissions with an oceanic general circulation and biogeochemistry model. Atmos. Chem. Phys. 15 (5), 2295–2312. http://dx.doi.org/10.5194/acp-15-2295-2015. Lee, C.L., Brimblecombe, P., 2016. Anthropogenic contributions to global carbonyl sulfide, carbon disulfide and organosulfides fluxes. Earth Sci. Rev. 160, 1–18. http://dx. doi.org/10.1016/j.earscirev.2016.06.005. Lejeune, Bernard, et al., 2017. Optimized approach to retrieve information on atmospheric carbonyl sulfide (OCS) above the Jungfraujoch station and change in its abundance since 1995. J. Quant. Spectrosc. Radiat. Transf. 186, 81–95. Lennartz, S.T., Marandino, C.A., von Hobe, M., Cortes, P., Quack, B., Simo, R., Booge, D., Pozzer, A., Steinhoff, T., Arevalo-Martinez, D.L., Kloss, C., Bracher, A., Röttgers, R., Atlas, E., Krüger, K., 2016. Oceanic emissions unlikely to account for the missing source of atmospheric carbonyl sulfide. Atmos. Chem. Phys. Discuss. 1–23. http://dx. doi.org/10.5194/acp-2016-778. mixi (September). Montzka, S.A., Aydin, M., Battle, M., Butler, J.H., Saltzman, E.S., Hall, B.D., Clarke, A.D., Mondeel, D., Elkins, J.W., 2004. A 350-year atmospheric history for carbonyl sulfide inferred from Antarctic firn air and air trapped in ice. J. Geophys. Res. 109, D22302. http://dx.doi.org/10.1029/2004JD004686. Montzka, S.A., Calvert, P., Hall, B.D., Elkins, J.W., Conway, T.J., Tans, P.P., Sweeney, C., 2007. On the global distribution, seasonality, and budget of atmospheric carbonyl sulfide (COS) and some similarities to CO2. J. Geophys. Res. 112 (D9). http://dx.doi.

19