[1] Solar absorption spectra recorded over a 24-year time span have been analyzed to retrieve average free tropospheric mixing ratios of carbonyl sulfide (OCS). The measurements were recorded with the Fourier transform spectrometer located in the U.S. National Solar Observatory McMath solar telescope facility on Kitt Peak (altitude 2.09 km, lat. 31.9°N, long. 111.6°W), southwest of Tucson, Arizona, and were obtained on 167 days between May 1978 and February 2002, typically at 0.01-cm⁻¹ spectral resolution. A best fit to the time series shows an average mixing ratio of 566 pptv (1 pptv = 10⁻¹² per unit volume) between 2.09 and 10 km, a small but statistically significant long-term decrease equal to (−0.25 ± 0.04)% yr⁻¹, 1 sigma, and a seasonal variation with a summer maximum, a winter minimum, and a peak amplitude of (1.3 ± 0.4)%, 1 sigma, relative to the mean. Although a statistically significant decline and seasonal variation have been detected, both are exceedingly small. The present results confirm and extend earlier studies showing that the OCS free tropospheric abundance at northern midlatitudes has remained nearly constant over the last decades.

1. Introduction

[2] Carbonyl sulfide (OCS) is the predominant sulfur-bearing gas in the remote troposphere with a global average concentration of about 500 parts per trillion by volume (pptv or 10⁻¹² ppv) [e.g., Sandalls and Penkett, 1977; Maroulis et al., 1977; Torres et al., 1980; Rasmussen et al., 1982a, 1982b; Carroll, 1985; Johnson and Harrison, 1986; Johnson et al., 1993; Mihalopoulos et al., 1991; Thornton et al., 1996; Notholt et al., 2000] and a lifetime of about 4 years [Chin and Davis, 1995; Watts, 2000], though estimates from measurements and models range from 1–7 years. Despite a high tropospheric abundance, its low reactivity and limited knowledge of the strengths of its sources and sinks have resulted in an uncertainty in its chemistry and long-term trend. Higher net emission fluxes than sinks in most models [e. g., Chin and Davis, 1993, 1995; Klelliström, 1998] should lead to an increase in OCS tropospheric concentrations as a function of time. However, measured trends near zero have been reported from analyses of multiyear tropospheric and stratospheric mixing ratio [Chin and Davis, 1995; Bandy et al., 1992; Mihalopoulos et al., 1991; Thornton et al., 1996], ground-based total column abundance [Rinsland et al., 1992; Mahieu et al., 1997; Griffith et al., 1998], and lower stratospheric mixing ratio [Zander et al., 1998; Rinsland et al., 1996] data sets. The discrepancy may be explained by an enhanced role for atmospheric removal by soils [Castro and Galloway, 1991; Kuhn et al., 1999; Kesselmeier et al., 1999; Simmonds et al., 1999] and a revision in the net influence by open and coastal oceans [Watts, 2000], though large uncertainty remains in both the OCS source and sink budgets.

[3] The upward flux of sulfur through the tropical tropopause to the stratosphere provides an in situ source of aerosol through its photooxidation to form SO₂, which is photochemically destroyed to form aerosol particles that contribute to the background stratospheric aerosol layer [Crutzen, 1976; World Meteorological Organization (WMO), 1986]. Estimates of the OCS flux from the tropical tropopause to the stratosphere vary widely [Klelliström, 1998, section 3.4, Table VII], but the range suggests its contribution to the stratospheric aerosol loading is less than initially estimated [Crutzen, 1976]. As the stratospheric aerosol layer plays a significant role in the Earth's radiative balance and climate [e.g., Lacis et al., 1992] and provides a surface for heterogeneous chemistry with an important influence on ozone [Hofmann and Solomon, 1989; Rodriguez et al., 1994], changes in tropospheric OCS loading or vertical transport through the tropical tropopause could lead to changes in OCS concentrations and other key lower stratospheric species such as NO₂ and HNO₃ (see the results of Hofmann and Solomon [1989]) as a result of heterogeneous chemistry and sulfate aerosol production. Although an increase in the background stratospheric sulfuric acid aerosol mass at northern midlatitudes was reported [Hofmann and Rosen, 1980; Hofmann, 1990], the observed increase rate is too rapid to be due to OCS alone [Hofmann, 1991]. It was hypothesized that additional sources, particularly sulfur emitted by jet aircraft flying in the 11–12 km region, must be responsible for most of the increase [Hofmann, 1991]. However, a study comparing aerosol trends during 1979 and 1989–1991 concluded that the 1989–1991 period, though low, still contained a significant volcanic signature that could not be completely attributed to nonvolcanic sources [Thomason et al., 1997].

[4] In view of these studies, precise, long-term sets of tropospheric and stratospheric OCS measurements are needed for trend evaluation and to obtain insights into source and sink strengths and their geographic locations. As most previous studies were limited in time duration to about one decade, an extended time series might reveal a previously undetected trend or seasonal changes. Although a previous analysis suggested that the main sources of OCS are natural, with oceans, microbial processes in soils, and the conversion of CS₂ (also predominately natural) contributing about 30, 20, and 30%, respectively [e.g., Khalil and Rasmussen, 1984], these estimates are quite uncertain. Khalil and Rasmussen [1984] inferred a northern-to-southern interhemispheric ratio (IHR) of 1.05 from published OCS measurements [Torres et al., 1980], and used this result to deduce that only about 25% of the OCS in the atmosphere results from direct and indirect (CS₂ to OCS conversion) anthropogenic activities. The largest anthropogenic source was attributed to emissions from biomass burning, about 10%, followed by releases from coal-fired power plants (about 4%), and automobiles, chemical industry, and sulfur recovery processes (about 3%) [Khalil and Rasmussen, 1984]. However, Turco et al. [1980] concluded that as much as 50% of OCS emissions might originate from anthropogenic emissions. Other estimates of the mean IHR are as high as 1.25 [Bingemer et al., 1990], though most estimates are lower [Carroll, 1985; Johnson and Harrison, 1986; Mihalopoulos et al., 1991; Griffith et al., 1998; Xu et al., 2001; Sturges et al., 2001].

[5] In this investigation we report OCS average free tropospheric mixing ratios derived from the analysis of high-spectral resolution ground-based infrared solar absorption spectra recorded with the Fourier transform spectrometer operated at the U.S. National Solar Observatory facility on Kitt Peak in southern Arizona, a remote high-altitude station located in a semi-arid region with a sparse population. The time series covers May 1978 to February 2002 and extends the results reported previously [Rinsland et al., 1992] by more than a decade. With respect to the previous work which dealt with total vertical column abundances, the analysis has been performed with an improved retrieval algorithm and an updated set of spectroscopic parameters with the goal of providing a more accurate and extended time series of tropospheric abundances, long-term, trends and seasonal variations.

2. Observations and Analysis

[6] The Kitt Peak solar spectra were recorded with the custom-made folded Michelson Fourier transform spectrometer (FTS) [Brault, 1978] located in the McMath solar telescope complex of the U.S. National Solar Observatory (altitude 2.09 km, lat. 31.9°N, long. 111.6°W), southwest of Tucson, Arizona. Individual spectra were recorded with two InSb detectors and a KCl or CaF₂ beam splitter with a weak apodizing function, mostly at 0.01-cm⁻¹ resolution, where the unapodized spectral resolution is defined as 1/2Δ_max and Δ_max is the maximum optical path difference. Retrievals were derived by fitting 3 spectral regions in the OCS ν₃ fundamental band, which is 2 orders of magnitude stronger than any other OCS IR band [cf. Kagann, 1982]. The typical signal-to-noise in an individual spectrum is 2000. The number of measurement days per year ranged from 1 to 13 with more frequent observations obtained during the last five years as a consequence of an increased commitment to the Network for the Detection of Stratospheric Change (NDSC) [Kurylo, 1991] (available at http://www.ndsc.ws) of which Kitt Peak is an official complementary site.

2.1. Analysis Method

[7] The measurements were analyzed with version 3.8 of the SFIT2 algorithm [Pougatchev et al., 1995; Rinsland et al., 1998], based on a semi-empirical application [Parrish et al., 1992; Connor et al., 1995] of the optimal estimation method [Rodgers, 1990]. Volume mixing ratio profiles of one or more molecules are retrieved by simultaneously fitting one or more narrow spectral intervals (microwindows) in one or more solar spectra. Examples of its use have been reported in recent papers [Jones et al., 2001; Rinsland et al., 2000, 2001a, 2001b].

[8] Profiles were retrieved with a 29-layer forward model that extended from the surface to 100 km with a layer vertical thickness of less than 1 km in the lowest layer, 1 km between 3 and 14 km, increasing to 2 km in the lower stratosphere, and even thicker above. Refractive ray-tracing, air masses, and density-weighted effective pressures in each layer were performed with an algorithm based on the FSCATM program [Gallery et al., 1983] modified and updated for ground-based retrievals from NDSC stations with a set of a priori profiles consistent with typical recent measurements [Meier et al., 2002].

[9] The stratospheric portion of the a priori OCS mixing ratio profile was an average of ATMOS (Atmospheric Trace Molecule Spectroscopy) version 3 measurements recorded during the final November 1994 mission between latitudes of 22°N to 42°N [Irion et al., 2002]; it was extended below 13.5 km assuming a constant volume mixing ratio of 495 pptv based on the mean of measurements of background marine air in the tropical north Pacific in September–October 1991 covering 0–40°N latitude [Thornton et al., 1996]. The assumption of a constant tropospheric mixing ratio is consistent with previous measurements [e.g., Rasmussen et al., 1982b, Figure 2] and its long tropospheric lifetime. A priori vertical profiles of other gases were taken mostly from reference profiles, as tabulated by [Meier et al., 2002].

[10] Temperature profiles assumed in the analysis were daily mean compilations performed by the National Centers for Environmental Prediction (NCEP) for the location of Kitt Peak and the measurement dates. They were smoothly connected to the U.S. Standard Atmosphere [1976] above 65 km.

[11] Table 1 lists the spectral ranges and retrieval parameters for the three microwindows fitted simultaneously in the analysis. Retrievals were performed from solar spectra recorded at astronomical zenith angle of 85° or less. The OCS mixing ratio uncertainty normalized to the a priori value was assumed to increase with altitude from 0.15 near the surface to 1.5 in the stratosphere and above. The vertical profile of O₃ was also retrieved from the spectra. The OCS covariance matrix was assumed diagonal. Total columns of H₂O and CO₂ were both retrieved by scaling the a priori profile of each molecule by a single factor. In addition, 4 parameters were adjusted to model the absorption by solar CO lines [Rinsland et al., 1998, section 3.2.2], which is important for the present investigation because of the strong absorption by solar CO in two of the 3 microwindows used here.

Table 1. Microwindows, Primary Target Lines, Principal Interferences, and OCS Spectral Parameters Adopted in the Analysis^a

Microwindow	Interfering Molecules	OCS Line Position	Spectroscopic Assignment	Intensity	γ(air)	E″	n(air)
2045.30–2045.67	CO₂, O₃, solar CO	2045.578488	P(37)	7.058E-19	0.0876	285.1308	0.60
2051.18–2051.48	O₃, CO₂, H₂O	2051.331394	P(25)	1.008E-18	0.0919	131.8385	0.82
2055.64–2055.96	solar CO, O₃, CO₂	2055.860551	P(15)	9.075E-19	0.0964	48.6831	0.88

a Standard HITRAN database units [Rothman et al., 1998, Table 3]. Positions and lower state energies (E″) are in cm⁻¹, intensity are cm⁻¹/(molecule cm⁻²) at 296 K, air-broadening coefficients γ(air) are in cm⁻¹atm⁻¹ at 296 K, and n(air), the coefficient of the temperature dependence of the air-broadening coefficient is dimensionless. All target OCS lines are in the ν₃ vibrational-rotational band.

2.2. Spectroscopic Parameters

[12] Spectroscopic parameters adopted in the present study originated from several sources. Parameters for H₂O were taken from HITRAN 2001 (available from http://www.hitran.com). The bands of O₃ at 4.7 μm are about one third as strong as those at 10 μm and hence are important absorbers, with significant absorption by the rare O₃ isotopes in the regions used to retrieve OCS [Arlander et al., 1994; Goldman et al., 1998, 2000, 2002]. Line parameters for the ν₁ + ν₃ bands of ¹⁶O¹⁸O¹⁶O and ¹⁶O¹⁶O¹⁸O [Flaud et al., 1994] and ¹⁶O¹⁷O¹⁶O and ¹⁶O¹⁶O¹⁷O [Perrin et al., 2001] were included in the database; absorptions by the latter two isotopes are weak but observable in some regions of high spectral resolution balloon-borne and ground-based infrared atmospheric spectra [Goldman et al., 2002]. Spectral parameters for OCS from HITRAN 2000 were adopted with a total of 5 isotopic species and additional hot bands in the 5-μm region as compared to earlier lists [Goldman et al., 2000]. The air-broadening coefficients for OCS at 296 K are based on a fit to the measurements in Table II of Bouanich et al. [1987]. The temperature dependence of the air-broadening coefficients was taken from the HITRAN 2002 file without modification. The pressure-shift coefficient and its temperature dependence were assumed equal to zero for all OCS lines. Note that this band is sometimes labeled ν₁ rather than ν₃ [e.g., Wells et al., 1990]. The spectroscopic parameters described above constitute a consistent set of the best currently available spectroscopic parameters for the target and interfering lines to be fitted in the selected microwindows. Solar CO lines were simulated with a single layer model assuming an effective temperature of 4500 K with adjustable parameters to fit the depths of the solar CO features [Rinsland et al., 1998, section 3.1].

2.3. Retrieval Characterization

[13] Averaging kernels provide a direct assessment of the theoretical altitude sensitivity of the observations in the absence of errors in the measurements and model parameters [Rodgers, 1990, section 4]. They are a function of the retrieval intervals selected, the spectral resolution of the observations, an assumed signal-to-noise ratio, and the selections of the retrieval parameters, such as the a priori profile and its covariance matrix.

[14] Averaging kernels for mixing ratio profile retrievals of 2.09–10 km and 10–100 km merged layers are plotted in Figure 1. The 10-km limit was selected to provide sampling of the free troposphere with only minor, infrequent contributions from stratospheric layers. The annual average tropopause altitude above Kitt Peak is 14 km [Rinsland et al., 1998]. Daily average tropopause altitudes calculated from a 2.5° × 2.5° NCEP analysis for 2000 and 2001 interpolated to the location of Kitt Peak indicate that the tropopause height is located between 7.6 and 17.5 km, but less than 2% of the occurrences during this two year period were below 10 km (S. Zhou, and A. Miller, NCEP, private communication, 2002).

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

Volume mixing ratio (VMR) averaging kernels for merged layers from 2.09–10 km and 10–100 km altitude vs. height.

[15] Calculations were performed for the typical spectral resolution of 0.01 cm⁻¹ and the 3 microwindows listed in Table 1 assuming a signal-to-noise of 250. The latter value is lower than the typical signal-to-noise of the solar spectra to empirically account for errors in the spectroscopic database and imperfections in the forward model (e.g., modeling of solar CO absorption), consistent with the semi-empirical retrieval approach adopted in SFIT2. The mixing ratio averaging kernels for the two layers are broad with maximum sensitivity at 4.9 km for the lower layer, the focus of this investigation. Although there is significant stratospheric sensitivity, we limit the focus of this work to the evaluation of the free tropospheric trend and seasonal cycle.

[16] Molecule-by-molecule simulations of the transmittance by the principal absorbing species in two of the three selected microwindows are shown in Figure 2. The simulated transmission spectra of OCS and all interfering gases (offset vertically for clarity) are calculated for the spectral resolution and solar zenith angle of the Kitt Peak measurement shown in the lowest trace in each frame. An arrow beneath the measured spectrum marks the location of the target OCS line. Each absorber in the simulation is identified on the right vertical axis including solar CO. Absorption by ¹⁶O¹⁶O¹⁷O and ¹⁶O¹⁷O¹⁶O is negligible in both selected intervals, and therefore omitted. Positions and strengths of the features in the measured spectrum and those in the simulated spectra show close correspondence. Note that solar CO absorbs significantly in the bottom microwindow of Figure 2. Molecule-by-molecule simulations for the interval containing the P(37) OCS line have been displayed previously [Rinsland et al., 1992, Figure 4].

[17] Table 2 lists the most important random and systematic sources of error and estimates of their contribution to the error budget for a typical spectrum. Values have been estimated as described previously [e.g., Rinsland et al., 1998, section 6]. Random error has been reduced by averaging mixing ratios from the 4 layers below 10 km to obtain the 2.09–10 km mean mixing ratio and then calculate daily averages from all measurements. The two most important sources of systematic error are uncertainties in the intensities and air-broadening coefficients of the target OCS lines, which are both estimated as 5%. Additional systematic sources of error are the uncertainty in the retrievals due to forward model approximations, estimated as 3% based on comparisons such as the one described by Goldman et al. [1999], and the bias from the a priori profile, 2%.

Table 2. Estimated OCS 2.09–10 km 1-Sigma Uncertainties for Daily Averages

Error Source	Uncertainty, %
Random Errors
Temperature profile	1
Finite signal-to-noise	<1
Interfering atmospheric lines	2
Zenith angle uncertainty	2
Zero level offsets	<1
RSS total random error	3

Systematic Errors
Spectroscopic parameters	5
A priori profile relative contribution	5
Forward model approximations	3
Instrument line shape function	<1
Zero transmittance offsets	<1
RSS total systematic error	8

3. Results and Discussion

[18] Figure 3 displays fits to the 3 OCS microwindows from the same Kitt Peak spectrum shown in Figure 2, which is typical of the time series. The retrieved daily average volume mixing ratios as well as related fits versus time are displayed in Figure 4. The time series spans 24 years, and to our knowledge, it is the longest time period of near continuous OCS measurements reported to date. After excluding noisy spectra and weak absorption by the target lines [Rinsland et al., 1998, Figure 9], the database contains 347 measurements on 167 days between May 1978 and February 2002.

[19] The time series of 2.09–10 km daily average mixing ratios were fitted with the expression

where C_A is the daily average 2.09–10 km mixing ratio at time t, a₀ is the mean 2.09–10 km mixing ratio of the time series, a₁ is the trend, a₂ is the amplitude of the seasonal cycle, and φ is the phase corresponding to the seasonal maximum. The best fit to the 24-year time series shows a long-term 2.09–10 km volume mixing ratio trend of (−1.41 ± 0.25) pptv yr⁻¹ and a seasonal cycle with a peak amplitude of (7.2 ± 2.2) pptv with φ equal to (0.103 ± 0.09 yr), 1 sigma, relative to the first measurement in the time series in May 1978. Normalized to the mean mixing ratio of 566 pptv, the results indicate a statistically significant long-term decrease in the 2.09–10 km average mixing ratio OCS equal to (−0.25 ± 0.04)% yr⁻¹, and a seasonal cycle amplitude of (1.28 ± 0.40)% with a June maximum and a December minimum. The best fit long-term trend and seasonal cycle are both displayed in Figure 4.

[20] Sets of multi-year ground-based OCS total columns measurements of from high spectral resolution solar absorption measurements were obtained from two northern hemisphere stations, Kitt Peak and the International Scientific Station of the Jungfraujoch, (ISSJ) (46.5°N, 8.0°E, 3.58 km altitude) [Rinsland et al., 1992; Mahieu et al., 1997] and from two midlatitude southern hemisphere stations, Lauder, New Zealand (45.0°S, 169.7°E, 0.37 km altitude) and Wollongong, Australia (34.45°S, 150.88°E, 30 m altitude) [Griffith et al., 1998]. Although a statistically significant decline of (0.3 ± 0.1)% yr⁻¹ was found from a fit to the 1984–1995 ISSJ measurements, large variability led to the conclusion that no unambiguous evidence existed for either a seasonal variation or a long-term trend above that site. The southern hemisphere studies also found no statistically significant long-term OCS total column trend over time periods as long as 4 years. However, after correcting for the seasonal variation in the tropopause height above Wollongong, a statistically significant 12% seasonal variation was identified with a late summer/early autumn maximum (March), which was tentatively attributed to the effects of lower tropospheric air transported from the warm nearby ocean by the prevailing winds towards the station [Griffith et al., 1998, Figure 5]. More recent solar absorption measurements remain consistent with the earlier Wollongong observations and this interpretation (D. W. T. Griffith, unpublished results, 2002).

[21] The relative amplitude of the seasonal variation in the present study is less than reported previously from sites distant from strong sources and sinks. The seasonal variation found from the Kitt Peak time series fit may result from a residual influence of tropopause height changes, as the tropospheric kernel is broad (Figure 1), as already mentioned, or real variations in source or sink strengths. The summer maximum and winter minimum obtained from the time series fit is consistent with that expected from uptake by soils, now recognized as a strong OCS sink. No significant seasonal changes in the OCS total column above 12 km were found from aircraft solar absorption measurements recorded during aircraft flights between 2°N to 48°N latitude during summer and winter time periods in 1978, [Mankin et al., 1979], though additional measurements are needed to confirm this conclusion in view of their limited spatial, temporal and seasonal sampling.

[22] Long-term firn air measurements in an Arctic location (72°N, 82°W, 1800m altitude) and two Antarctic stations (77°S, 10°W, 2300m altitude and 75°S, 123°E, 3240m altitude) have been reported recently [Sturges et al., 2001]. The air samples were analyzed by gas chromatography-mass spectrometry to derive OCS mixing ratios spanning atmospheric composition from the early through the mid-20th century up to the time of the sampling (1998/1999). Although the concentrations in Antarctica were found to be almost constant with time, the Arctic measurements appear to have declined over the last 10 years by 8 ± 5%, leading to a present interhemispheric ratio (IHR) close to unity. The authors hypothesized that the decline originated from a reduction in CS₂ emissions by the viscose-rayon industry, its principal anthropogenic source, though a shift of industrialized CS₂ emissions from Europe to Asia was also cited as a possible cause of the recent decline. An unpublished time series of OCS total columns from ground-based solar spectra obtained by one of us (J. Notholt) between 1993 and 2001 at the NDSC station in Ny Ålesund (78.9°N, 11.9°E, 10/20 m altitude) in Spitzbergen between days 140 and 270 of each year (selected to avoid time periods with perturbed chemistry and transport) also show evidence for a small OCS long-term total column decline with time. The Kitt Peak measurements are recorded sufficiently far south that no instances of perturbed winter-spring stratospheric chemistry are observed (e.g., elevated winter/spring ClONO₂).

[23] As mentioned earlier [Rinsland et al., 1992, section 3], it is reasonable to assume the maximum seasonal change in the free tropospheric OCS concentration is no more than a few percent in the absence of strong local sources or sinks as the lifetime of OCS is several years. The present results are consistent with that limit.

[24] Our measurements over a 24-year time span place a strong constraint on the long-term OCS growth rate with a higher vertical resolution than the previous solar absorption measurements, which only reported total columns. Our mean 2.09–10 km mixing ratio of 566 pptv is slightly higher than most northern hemisphere measurements [Sandalls and Penkett, 1977; Maroulis et al., 1977; Torres et al., 1980; Rasmussen et al., 1982a, 1982b; Carroll, 1985; Johnson and Harrison, 1986; Sturges et al., 2001], but consistent within our estimated 6% total systematic error.

4. Summary and Conclusions

[25] A 24-year time series of solar absorption spectra recorded typically at a spectral resolution of 0.01 cm⁻¹ has been analyzed to retrieve carbonyl sulfide average free tropospheric mixing ratios for the 2.09–10 km altitude region. The time series was recorded from the U.S. National Solar Observatory and analyzed with the SFIT2 algorithm based on a semi-empirical implementation of the Rodgers [1990] optimal estimation method. A fit to the 167 daily averages from 347 measurements between May 1978 to February 2002 yields a mean mixing ratio of 566 pptv, a long-term trend of (−0.25 ± 0.04)% yr⁻¹, 1 sigma, superimposed on a seasonal cycle with a peak amplitude of (1.28 ± 0.40)%, 1 sigma, relative to the mean, with a June maximum and a December minimum. The present database reduces the impact of seasonal changes in the tropopause height by restricting the measurements to 2.09 to 10 km, significantly below the 14 km mean tropopause height above Kitt Peak. The results are consistent with previous shorter duration studies that found no significant long-term change in the free tropospheric concentration of OCS at northern midlatitudes, though the small statistically significant decline detected by fitting the extended time span is consistent with other recently reported long-term time series.

Acknowledgments

[26] Research at the NASA Langley Research Center was supported by NASA's Upper Atmosphere Research and the Atmospheric Chemistry and Modeling programs. Research at the University of Denver was supported in part by the NASA and in part by the National Science Foundation. The National Solar Observatory (NSO) is operated by the Association of Universities for Research in Astronomy, Inc. (AURA) under cooperative agreement with the National Science Foundation (NSF) with the present measurements also supported by NASA. Research at the University of Liège was primarily supported by the Federal Office for Scientific, Technical, and Cultural Affairs (OSTC) and by the European Community (EC), both in Brussels, Belgium. We thank Alvin Miller and Shuntai Zhao for providing the NCEP tropopause heights above Kitt Peak for the 2000 and 2001 time period.

References

Arlander, D. W., A. Barbe, M. T. Bourgeois, A. Hamdouni, J.-M. Flaud, C. Camy-Peyret, and P. Demoulin, The identification of ¹⁶O¹⁸O¹⁶O and ¹⁶O¹⁶O¹⁸O ozone isotopes in high resolution ground-based FTIR spectra, J. Quant. Spectrosc. Radiat. Transfer, 52, 267–271, 1994.
10.1016/0022-4073(94)90156-2
CASADSWeb of Science®Google Scholar
Bandy, A. R., D. C. Thornton, D. L. Scott, M. Lalevic, E. E. Lewis, and A. R. Driedger III, A time series for carbonyl sulfide in the northern hemisphere, J. Atmos. Chem., 14, 527–534, 1992.
10.1007/BF00115256
CASWeb of Science®Google Scholar
Bingemer, H. G., S. Bürgermeister, R. L. Zimmermann, and H.-W. Georgii, Atmospheric OCS: Evidence for a contribution of anthropogenic sources? J. Geophys. Res., 95, 20,617–20,622, 1990.
10.1029/JD095iD12p20617
CASADSWeb of Science®Google Scholar
Bouanich, J.-P., J. Walrand, S. Alberty, and G. Blanquet, Diode-laser measurements of oxygen-broadened line widths in the v₁ band of OCS, J. Mol. Spectrosc., 123, 37–47, 1987.
10.1016/0022-2852(87)90261-X
CASADSWeb of Science®Google Scholar
Brault, J. W., Solar Fourier transform spectroscopy, Osser. Mem. Osserv. Astrofis. Arcetri, 106, 33–52, 1978.
Google Scholar
Carroll, M. A., Measurements of OCS and CS₂ in the free troposphere, J. Geophys. Res., 90, 10,483–10,486, 1985.
10.1029/JD090iD06p10483
ADSWeb of Science®Google Scholar
Castro, M. S., and J. N. Galloway, A comparison of sulfur-free and ambient air enclosure techniques for measuring the exchange of reduced sulfur gases between soils and the atmosphere, J. Geophys. Res., 96, 427–437, 1991.
10.1029/91JD01399
ADSWeb of Science®Google Scholar
Chin, M., and D. D. Davis, Global sources and sinks of OCS and CS₂ and their distributions, Global Biogeochem. Cycles, 7, 321–337, 1993.
10.1029/93GB00568
CASADSWeb of Science®Google Scholar
Chin, M., and D. D. Davis, A reanalysis of carbonyl sulfide as a source of stratospheric background sulfur aerosol, J. Geophys. Res., 100, 8993–9005, 1995.
10.1029/95JD00275
CASADSWeb of Science®Google Scholar
Connor, B. J., A. Parrish, J.-T. Tsou, and M. P. McCormick, Error analysis for the ground-based microwave measurements during STOIC, J. Geophys. Res., 100, 9283–9291, 1995.
10.1029/94JD00413
CASADSWeb of Science®Google Scholar
Crutzen, P. J., The possible importance of CSO for the sulfate layer of the stratosphere, Geophys. Res. Lett., 3, 73–76, 1976.
10.1029/GL003i002p00073
CASADSWeb of Science®Google Scholar
Flaud, J.-M., M. T. Bourgeois, A. Barbe, J. J. Plateaux, and C. Camy-Peyret, The ν1 + ν3 bands of ¹⁶O¹⁸O¹⁶O and ¹⁶O¹⁸O¹⁶O, J. Mol. Spectrosc., 165, 464–469, 1994.
10.1006/jmsp.1994.1149
CASADSWeb of Science®Google Scholar
Gallery, W. O., F. X. Kneizys, and S. A. Clough, Air mass computer program for atmospheric transmittance/radiance calculation: FSCATM, Environ. Res. Pap. 828 (AFGL-TR-0065), 145 pp., Air Force Geophys. Lab., Bedford, Mass., 1983.
Google Scholar
Goldman, A., W. G. Schoenfield, T. M. Stephen, F. J. Murcray, C. P. Rinsland, A. Barbe, A. Hamdouni, J.-M. Flaud, and C. Camy-Peyret, Isotopic ozone in the 5 μ region from high resolution balloon-borne and ground-based FTIR solar spectra, J. Quant. Spectrosc. Radiat. Transfer, 59, 231–244, 1998.
10.1016/S0022-4073(97)00118-0
CASADSWeb of Science®Google Scholar
Goldman, A., C. Paton Walsh, W. Bell, G. C. Toon, J.-F. Blavier, B. Sen, M. T. Coffey, J. W. Hannigan, and W. G. Mankin, Network for the detection of stratospheric change Fourier transform infrared intercomparison at Table Mountain facility, November 1996, J. Geophys. Res., 104, 30,481–30,503, 1999.
10.1029/1999JD900879
CASADSWeb of Science®Google Scholar
Goldman, A., M. T. Coffey, T. M. Stephen, C. P. Rinsland, W. G. Mankin, and J. W. Hannigan, Isotopic OCS from high-resolution balloon-borne and ground-based infrared solar absorption spectra, J. Quant. Spectrosc. Radiat. Transfer, 67, 447–455, 2000.
10.1016/S0022-4073(00)00012-1
CASADSWeb of Science®Google Scholar
Goldman, A., et al., Weak lines absorption in the 5 μm region from high resolution FTIR solar spectra, J. Quant. Spectrosc. Radiat. Transfer, 74, 133–138, 2002.
10.1016/S0022-4073(01)00185-6
CASADSWeb of Science®Google Scholar
Griffith, D. W. T., N. B. Jones, and W. A. Matthews, Interhemispheric ratio and annual cycle of carbonyl sulfide (OCS) total column from ground-based solar FTIR spectra, J. Geophys. Res., 103, 8447–8454, 1998.
10.1029/97JD03462
CASADSWeb of Science®Google Scholar
Hofmann, D. J., Increase in the stratospheric background sulfuric acid aerosol mass in the past 10 years, Science, 248, 996–1000, 1990.
10.1126/science.248.4958.996
CASADSPubMedWeb of Science®Google Scholar
Hofmann, D. J., Aircraft sulphur emissions, Nature, 349, 659, 1991.
10.1038/349659a0
ADSWeb of Science®Google Scholar
Hofmann, D. J., and J. M. Rosen, Stratospheric sulfuric acid layer: Evidence for an anthropogenic component, Science, 208, 1368–1370, 1980.
10.1126/science.208.4450.1368
CASADSPubMedWeb of Science®Google Scholar
Hofmann, D. J., and S. Solomon, Ozone destruction through heterogeneous chemistry following the eruption of El Chichón, J. Geophys. Res., 94, 5029–5041, 1989.
10.1029/JD094iD04p05029
CASADSWeb of Science®Google Scholar
Irion, F. W., et al., The Atmospheric Trace Molecule Spectroscopy Experiment (ATMOS) version 3 data retrievals, Appl. Opt., in press, 2002.
10.1364/AO.41.006968
ADSWeb of Science®Google Scholar
Johnson, J. E., and H. Harrison, Carbonyl sulfide concentrations in the surface waters and above the Pacific Ocean, J. Geophys. Res., 91, 7883–7888, 1986.
10.1029/JD091iD07p07883
CASADSWeb of Science®Google Scholar
Johnson, J. E., A. Bandy, D. C. Thornton, and T. S. Bates, Measurements of atmospheric carbonyl sulfide during the NASA chemical instrumentation and test evaluation project: Implementations for the global COS budget, J. Geophys. Res., 98, 23,443–23,448, 1993.
10.1029/92JD01911
CASADSWeb of Science®Google Scholar
Jones, N. B., C. P. Rinsland, and J. B. Liley, Correlation of aerosol and carbon monoxide at 45°S: Evidence of biomass burning emissions, Geophys. Res. Lett., 28, 709–712, 2001.
10.1029/2000GL012203
CASADSWeb of Science®Google Scholar
Kagann, R. H., Infrared absorption intensities for OCS, J. Mol. Spectrosc., 94, 192–198, 1982.
10.1016/0022-2852(82)90303-4
CASADSWeb of Science®Google Scholar
Kesselmeier, J., N. Teusch, and U. Kuhn, Controlling variables in the uptake of atmospheric carbonyl sulfide by soils, J. Geophys. Res., 104, 11,577–11,584, 1999.
10.1029/1999JD900090
CASADSWeb of Science®Google Scholar
Khalil, M. A. K., and R. A. Rasmussen, Global sources, lifetimes and mass balances of carbonyl sulfide (OCS) and carbonyl disulfide (CS₂) in the Earth's atmosphere, Atmos. Environ., 18, 1805–1813, 1984.
10.1016/0004-6981(84)90356-1
CASADSWeb of Science®Google Scholar
Klelliström, E., A three-dimensional global model study of carbonyl sulfide in the troposphere and lower stratosphere, J. Atmos. Chem., 29, 151–177, 1998.
10.1023/A:1005976511096
Web of Science®Google Scholar
Kuhn, U., C. Ammann, A. Wolf, F. X. Meixner, M. O. Andreae, and J. Kesselmeier, Carbonyl sulfide exchange on an ecosystem scale: Soil represents a dominant sink for atmospheric COS, Atmos. Environ., 33, 995–1008, 1999.
10.1016/S1352-2310(98)00211-8
CASADSWeb of Science®Google Scholar
Kurylo, M. J., Network for the detection of stratospheric change, Proc. Soc. Photo. Opt. Instrum. Eng., 1491, 169–174, 1991.
Google Scholar
Lacis, A., J. Hansen, and M. Sato, Climate forcing by stratospheric aerosols, Geophys. Res. Lett., 19, 1607–1610, 1992.
10.1029/92GL01620
ADSWeb of Science®Google Scholar
Mahieu, E., R. Zander, L. Delbouille, P. Demoulin, G. Roland, and C. Servais, Observed trends in total vertical column abundances of atmospheric gases from IR solar spectra recorded from the Jungfraujoch, J. Atmos. Chem., 28, 227–243, 1997.
10.1023/A:1005854926740
CASWeb of Science®Google Scholar
Mankin, W. G., M. T. Coffey, D. W. T. Griffith, and S. R. Drayson, Spectroscopic measurement of carbonyl sulfide (OCS), Geophys. Res. Lett., 6, 853–856, 1979.
10.1029/GL006i011p00853
CASADSWeb of Science®Google Scholar
Maroulis, P. J., A. L. Torres, and A. R. Bandy, Atmospheric concentrations of carbonyl sulfide in the southwestern and eastern United States, Geophys. Res. Lett., 4, 510–512, 1977.
10.1029/GL004i011p00510
CASADSWeb of Science®Google Scholar
Meier, A., A. Goldman, P. S. Manning, T. M. Stephen, C. P. Rinsland, N. B. Jones, and S. W. Wood, Improvements to air mass calculations from ground-based infrared measurements, J. Quant. Spectrosc. Radiat. Transfer, in press, 2002.
Google Scholar
Mihalopoulos, N., J. P. Putaud, B. C. Nguyen, and S. Belviso, Annual variation of atmospheric carbonyl sulfide in the marine atmosphere in the southern Indian ocean, J. Atmos. Chem., 13, 73–82, 1991.
10.1007/BF00048101
CASWeb of Science®Google Scholar
Notholt, J., G. C. Toon, C. P. Rinsland, N. S. Pougatchev, N. B. Jones, B. J. Connor, R. Weller, M. Gautrois, and O. Schrems, Latitudinal variations of trace gas concentrations in the free troposphere measured by solar absorption spectroscopy during a ship cruise, J. Geophys. Res., 105, 1337–1349, 2000.
10.1029/1999JD900940
CASADSWeb of Science®Google Scholar
Parrish, A., B. J. Connor, J. J. Tsou, I. S. McDermid, and W. G. Chu, Ground-based microwave monitoring of stratospheric ozone, J. Geophys. Res., 97, 2541–2546, 1992.
10.1029/91JD02914
CASADSWeb of Science®Google Scholar
Perrin, A., J.-M. Flaud, F. Keller, M. A. H. Smith, C. P. Rinsland, V. Malathy Devi, D. C. Benner, T. M. Stephen, and A. Goldman, The ν₁ + ν₃ bands of the ¹⁶O¹⁷O¹⁶O and ¹⁶O¹⁶O¹⁷O isotopomers of ozone, J. Mol. Spectrosc., 207, 54–59, 2001.
10.1006/jmsp.2001.8320
CASADSPubMedWeb of Science®Google Scholar
Pougatchev, N. S., B. J. Connor, and C. P. Rinsland, Infrared measurements of the ozone vertical distribution above Kitt Peak, J. Geophys. Res., 100, 16,689–16,697, 1995.
10.1029/95JD01296
CASADSWeb of Science®Google Scholar
Rasmussen, R. A., M. A. K. Khalil, and S. D. Hoyt, The oceanic source of carbonyl sulfide (OCS), Atmos. Environ., 16, 1591–1594, 1982a.
10.1016/0004-6981(82)90111-1
ADSWeb of Science®Google Scholar
Rasmussen, R. A., M. A. K. Khalil, A. J. Crawford, and P. J. Fraser, Natural and anthropogenic trace gases in the southern hemisphere, Geophys. Res. Lett., 9, 704–707, 1982b.
10.1029/GL009i006p00704
CASADSWeb of Science®Google Scholar
Rinsland, C. P., R. Zander, E. Mahieu, P. Demoulin, A. Goldman, D. H. Ehhalt, and J. Rudolph, Ground-based infrared measurements of carbonyl sulfide total column abundances: Long-term trends and variability, J. Geophys. Res., 97, 5995–6002, 1992.
10.1029/92JD00040
ADSWeb of Science®Google Scholar
Rinsland, C. P., et al., Trends of OCS, HCN, SF₆, CHClF₂ (HCFC-22) in the lower stratosphere from 1985 and 1994 atmospheric trace molecule spectroscopy experiment measurements near 30°N latitude, Geophys. Res. Lett., 23, 2349–2352, 1996.
10.1029/96GL01234
CASADSWeb of Science®Google Scholar
Rinsland, C. P., et al., Northern and Southern Hemisphere ground-based infrared spectroscopic measurements of tropospheric carbon monoxide and ethane, J. Geophys. Res., 103, 28,197–28,217, 1998.
10.1029/98JD02515
CASADSWeb of Science®Google Scholar
Rinsland, C. P., E. Mahieu, R. Zander, P. Demoulin, J. Forrer, and B. Buchmann, Free tropospheric CO, C₂H₆, and HCN above central Europe: Recent measurements from the Jungfraujoch station including the detection of elevated columns during 1998, J. Geophys. Res., 105, 24,235–24,249, 2000.
10.1029/2000JD900371
CASADSWeb of Science®Google Scholar
Rinsland, C. P., A. Goldman, R. Zander, and E. Mahieu, Enhanced tropospheric HCN columns above Kitt Peak during the 1982–1983 and 1997–1998 El Niño warm phases, J. Quant. Spectrosc. Radiat. Transfer, 69, 3–8, 2001a.
10.1016/S0022-4073(00)00122-9
CASADSWeb of Science®Google Scholar
Rinsland, C. P., A. Meier, D. W. T. Griffith, and L. S. Chiou, Ground-based measurements of tropospheric CO, C₂H₆, and HCN from Australia at 34°S latitude, J. Geophys. Res., 106, 20,913–20,924, 2001b.
10.1029/2000JD000318
CASADSWeb of Science®Google Scholar
Rodgers, C. D., Characterization and error analysis of profiles retrieved from remote sounding measurements, J. Geophys. Res., 95, 5587–5595, 1990.
10.1029/JD095iD05p05587
ADSWeb of Science®Google Scholar
Rodriguez, J. M., M. K. W. Ko, N. D. Sze, and C. W. Heisey, Ozone response to enhanced heterogeneous processing after the eruption of Mt. Pinatubo, Geophys. Res. Lett., 21, 209–212, 1994.
10.1029/93GL03537
CASADSWeb of Science®Google Scholar
Rothman, L. S., et al., The HITRAN molecular spectroscopic database and HAWKS (HITRAN atmospheric workstation): 1996 edition, J. Quant. Spectrosc. Radiat. Transfer, 60, 665–710, 1998.
10.1016/S0022-4073(98)00078-8
CASADSWeb of Science®Google Scholar
Sandalls, F. J., and S. A. Penkett, Measurements of carbonyl sulfide and carbon disulphide in the atmosphere, Atmos. Environ., 11, 197–199, 1977.
10.1016/0004-6981(77)90227-X
CASADSWeb of Science®Google Scholar
Simmonds, J. S., L. Klemedsson, H. Hultberg, and M. E. Hines, Consumption of atmospheric carbonyl sulfide by coniferous boreal forest soils, J. Geophys. Res., 104, 11,569–11,576, 1999.
10.1029/1999JD900149
ADSGoogle Scholar
Sturges, W. T., S. A. Penkett, J.-M. Barnola, J. Chappellaz, E. Atlas, and V. Stroud, A long-term measurement record of carbonyl sulfide (OCS) in two hemispheres from firn air measurements, Geophys. Res. Lett., 28, 4095–4098, 2001.
10.1029/2001GL013958
CASADSWeb of Science®Google Scholar
Thomason, L. W., G. S. Kent, C. R. Trepte, and L. R. Poole, A comparison of the stratospheric aerosol background periods of 1979 and 1989–1991, J. Geophys. Res., 102, 3611–3616, 1997.
10.1029/96JD02960
ADSWeb of Science®Google Scholar
Thornton, D. C., A. R. Bandy, B. W. Blomquist, and B. E. Anderson, Impact of anthropogenic and biogenic sources and sinks of carbonyl sulfide in the north Pacific troposphere, J. Geophys. Res., 101, 1873–1881, 1996.
10.1029/95JD00617
CASADSWeb of Science®Google Scholar
Torres, A. L., P. J. Maroulis, A. B. Goldberg, and A. R. Bandy, Atmospheric OCS measurements on project Gametag, J. Geophys. Res., 85, 7357–7360, 1980.
10.1029/JC085iC12p07357
CASADSWeb of Science®Google Scholar
Turco, R. P., R. C. Whitten, O. B. Toon, J. B. Pollack, and P. Hamill, OCS, stratospheric aerosols and climate, Nature, 283, 283–286, 1980.
10.1038/283283a0
CASADSWeb of Science®Google Scholar
U.S. Standard Atmosphere (1976), NOAA/NASA/U.S. Air Force, U.S. Government Printing Office, Washington, D.C., 1976.
Google Scholar
Watts, S. F., The mass budgets of carbonyl sulfide, dimethyl sulfide, carbon disulfide, and hydrogen sulfide, Atmos. Environ., 34, 761–779, 2000.
10.1016/S1352-2310(99)00342-8
CASADSWeb of Science®Google Scholar
Wells, J. S., M. Schneider, and A. G. Maki, Heterodyne frequency measurements on OCS near 61.76 THz (2060 cm⁻¹), J. Mol. Spectrosc., 140, 170–176, 1990.
10.1016/0022-2852(90)90015-I
CASADSWeb of Science®Google Scholar
World Meteorological Organization, Atmospheric ozone 1985—Assessment of our understanding of the processes controlling its present distribution and change, Rep. 16, World Meteorol. Org., Geneva, 1986.
Google Scholar
Xu, H., G. Bingemer, H.-W. Georgii, U. Schmidt, and U. Bartell, Measurements of carbonyl sulfide (COS) in surface seawater and marine air, and estimates of the air-sea flux from observations during two Atlantic cruises, J. Geophys. Res., 106, 3491–3502, 2001.
10.1029/2000JD900571
CASADSWeb of Science®Google Scholar
Zander, R., C. P. Rinsland, C. B. Farmer, J. Namkung, R. H. Norton, and J. M. Russell III, Concentrations of carbonyl sulfide and hydrogen cyanide in the free upper troposphere and lower stratosphere deduced from ATMOS/Spacelab 3 infrared solar occultation spectra, J. Geophys. Res., 93, 1669–1678, 1998.
10.1029/JD093iD02p01669
ADSWeb of Science®Google Scholar

Citing Literature

Volume107, IssueD22

27 November 2002

Pages ACH 24-1-ACH 24-9

Ground-based infrared spectroscopic measurements of carbonyl sulfide: Free tropospheric trends from a 24-year time series of solar absorption measurements

Abstract

1. Introduction