Our changing atmosphere: evidence based on long-term infrared solar observations at the Jungfraujoch since 1950.

The Institute of Astrophysics of the University of Liège has been present at the High Altitude Research Station Jungfraujoch, Switzerland, since the late 1940s, to perform spectrometric solar observations under dry and weakly polluted high-mountain conditions. Several solar atlases of photometric quality, extending altogether from the near-ultra-violet to the middle-infrared, were produced between 1956 and 1994, first with grating spectrometers then with Fourier transform instruments. During the early 1970s, scientific concerns emerged about atmospheric composition changes likely to set in as a consequence of the growing usage of nitrogen-containing agricultural fertilisers and the industrial production of chlorine-bearing compounds such as the chlorofluorocarbons and hydro-chlorofluorocarbons. Resulting releases to the atmosphere with ensuing photolysis in the stratosphere and catalytic depletion of the protective ozone layer prompted a worldwide consortium of chemical manufacturing companies to solicit the Liège group to help in clarifying these concerns by undertaking specific observations with its existing Jungfraujoch instrumentation. The following pages evoke the main steps that led from quasi full sun-oriented studies to priority investigations of the Earth's atmosphere, in support of both the Montreal and the Kyoto Protocols.

Infrared measurements of HF and HCl total column abundances above Kitt Peak, 1977-1990: seasonal cycles, long-term increases, and comparisons with model calculations.

Series of high-resolution (approximately 0.01 cm-1) solar absorption spectra recorded with the McMath Fourier transform spectrometer on Kitt Peak (altitude 2.09 km, 31.9 degrees N, 111.6 degrees W) have been analyzed to deduce total column amounts of HF on 93 different days and HCl on 35 different days between May 1977 and June 1990. The results are based on the analysis of the HF and H35Cl (1-0) vibration-rotation band R(1) lines which are located at 4038.9625 and 2925.8970 cm-1, respectively. All of the data were analyzed using a multilayer, nonlinear least squares spectral fitting procedure and a consistent set of spectroscopic line parameters. The results indicate a rapid increase in total HF and a more gradual increase in total HCl with both trends superimposed on short-term variability. In addition, the total columns of both gases undergo a seasonal cycle with an early spring maximum and an early fall minimum, with peak-to-peak amplitudes equal to 25% for HF and 13% for HCl. In the case of HF, the changes over the 13 years of measurement are sufficiently large to determine that a better fit is obtained assuming a linear rather than an exponential increase with time. For HCl, linear and exponential models fit the data equally well. Referenced to calendar year 1981.0 and assuming a sinusoidal seasonal cycle superimposed on a linear total column increase with time, HF and HCl increase rates of (10.9 +/- 1.1)% yr-1 and (5.1 +/- 0.7)% yr-1 and total columns of (3.17 +/- 0.11) x 10(14) and (1.92 +/- 0.06) x 10(15) molecules cm-2 (2 sigma) are derived, respectively; the corresponding best fit mean exponential increase rates are equal to (7.6 +/- 0.6)% yr-1 and (4.2 +/- 0.5)% yr-1 (2 sigma). Over the 13-year observing period, the HF and HCl total columns increased by factors of 3.2 and 1.8, respectively. Based on HF and HCl total columns deduced from measurements on the same day, the HF/HCl total columns ratio increased from 0.14 in May 1977 to 0.23 in June 1990. Short-term temporal variations in the HF and HCl total columns are highly correlated; these fluctuations are believed to be caused by dynamical variability in the lower stratosphere. The results of this investigation are compared with previously reported measurements and with time-dependent, two-dimensional model calculations of HF and HCl total columns based on emission histories and photo-oxidation rates for the source molecules.

ATMOS data processing and science analysis methods.

The ATMOS (atmospheric trace molecule spectroscopy) instrument, a high speed Fourier transform spectrometer operating in the middle IR (2.2-16 microm), recorded more than 1500 solar spectra at approximately 0.0105-cm(-1) resolution during its first mission onboard the shuttle Challenger in the spring of 1985. These spectra were acquired during high sun conditions for studies of the solar atmosphere and during low sun conditions for studies of the earth's upper atmosphere. This paper describes the steps by which the telemetry data were converted into spectra suitable for analysis, the analysis software and methods developed for the atmospheric and solar studies, and the ATMOS data analysis facility.

Line mixing effects in solar occultation spectra of the lower stratosphere: measurements and comparisons with calculations for the 1932-cm(-1) CO(2) Q branch.

Line mixing effects have been observed in a CO(2)Q branch recorded in 0.01-cm(-1) resolution IR solar occultation spectra of the lower stratosphere. The spectral data were obtained by the Atmospheric Trace Molecule Spectroscopy Fourier transform spectrometer during the Spacelab 3 shuttle mission in the spring of 1985. Analysis of the 1932.47-cm(-1) (11102)?(00001) band Q branch of (12)C(16)O(2) shows absorption coefficients below the band origin approximately 0.62 times those calculated using a standard Voigt line shape function. Calculations of line mixing using the Lorentz halfwidths of the lines and a simple energy gap scaling law to parametrize rotational energy transfer reproduce the observed absorption coefficients to ~10%. The present results provide the first quantitative information on air-broadened line mixing effects in a Q branch at low temperatures (~210 K) and show that these effects are significant even at the low pressures of the lower stratosphere (~100 mbar).

Infrared spectroscopic measurements of tropospheric trace gases.

Absorption features of several trace gases have been detected in 0.017-cm -(1)-resolution infrared spectra recorded over surface-level paths of 0.5 and 1.5 km at the National Solar Observatory on Kitt Peak, near Tucson, Arizona. Measurements of 0(3), HCOOH, NH(3), and H(2)CO are briefly discussed.

Measurements of absolute line intensities in carbon dioxide bands near 5.2 microm.

A nonlinear least-squares spectral fitting procedure has been used to derive experimental absolute intensities for over 300 unblended lines belonging to twelve bands of (12)C(16)O2, (13)C(16)O2, (16)O(12)C(18)O, (16)O(12)C(17)O, and (16)O(13)C(18)O in the 5.2-microm region. The spectral data were recorded at 0.01-cm(-1) resolution and room temperature with the Fourier transform spectrometer in the McMath solar telescope complex at the National Solar Observatory on Kitt Peak and have a signal-to-rms noise ratio of 2000-4000. A natural sample of carbon dioxide was used as the sample gas. For each band, the measured line intensities have been analyzed to derive the vibrational band intensity and coefficients of the F factor. The results are compared to the values used to calculate the intensities in the 1982 Air Force Geophysics Laboratory line parameters compilation.

Measurements of air-, N(2)-, and O(2) -broadened halfwidths and pressure-induced line shifts in the v(3) band of (13)CH(4).

Air-, nitrogen-, and oxygen-broadened halfwidth and pressure-induced line shift coefficients have been measured for over seventy-five individual vibration-rotation transitions in the v(3) fundamental band of (13)CH(4)at 3 mum from room temperature IR laboratory absorption spectra recorded at a 0.01-cm(-1) resolution with a Fourier transform spectrometer. Transitions up to J" = 13 in the P-branch and J" = 7 in the R-branch were analyzed using a nonlinear least-squares curve fitting technique assuming a Voigt line shape. From the analysis, we have determined the mean ratios of air- to nitrogen-broadened and air- to oxygen-broadened halfwidth coefficients obtained for the same transitions as 1.00 +/- 0.03 and 1.06 +/- 0.04, respectively. The measured pressure shift coefficients were predominantly negative, and the mean and standard deviation values for air-, nitrogen-, and oxygen-broadened shifts were -0.0067 +/- 0.0032, -0.0068 +/- 0.0041, and -0.0066 +/- 0.0028 cm(-1) atm(-1), respectively. Comparisons are made with the results obtained for similar transitions in the v(3) band of (12)CH(4) and the v(4) band of (13)CH(4).

Measurements of Lorentz air-broadening coefficients and relative intensities in the H(2)(16)O pure rotational and v(2) bands from long horizontal path atmospheric spectra.

Lorentz air-broadening coefficients and relative intensities have been measured for forty-three lines in the pure rotational band and twenty lines in the v(2) band of H(2)(16)O between 800 and 1150 cm(-1). The results were derived from analysis of nine 0.017-cm(-1) resolution atmospheric absorption spectra recorded over horizontal paths of 0.5-1.5 km with the McMath Fourier transform spectrometer and main solar telescope operated on Kitt Peak by the National Solar Observatory. A nonlinear least-squares spectral fitting technique was used in the spectral analysis. The results are compared with previous measurements and calculations. In most cases, the measured pressure-broadening coefficients and intensities are significantly different from the values in the 1986 HITRAN line parameters compilation.

Measurements of air-broadened and nitrogen-broadened Lorentz width coefficients and pressure shift coefficients in the nu(4) and nu(2) bands of (12)CH(4).

Air-broadened and N(2)-broadened halfwidth and pressure shift coefficients of 294 transitions in the nu(4) and nu(2) bands of (12)CH(4) have been measured from laboratory absorption spectra recorded at room temperature with the Fourier transform spectrometer in the McMath solar telescope facility of the National Solar Observatory. Total pressures of up to 551 Torr were employed with absorption paths of 5-150 cm, CH(4) volume mixing ratios of 2.6% or less, and resolutions of 0.005 and 0.01 cm(-1). A nonlinear least-squares spectral fitting technique has been utilized in the analysis of the twenty-five measured spectra. Lines up to J' = 18 in the nu(4) band and J' = 15 in the nu(2) band have been analyzed.

Absolute intensities of CO(2) lines in the 3140-3410-cm(-1) spectral region.

Absolute intensities for 430 transitions belonging to eleven rotation-vibration bands of (12)C(16)O(2),(13)C(16)O(2) and(16)O(12)C(18)O in the 3140-3410-cm(-1) spectral region have been determined by analyzing spectra recorded at 0.01-cm(-1) resolution with the Fourier transform spectrometer in the McMath solar telescope complex at the National Solar Observatory on Kitt Peak. The data were recorded at room temperature and low pressures (<10 Torr) using a natural sample of carbon dioxide. Intensities were derived using a nonlinear least-squares spectral fitting procedure, and the values obtained for each band have been analyzed to determine the vibrational band intensity and nonrigid rotor coefficients. An alternative mathematical formulation is shown in the case of bands for which the Coriolis effect is large and the Q-branch line intensities were not determinate either because they were severely blended or absent from the spectra. Comparisons are made between the results obtained in this study and other published values.

Air-broadened Lorentz halfwidths and pressure-induced line shifts in the nu(4) band of (13)CH(4).

Air-broadened halfwidths and pressure-induced line shifts in the nu(4) fundamental of (13)CH(4) were determined from spectra recorded at room temperature and at 0.01-cm(-1) resolution using a Fourier transform spectrometer. Halfwidths and pressure shifts were determined for over 180 transitions belonging to J''values of

Measurements of (12)CH4 nu4 band halfwidths using a tunable diode laser system and a Fourier transform spectrometer.

Air-broadened and N2-broadened halfwidths at room temperature for twenty-five transitions in the 4 fundamental band of '2CH4 have been determined from IR absorption spectra recorded with a tunable diode laser spectrometer. Two tunable diode lasers operating in the 1250-1380-cm(-1) region were used to obtain these data. Air-broadened halfwidths for twenty of these lines were also determined from additional spectra recorded at 0.01-cm(-1) resolution with the Fourier transform spectrometer in the McMath solar telescope complex on Kitt Peak. The air-broadened halfwidths obtained from these two techniques are very consistent with agreement better than 3% in most cases.

Measurements of argon broadened Lorentz width and pressureinduced line shift coefficients in the nu(4) band of (12)CH(4).

Room temperature argon broadened halfwidth and pressure-induced line shift coefficients have been determined for 118 transitions in the nu(4) band of (12)CH(4) from analysis of high resolution laboratory absorption spectra recorded with the McMath Fourier transform spectrometer operated on Kitt Peak by the National Solar Observatory. Transitions up to J''= 12 have been measured using a nonlinear least-squares spectral fitting procedure. The variation of the measured halfwidth coefficients with symmetry type and rotational quantum number is very similar to that measured previously for N(2) and air broadening, but the absolute values of the argon broadening coefficients are all smaller. On average, the ratio of the argon broadened halfwidth coefficient to the corresponding N(2) broadened halfwidth coefficient is 0.877 +/- 0.017 (2sigma). More than 95% of the pressure-induced shifts are negative with values ranging from -0.0081 to +0.0055 cm(-1) atm(-1). The pressure shifts in argon are nearly equal to corresponding values measured previously in N(2) and air.

Spectroscopic line parameters for the nu(6) band of carbonyl fluoride.

New measurements and analysis of high resolution(0.0025 cm(-1)) laboratory spectra of the carbonyl fluoride v6 band are described. The data are used to generate line parameters suitable for high resolution atmospheric studies.

Observations of the infrared solar spectrum from space by the ATMOS experiment.

The final flight of the Atmospheric Trace Molecule Spectroscopy experiment as part of the Atmospheric Laboratory for Applications and Science (ATLAS-3) Space Shuttle mission in 1994 provided a new opportunity to measure broadband (625-4800 cm(-1), 2.1-16 µm) infrared solar spectra at anunapodized resolution of 0.01 cm(-1) from space. The majority of the observations were obtained as exoatmospheric, near Sun center, absorption spectra, which were later ratioed to grazing atmospheric measurements to compute the atmospheric transmission of the Earth's atmosphere and analyzed for vertical profiles of minor and trace gases. Relative to the SPACELAB-3 mission that produced 4800 high Sun spectra (which were averaged into four grand average spectra), the ATLAS-3 mission produced some 40,000 high Sun spectra (which have been similarly averaged) with an improvement in signal-to-noise ratio of a factor of 3-4 in the spectral region between 1000 and 4800 cm(-1). A brief description of the spectral calibration and spectral quality is given as well as the location of electronic archives of these spectra.

Remote sensing of the Earth's atmosphere from space with high-resolution Fourier-transform spectroscopy: development and methodology of data processing for the Atmospheric Trace Molecule Spectroscopy experiment.

The methodology of spectroscopic remote sensing with high-resolution Fourier-transform spectra obtained from low Earth orbit by the Atmospheric Trace Molecule Spectroscopy (ATMOS) experiment is discussed. During the course of the Atmospheric Laboratory for Applications and Science (ATLAS) shuttle missions (1992-1994) a flexible, yet reproducible, retrieval strategy was developed that culminated in the near-real-time processing of telemetry data into vertical profiles of atmospheric composition during the ATLAS-3 mission. The development, evolution, robustness, and validation of the measurements are presented and assessed with a summary comparison of trace-gas observations within the Antarctic polar vortex in November 1994.

Pressure sounding of the middle atmosphere from ATMOS solar occultation measurements of atmospheric CO(2) absorption lines.

A method for retrieving the atmospheric pressure corresponding to the tangent point of an infrared spectrum recorded in the solar occultation mode is described and applied to measurements made by the Atmospheric Trace Molecule Spectroscopy (ATMOS) Fourier-transform spectrometer. Tangent pressure values are inferred from measurements of isolated CO(2) lines with temperature-insensitive strengths by measuring the slant-column CO(2) amount and by adjusting the viewing geometry until the calculated column matches the observed column. Tangent pressures are determined with a spectroscopic precision of l%-3%, corresponding to a tangent-point height precision of 70-210 m. The total uncertainty is limited primarily by the quality of the spectra and ranges between 4% and 6% (280-420 m) for spectra with signal-to-noise ratios of 300:1 and between 4% and 10% for spectra with signal-to-noise ratios of 100:1. The retrieval of atmospheric pressure increases the accuracy of the retrieved-gas concentrations by minimizing the effect of systematic errors introduced by climatological pressure data, ephemeris parameters, and the uncertainties in instrumental pointing.

Absolute intensity measurements of the (11(1)0)II <--00(0)0 band of 12C16O2 at 5.2 microm.

A nonlinear least-squares fitting procedure has been used to derive absolute intensities for lines in the P, R, and Q branches of the (11(1)0)II <-- 00(0)0 band of 12Cl602 (band center = 1932 cm(-1)) from long-path 0.01-cm(-1) resolution laboratory spectra. The spectral data were recorded at room temperature and low pressure (<10 Torr) with the Fourier transform spectrometer in the McMath solar telescope complex at Kitt Peak National Observatory. The observed line intensities were analyzed to derive the vibrational band intensity and F-factor coefficients. To obtain a good fit to the data, it was necessary to include terms in the expression for the F factor, which account for Coriolis-type and Fermi-type interactions and centrifugal distortion effects.

Infrared measurements of increased CF(2)Cl(2) (CFC- 12) absorption above the South Pole.

High-resolution ground-based solar spectra recorded at the Amundsen-Scott South Pole station in Dec. 1980 and Nov. 1986 have been analyzed in the region of the CF(2)Cl(2) (chlorofluorocarbon 12) nu(8) band Q branches at 1161 cm(-1). An increase in the CF(2)Cl(2) total vertical column above the South Pole of 1.24 +/- 0.15 over the 6-yr period, corresponding to an average rate of increase of 3.6 +/- 2.1%, is derived. This rate of increase is lower than indicated by in situ measurements at the South Pole over the same time period, but there is agreement when the rather error bars of the spectral measurement results are considered. Spectroscopic parameters that can successfully model CF(2)C1(2) absorption at low temperatures are needed to improve retrieval accuracies and could be applied to a number of pre-1980 atmospheric spectral data sets in the literature to obtain an improved record of early CF(2)Cl(2) concentration trends for comparison with estimates of historical release rates.

Improved line parameters for ozone bands in the 10-microm spectral region.

A complete update of spectroscopic line parameters for the 10-microm bands of ozone is reported. The listing contains calculated positions, intensities, lower state energies, and air- and self-broadened halfwidths of more than 53,000 lines. The results have been generated using improved spectroscopic parameters obtained in a number of recent high resolution laboratory studies. A total of eighteen bands of (16)O(3) (sixteen hot bands plus the nu(1) and nu(3) fundamentals) are included along with the nu(1) and nu(3) fundamentals of both (16)O(16)O(18)O and (16)O(18)O(16)O. As shown by comparisons of line-by-line simulations with 0.003-cm(-1) resolution balloon-borne stratospheric solar spectra, the new parameters greatly improve the accuracy of atmospheric calculations in the 10-microm region, especially for the isotopic (16)O(16)O(18)O and (16)O(18)O(16)O lines.