Search

• Innovative setup for the simultaneous measurement of ozone in 5 µm and 10 µm regions. • Low-pressure measurements on pure ozone traceable to new cross section at 253.65 nm. • Comprehensive uncertainty budget with uncertainty of line positions at 3 · 10 − 5 cm − 1 • First relative standard uncertainty of line intensities equal or better than 1.1%. • High data quality allows to resolve UV-IR discrepancy and to re-evaluate databases. [Display omitted] This is the first in a series of two papers concerned with new accurate high resolution spectroscopic measurements of line positions and intensities of ozone in the 5 µm and 10 µm regions using a commercial Michelson Fourier transform spectrometer. Measurements in both wavelength regions have been made quasi simultaneously using a special H-shaped absorption cell with two parallel absorption paths of different lengths that remained connected during the experiments. Accurate ozone partial pressures were monitored on the shorter of the two paths using UV absorption at the 253.65 nm Hg reference line. In this paper, we report in detail on the experimental setup and provide a comprehensive uncertainty analysis. The spectral analysis and retrieved line parameters are provided in the second paper, which also contains the comparison to the literature data. Line positions have been measured with standard uncertainties of u (ν 0) = 3 · 10 − 5 cm-1 in both spectral regions, mainly in the ν 3 and the ν 1 + ν 3 bands of ozone. Intensities of most of the transitions in the ν 3 band have a relative standard uncertainty of u r (S) = 1.1 · 10 − 2 . The uncertainty is slightly smaller for intensities in the ν 1 + ν 3 band at 5 µm. The highly accurate experimental data provide a benchmark for molecular calculations and the 1.1 % line intensity uncertainty will help resolving the discrepancy between UV and IR atmospheric ozone retrievals and correcting current biases in spectroscopic databases. [ABSTRACT FROM AUTHOR]

• Accurate measurements of ozone line positions and intensities at 10 µ m and 5 µ m. • Detailed comparison of new results with databases and literature values. • Systematic biases at 10 µ m and at 5 µ m are revealed in current versions of databases. Ozone line parameters have been retrieved from high-resolution Fourier transform spectra in the 5- µ m and 10- µ m spectral regions using a H-shaped cell with two parallel paths described in the first paper of the series (Multi-spectral investigation of ozone: Part I. Setup & uncertainty budget). The experimental arrangement allows for quasi simultaneous acquisition of spectra at 5 µ m and 10 µ m with consistent ozone pressure. A multispectrum fitting procedure has been used to accurately measure line positions and intensities. Special care has been taken to determine the instrument line shape and its effect on line intensities and sources of uncertainties from the assumption of a molecular line profile have been investigated. Line positions and intensities have been retrieved for 497 transitions at 10 µ m, mostly (476) from the ν 3 band, but also from the ν 1 band (18 transitions) and from the ν 2 +ν 3 –ν 2 hot band (3 transitions). At 5 µ m, 319 transitions have been determined, of which 316 belong to the ν 1 +ν 3 and 3 to the 2ν 3 bands. Comparisons with previous experiments and calculations as well as with databases are presented, showing systematic discrepancies. The new intensity data at the unprecedented 1% accuracy level unambiguously indicates that atmospheric databases currently suffer from systematic biases up to 4.2% in the ν 1 +ν 3 band and 2.2% in the ν 3 band. The present results will help to resolve the long-standing UV-IR discrepancy observed in atmospheric and laboratory ozone studies. Comparisons with yet unpublished measurements and Hamiltonian-type calculations, as well as with recently published ab-initio calculations show an average discrepancy between 0.1% and 1.2% with sub-percent (0.4–0.8%) dispersion. The small scatter of the comparison confirms the sub-percent precision of both, the calculations, and the measurements. [Display omitted] [ABSTRACT FROM AUTHOR]