Non-Voigt line-shape effects on retrievals of atmospheric ozone: Collisionally isolated lines

This paper addresses the question of errors in retrievals of vertical profiles of ozone from atmospheric spectra caused by assuming that the absorption lines have pure Voigt line shapes. The case of collisionally isolated transitions (no line mixing) is treated by considering only the effects of the speed dependence (SD) of the pressure broadening. The case of O3 retrievals from a sequence of limb transmission spectra is first treated theoretically. The results show that the influence of SD is very small, leading to changes in the residuals and in the retrieved O3 mixing ratios smaller than 1%. These findings are then confirmed by treating a series of spectra recorded by the ACE-FTS instrument. Similar exercises are also made for other observation techniques by treating simulated or measured limb and nadir emission spectra as well as ground-based solar and in-situ laser transmission data. All lead to the general conclusion that SD (and Dicke narrowing) can be neglected in retrievals of ozone amounts from recorded atmospheric spectra. Indeed, the biases caused in the ozone profiles by the use of pure Voigt line shapes still remain significantly smaller than the total error/uncertainty from other sources such as improper line intensities and widths, uncertainty in the instrument function, errors in the pressure and temperature profiles and so forth.