Comment on amt-2022-257

Abstract

A new global O3 data product retrieved from Michelson Interferometer for Passive Atmospheric Sounding (MIPAS) spectra with the IMK/IAA MIPAS data processor has been released. These data are based on ESA version 8 recalibration of radiance spectra which takes detector aging into consideration to minimize drifts. The new ozone retrievals use improved temperatures and thus suffer less from the propagation of related errors. Changes in the level-2 processing with respect to previous data versions and relevant to ozone include: (1) The background continuum is now considered up to 58 km. (2) A priori information is now used to constrain the retrieval of the radiance offset. (3) Water vapour is jointly retrieved along with ozone mixing ratios. (4) A more adequate regularization has been chosen. (5) Ozone lines in the MIPAS A band (685–980 cm-1) are used almost exclusively because of inconsistencies in spectroscopic data of the MIPAS AB band (1010–1180 cm-1). Only at altitudes above 50 km, where A band ozone lines do not provide sufficient information, ozone lines in the MIPAS AB band are used. (6) Temperature-adjusted climatologies of vibrational temperatures of O3 and CO2 are considered to account for non-local thermodynamic equilibrium radiation. Ozone errors are estimated to be less than 10 % in the altitude range 20–50 km. The error budget is dominated by the spectroscopic errors of ozone and carbon dioxide. The latter error contribution is propagated from the results of temperature and line-of-sight retrievals. Further notable contributions are the uncertainty of the instrumental line shape function, the gain calibration error, and the spectral noise, directly in the ozone lines and propagated via the previously retrieved temperature and line-of-sight. The error contribution of interfering gases is almost negligible. The vertical resolution in terms of the full width at half maximum of the averaging kernel rows depends on altitude and atmospheric conditions. For the measurement period 2002–2004 it varies between 2.5 km at the lowest altitudes and 6 km at 70 km, while in 2005–2012 it covers 2 to 5.5 km in the same altitude range. The horizontal smearing in terms of the full width at half maximum of the horizontal component of the 2-dimensional averaging kernel matrix is smaller than, or approximately equal to, the distance between two subsequent limb scans, at all altitudes. This implies that the horizontal resolution is sampling-limited or optimal, respectively. Along with the regular representation of the data, that have non-unity averaging kernels, a resampled data version is made available that is free of formal a priori information and thus more user-friendly for certain applications. Version 8 ozone results show a better consistency between the two MIPAS measurement periods. They seem to be more realistic than preceding data versions in terms of long-term stability, as at least a part of the drift is corrected. Further, the representation of elevated stratopause situations is improved, but there is still some indication of a positive bias in the upper stratosphere.