Retrieval of stratospheric ozone profiles from MIPAS/ENVISAT limb emission spectra: a sensitivity study.

We report on the dependance of ozone volume mixing ratio profiles, retrieved from spectra of the Michelson Interferometer for Passive Atmospheric Sounding (MIPAS), on different retrieval setups such as different a-priori profiles, regularization strengths and spectral regions used for analysis. MIPAS is a spaceborne limb-viewing Fourier transform infrared (FTIR) emission spectrometer, by which vertical profiles of various trace gases can be measured simultaneously. Purpose of this investigation is to check and to optimize the current retrieval setup. The choice of different a-priori profiles, of a different approach to retrieve the continuum radiation, and of a weaker regularization than in the reference data version (V2_O3_2) causes only small to moderate deviations of up to ±0.2, -0.3 and ±0.5 ppmv, respectively, in the retrieved ozone volume mixing ratios below 60 km altitude. Use of different microwindow sets optimized for polar, mid-latitude and tropical conditions results in deviations of up to ±1.5 ppmv in the altitude region of the ozone maximum, exceeding the total estimated retrieval error of 0.65 ppmv (polar regions) -1.2 ppmv (tropics) in this height region. Therefore, to avoid latitudinal artefacts, one fixed set of microwindows is considered more appropriate for retrieval of a whole orbit rather than a latitude-dependent microwindow selection. For this task the microwindow set optimized for the polar atmosphere was found to be better suitable than its midlatitude and tropical counterparts. The results from the different microwindow sets, which variably cover MIPAS spectral bands A (685-970 cm-1) and AB (1020-1170 cm-1), indicated a positive bias of up to 1 ppmv between the ozone maxima retrieved from the ozone emission in MIPAS band AB only and from combined analysis of MIPAS bands A and AB. Further investigations showed that this discrepancy can be caused by a bias between the radiance calibration of level-1B spectra of bands A and AB or by a bias between the spectroscopic data used in bands A and AB. [ABSTRACT FROM AUTHOR]