Impact of using a new ultraviolet ozone absorption cross-section dataset on OMI ozone profile retrievals

Abstract

We evaluate different sets of high-resolution ozone absorption cross-section data for use in atmospheric ozone profile measurements in the Hartley and Huggins bands with a particular focus on Brion-Daumont-Malicet et al. (1995) (BDM) currently used in our retrievals, and a new laboratory dataset by Birk and Wagner (BW) (2018). The BDM cross-section data have been recommended to use for retrieval of ozone profiles using spaceborne nadir viewing Backscattered UltraViolet (BUV) measurements since its improved performance was demonstrated against other cross-sections including Bass and Paur (1985) (BP) and those of Serdyuchenko et al (2014) and Gorshelev et al. (2014) (SER) by the 'Absorption Cross-Sections of Ozone' (ACSO) activity. The BW laboratory data were recently measured within the framework of the ESA project SEOM-IAS (Scientific Exploitation of Operational Missions - Improved Atmospheric Spectroscopy Databases) to provide an advanced absorption cross-section database. The BW cross-sections are made from measurements at more temperatures and in a wider temperature range than BDM, especially for low temperatures. Compared to BW, BDM cross-sections are positively biased from ~2 % at shorter UV to ~5 % at longer UV at warm temperatures. Furthermore, these biases dynamically increase by up to ± 40 % at cold temperatures due to no BDM measurements below 218 K. We evaluate the impact of using different cross-sections on ozone profile retrievals from Ozone Monitoring Instrument (OMI) measurements. Correspondingly, this impact leads to significant differences in individual ozone retrievals by up to 50 % in the tropopause where the coldest atmospheric temperature is observed. Bottom atmospheric layers illustrate the significant change of the retrieved ozone values with biases of 20 % in low latitudes, which is not the case in high latitudes because the ozone retrievals are mainly controlled by a priori ozone information in high latitudes due to less photon penetration down to the lower troposphere. Validation with ozonesonde observations demonstrates that BW and BDM retrievals show altitude-dependent bias oscillations of similar magnitude relative to ozonesonde measurements, much smaller than those of both BP and SER retrievals. However, compared to BDM, BW retrievals show significant reduction in standard deviation by up to 15 %, especially at the coldest atmospheric temperature. Such improvement is achieved mainly by th better characterization of the temperature dependence of ozone absorption