Tropospheric ozone retrieval from thermal infrared nadir satellite measurements: Towards more adaptability of the constraint using a self-adapting regularization. (November 2019)

• Record Type: Journal Article
• Authors: Eremenko, Maxim; Sgheri, Luca; Ridolfi, Marco; Cuesta, Juan; Costantino, Lorenzo; Sellitto, Pasquale; Dufour, Gaëlle
• Is Part Of: Journal of quantitative spectroscopy & radiative transfer. Volume 238(2019)
• Publication Date: 2019-11
• Subjects: Atmospheric composition -- Air pollution monitoring -- Ozone -- Passive remote sensing -- Inverse problems; Spectrum analysis -- Periodicals; Radiation -- Periodicals; Analyse spectrale -- Périodiques; Rayonnement -- Périodiques; Radiation; Spectrum analysis; Periodicals; 543.0858
• DOI: 10.1016/j.jqsrt.2019.106577
• Languages: English
• ISSNs: 0022-4073

Abstract

Highlights: The study presents a retrieval method developing a variable in altitude constraint that adapts for each individual infrared nadir measurement. This retrieval approach performs especially well in situations where standard algorithms may have difficulties according to the validation studies. The retrievals are tested on the synthetic observations representing the future IASI-NG satellite measurements. Abstract: We developed a Self-Adapting Constraint Retrieval Scheme (SACRS) to retrieve ozone profiles from nadir infrared satellite measurements. In this algorithm, the constraint is variable in altitude and adapted automatically for each individual measurement. The algorithm is tested on synthetic observations representing the future IASI-NG satellite observations and considering either ozonesonde measurements or chemistry-transport model ozone simulations to represent the true ozone (pseudo-reality). The ozone retrievals are evaluated mainly for the troposphere with a specific focus on the lower troposphere between the surface and 6 km. Compared to a previous algorithm based on a fixed constraint retrieval scheme (FCRS), the biases, correlation and error estimates are improved with the SACRS. The bias is reduced by 40% and the correlation coefficient increases from 0.72 to 0.80. The SACRS algorithm also leads to an enhanced sensitivity in the lower troposphere with degrees of freedom for signal up to 0.83, increased by 11% compared to the FCRS. The SACRS performs especially well where current algorithms usually fail, namely for polar and tropical air masses. The bias is reduced from 8.6% to 0.5% in the troposphere (surface-9 km) when considering polar cases and from 24.4% to 10.1% in the upper troposphere - lower troposphere column (12–18 km) in the tropics.