Physical inversion of the full IASI spectra: further validation and inter-comparison of O3 and OCS products

TL;DR

This study enhances the validation and inter-comparison of IASI satellite spectra inversion products for atmospheric gases, demonstrating high accuracy for ozone and good correlation for OCS with in situ data over the Pacific Ocean.

Contribution

It provides a comprehensive validation and inter-comparison of IASI O3 and OCS products against other measurements, improving confidence in satellite-based atmospheric gas retrievals.

Findings

IASI O3 products agree well with OMI, with 0.97 correlation and 1 Dobson Unit bias.

IASI OCS retrievals show 0.75 correlation with in situ data, capturing seasonal cycles.

A negative bias of -74 pptv for OCS is identified, linked to spectroscopic errors.

Abstract

Results from the physical, simultaneous, inversion of the full IASI (Infrared Atmospheric Sounder Interferometer) spectra applied to the retrieval of surface temperature and emissivity, temperature, water vapour, ozone and HDO vertical profiles, and gas column abundance of CO2, CO, CH4, SO2, N2O, HNO3, NH3, OCS and CF4 have been updated with respect to the validation of OCS and O3 products. A case study has been set up and analysed, which consists of a two years long record of IASI spectra over sea surface in the Pacific Ocean close to the Mauna Loa (Hawaii) validation station. IASI O3 products have been intercompared with OMI (Ozone Monitoring Instrument) O3 retrievals, which, in contrast to IASI that uses the thermal infrared, are based on observation in the near ultra violet region of the electromagnetic spectrum. IASI OCS retrievals are compared to in situ flask observations. The IASI vs OMI comparison shows an excellent agreement with a correlation of 0.97 and a relative bias of approximately 1 Dobson Unit. For the case of OCS, we have found a good correlation of 0.75 for night time soundings. IASI captures the OCS seasonal cycle, although a relatively large negative bias of approximately -74 pptv is seen, which can be explained with a large spectroscopic error for OCS in the HITRAN 2008 version.