Validation of the IASI FORLI/EUMETSAT ozone products using satellite (GOME-2), ground-based (Brewer-Dobson, SAOZ, FTIR) and ozonesonde measurements

This paper assesses the quality of IASI (Infrared Atmospheric Sounding Interferometer)/Metop-A (IASI-A) and IASI/Metop-B (IASI-B) ozone (O3) products (total and partial O3 columns) retrieved with the Fast Optimal Retrievals on Layers for IASI Ozone (FORLI-O3; v20151001) software for 9 years (2008-July 2017) through an extensive intercomparison and validation exercise using independent observations (satellite, ground-based and ozonesonde). Compared with the previous version of FORLI-O3 (v20140922), several improvements have been introduced in FORLI-O3 v20151001, including absorbance look-up tables recalculated to cover a larger spectral range, with additional numerical corrections. This leads to a change of ĝ1/4 4 % in the total ozone column (TOC) product, which is mainly associated with a decrease in the retrieved O3 concentration in the middle stratosphere (above 30 hPa/25 km). IASI-A and IASI-B TOCs are consistent, with a global mean difference of less than 0.3 % for both daytime and nighttime measurements; IASI-A is slightly higher than IASI-B. A global difference of less than 2.4 % is found for the tropospheric (TROPO) O3 column product (IASI-A is lower than IASI-B), which is partly due to a temporary issue related to the IASI-A viewing angle in 2015. Our validation shows that IASI-A and IASI-B TOCs are consistent with GOME-2 (Global Ozone Monitoring Experiment-2), Dobson, Brewer, SAOZ (Système d'Analyse par Observation Zénithale) and FTIR (Fourier transform infrared) TOCs, with global mean differences in the range of 0.1 %-2 % depending on the instruments compared. The worst agreement with UV-vis retrieved TOC (satellite and ground) is found at the southern high latitudes. The IASI-A and ground-based TOC comparison for the period from 2008 to July 2017 shows the long-Term stability of IASI-A, with insignificant or small negative drifts of 1 %-3 % decadeĝ'1. The comparison results of IASI-A and IASI-B against smoothed FTIR and ozonesonde partial O3 columns vary with altitude and latitude, with the maximum standard deviation being seen for the 300-150 hPa column (20 %-40 %) due to strong ozone variability and large total retrievals errors. Compared with ozonesonde data, the IASI-A and IASI-B O3 TROPO column (defined as the column between the surface and 300 hPa) is positively biased in the high latitudes (4 %-5 %) and negatively biased in the midlatitudes and tropics (11 %-13 % and 16 %-19 %, respectively). The IASI-A-To-ozonesonde TROPO comparison for the period from 2008 to 2016 shows a significant negative drift in the Northern Hemisphere of ĝ'8.6±3.4 % decadeĝ'1, which is also found in the IASI-A-To-FTIR TROPO comparison. When considering the period from 2011 to 2016, the drift value for the TROPO column decreases and becomes statistically insignificant. The observed negative drifts of the IASI-A TROPO O3 product (8 %-16 % decadeĝ'1) over the 2008-2017 period might be taken into consideration when deriving trends from this product and this time period.