Evaluation of IASI total ozone accuracy by comparisons with SAOZ and all other satellites measurements in the tropics and at mid-latitudes
Abstract
High accuracy global total ozone measurements from space are key data for evaluating the
amplitude of ozone depletion, understanding its contribution and consequences on climate change
and predicting the future evolution of both ozone and climate. Long series of total ozone data are
available, but showing still systematic significant differences between them. Here most recent data
available from IASI A and B are compared to SAOZ NDACC UV-Vis ground-based measurements
at several stations in the tropics and mid-latitudes, as well as all other satellites data available since
the beginning of IASI operations in 2007. The list of satellites includes GOME, SCIAMACHY,
GOME2 and OMI reprocessed with the GODFIT V3 (GOME-Type Direct FITing) algorithm
developed in the frame of the ESA Ozone Climate Initiative (Lerot et al., 2014), the NASA SBUV
v8.6 (McPeters et al., 2013), the NASA-AURA-OMIT and OMID (Levelt et al, 2007).
According to these comparisons, the IASI A-B ozone columns are systematically overestimated by
2 - 6%, depending on the station and the satellite, and the difference between all of them is showing
more or less seasonality of amplitude depending also on the satellite and the station.
Mean biases are attributed to smaller ozone absorption cross-sections in the IR compared to the UV
and visible and also to longitudinal ozone profiles variations ignored in most satellite retrievals.
When corrected for mean biases, the seasonality of the differences between IASI, SAOZ and most
others is generally smaller than ± 1% that is of the order of magnitude of the dispersion and thus
random uncertainty of the measurements, except in a few specific cases:
a) the presence of a larger 2-3% peak IASI-Sat difference in the monthly mean in May at OHP on
both IASI A and B suggesting a retrieval error on all IASI data on that month but of totally
unknown origin;
b) a larger seasonality of the 4 CCI GOFDIT processed satellites compared to IASI and others
suggesting a slight systematic error related to the use of zonal mean O3 profiles climatologies
ignoring longitudinal variations;
c) larger dispersion, seasonality and moreover opposite seasonality phase of NPP at southern midlatitudes
indicative of an error in those data files in the southern hemisphere;
d) and, finally, an overestimation of IASI ozone in Reunion Island in coincidence with hurricanes,
suggesting that high altitude clouds are not properly removed in the IASI data sets.
Shown in the presentation will be the illustration of the IASI- SAOZ- Sat differences at all SAOZ
stations and possible explanations when significant differences are observed.