Validation of the version 4.5 MAESTRO ozone and NO₂ measurements

Abstract

Launched aboard the Canadian SCISAT satellite in August 2003, the Measurement of Aerosol Extinction in the Stratosphere and Troposphere Retrieved by Occultation (MAESTRO) instrument has been measuring solar absorption spectra in the ultraviolet (UV) and visible part of the spectrum for more than 20 years. The UV-channel measurements from MAESTRO are used to retrieve profiles of ozone from the short-wavelength end of the Chappuis band (UV-ozone) and NO2, while measurements made in the visible part of the spectrum are used to retrieve a separate ozone (Vis-ozone) product. The latest ozone and NO2 profile products, version 4.5, have been released, and they initially cover the period from February 2004 to December 2023, although they will continue to be updated. The version 4.5 retrieval algorithm represents an improvement from previous versions, with changes including updated pressure and temperature input information, an improved algorithm for high-Sun reference spectrum calculation, improved Rayleigh scattering modelling, and the change to a Twomey–Tikhonov inversion algorithm from a Chahine relaxation technique. Due to the buildup of an unknown contaminant, the UV-ozone and NO2 products are only viable up to June 2009 for NO2 and December 2009 for UV-ozone. This study presents comparisons of the version 4.5 MAESTRO ozone and NO2 measurements with coincident (both spatially and temporally) measurements from an ensemble of 11 other satellite limb-viewing instruments. In the stratosphere, the Vis-ozone product was found to possess a small high bias, with stratosphere-averaged relative differences between 2.3 % and 8.2 %, although good agreement with the comparison datasets was found overall. A similar bias, albeit with slightly poorer agreement, is found for the UV-ozone product in the stratosphere, with the average stratospheric agreement between MAESTRO and the other datasets ranging from 2.8 % to 11.9 %. For NO2, general agreement with the comparison datasets is only found in the range from 20 to 40 km. Within this range, MAESTRO is found to have a low bias for NO2, and most of the datasets agree to within 27.2 %, although the average agreement ranges from 8.5 % to 43.4 %.