Recovery and validation of Odin/SMR measurements of mesospheric CO and H2O

Abstract: Although life on Earth is concentrated in the troposphere, not only the processes happening here are of interest, since tropospheric processes are not the only ones that have an impact on the climate. Coupling mechanisms between different layers are such that changes in middle atmospheric dynamics and composition have an effect on what happens at lower altitudes. Therefore there is the need to extend climate models to include higher altitudes and to perform measurements of the middle atmosphere. Among the species that can be observed in the middle atmosphere, carbon monoxide (CO) and water vapour (H2O) are of high scientific interest thanks to their aptness to being used as circulation tracers. This is due to their long photochemical lifetime which is longer or equal to zonal, meridional and vertical transport time scales. Moreover, H2O plays a decisive role for O3 chemistry. In the mesosphere, meridional circulation is driven by momentum deposition from breaking gravity waves. This originates an annual cycle which affects mostly high latitudes: with measurements it can be observed how CO-rich dry air descends towards the lower mesosphere at the winter pole while moist and CO-poor air is uplifted towards upper mesosphere and lower thermosphere at the summer pole. Gravity waves also give rise to the Semi- Annual Oscillation (SAO) in zonal winds resulting in concentration oscillations that can be observed in CO and H2O measurements, mostly at the tropics. Among the satellite instruments currently performing remote sensing of the middle atmosphere, the Sub-Millimetre Radiometer (SMR) on board the Odin satellite is one of the most long-lived. In fact, SMR has been performing limb sounding of the middle atmosphere for 19 years, providing valuable long term datasets. However, the CO and H2O datasets are both affected by instrumental artifacts that resulted in a misestimation of the two species’ concentration. The two papers included in this thesis focus on identifying the causes and correcting such artifacts. CO observation modes were affected by a malfunctioning of the Phase-Lock Loop (PLL) of the local oscillator (LO) which was causing frequency shifts in the spectra and line broadening. An algorithm that shifts the line back to its theoretical position has been developed and the line broadening has been quantified and taken into account when performing new retrievals. Regarding the H2O observation modes, underestimation of sideband leakage resulted in artifacts in the spectra that caused the misestimation of the retrieved concentration and temperature. For this reason SMR agreed poorly with other instruments measurements. An improvement was brought by assuming a larger sideband leakage than previously thought. For both species new inversions have been performed using the Optimal Estimation Method (OEM) with the Atmospheric Radiative Transfer Simulator (ARTS) as forward model. The recovery and correction of these products resulted in two new long-term and global data sets that are now available to the scientific community to study middle atmospheric dynamics.