Time series of mesospheric species measured using ground-based microwave spectrometry: Retrieval and Error estimation

Sammanfattning: As the human emission of greenhouse gases continue to increase, the atmosphere is changing. The mesosphere region is particularly interesting as the magnitude of temperature change in this region is expected to be larger, than in the lower layers of the atmosphere. However, there is still much uncertainty related to the long-term effects of these changes. It is therefore important to have instruments that can provide long-term monitoring of the dynamics and chemistry of the middle atmosphere. Species with long photochemical lifetime can be used to study the dynamics of the mesosphere. At the Onsala Space Observatory two such species, CO and H2O, have been measured over a number of years with microwave spectrometers. These instruments can provide us with continuous measurements of the mesosphere over long time periods. This thesis concerns characterising and improving the measurements and retrievals from the two microwave spectrometers. This thesis puts forth a possible improvement in the inversion algorithm used for retrieving time-series from ground-based microwave spectrometers. By using a two dimensional retrieval method, the inversions can take into account the temporal correlation of the atmospheric state. The new method was tested on the 22 GHz spec- trometer measuring mesospheric water vapour at OSO. The result is a more flexible retrieval that removes the problem of ad-hoc selecting data averaging times before performing an inversion. An additional improvement offered by the new method is the ability to consistently interpolate the retrieved data in order to close measurement gaps or re-grid the data. The thesis also compares a six year long time series of mesospheric CO from the 115 GHz spectrometer at OSO to measurements from contemporary satellite instruments. The systematic errors of the instrument are estimated to ±40 % from 2002-2004 and ±20 % from 2004-2008. Compared to the satellite measurements, the measurements from the OSO radiometer have a mean difference of less than or equal to 25 %. This is consistent with the combined systematic errors of both the ground-based and satellite instruments. The thorough error characterisation, and the improved retrieval method, presented in this thesis ensure that the data produced by the microwave spectrometers at OSO remain relevant and important for the future study of the middle atmosphere.