Understanding The Global Sources And Sinks Of Atmospheric Carbonyl Sulfide In Order To Provide Insights Into Carbon Cycle Processes

This thesis focuses on the modelling and measurement of atmospheric OCS. Carbonyl sulfide (OCS) has emerged as a proxy for photosynthetic uptake of carbon dioxide (CO2) by plants. Multiple mechanistic and scaling approaches for inferring CO2 uptake, known as gross primary productivity (GPP), are emerging and the need for a global satellite dataset of OCS is crucial.

The atmospheric three-dimensional chemical transport model TOMCAT is used to model global atmospheric OCS. A leaf relative uptake (LRU) approach is applied to GPP fields in order to infer OCS vegetative uptake, estimated to be approximately 629 Gg S yr-1, which is at the lower end of estimates from previous studies (238 – 1335 Gg S yr-1). The remaining fluxes are extracted from the literature and scaled accordingly in order to close the OCS budget. A comparison is made with National Oceanographic and Atmospheric Administration - Earth System Research Laboratories (NOAA-ESRL) surface flask observations and the mean error reduction across all sites is approximately 25%, compared to a control simulation. TOMCAT is compared to profiles of OCS from the limb-sounding satellite instrument Atmospheric Chemistry Experiment – Fourier Transform Spectrometer (ACE-FTS), which show an excellent comparison throughout most of the atmosphere to within ±25 ppt, approximately 5% in the troposphere.

The University of Leicester IASI Retrieval Scheme (ULIRS) is adapted to retrieve profiles of OCS globally from atmospheric spectra recorded by the Infrared Atmospheric Sounding Interferometer (IASI), primarily over the oceans, OCSOCE. A spectrally resolved emissivity product is implemented in a first effort to improve land-based retrievals of OCS, OCSTER. Total column random errors are found to range between 18 and 38%. A comparison is made between OCSOCE and TOMCAT, agreeing to within ±5% throughout most of the tropics (30°S – 30°N). The seasonality of OCSOCE is compared to surface observations, showinga correlation coefficient of 0.92 with measurements at Mauna Loa, Hawaii. Agreement of OCSTER with TOMCAT to within ±5% is limited to 20°S – 20°N.

Finally, a flux inversion is performed by assimilating surface flask observations of OCS for 16 regions and 5 flux sectors, INVOCS. Mean INVOCS error reduction is 23% across all regions and the root mean square error is reduced by an average of 51% compared to the a priori mixing ratios. INVOCS shows a substantial shift in vegetative uptake from the tropics to the northern hemisphere extra tropics and the opposite for oceanic emissions. An independent comparison with ACE-FTS shows agreement within ±30 ppt throughout most of the troposphere, with a predominantly negative bias in the a priori mixing ratios, but positive bias in the a posteriori.