Abstract/Summary

Stratospheric sulphate aerosol injection has been proposed as a potential way to cool the climate and alleviate some of the adverse impacts of climate change. Both prudent climate policy-making and successful geoengineering monitoring would require robust knowledge of the detectability of the geoengineering effects amid other externally forced changes and internal variability. The first part of this thesis investigates how we could detect the global-mean cooling effect of 5 Tg yr−1 SO2 injection with optimal fingerprint total least squares techniques in a climate model. Both the classical technique that assumes unforced variability as the null hypothesis and a non-stationary alternative could be applied to detect the global-mean cooling response within 10 years of geoengineering. However, the detectability of the geoengineering signal would depend on the filtering technique used for distinguishing the forced signals from climate noise. The second part of this thesis identifies the best spatial scales on Earth’s surface for geoengineering detection in temperature within 10 years of hypothetical deployment. The global scale would give the highest chance of early geoengineering detection, followed by the Northern Hemisphere and the Northern mid-latitudes. Detection of the geoengineering effect in other latitudinal bands and smaller regions would likely (>50% of the studied cases) take longer than 10 years. Compared to surface temperatures, using the vertical temperature structure that includes the tropospheric cooling and tropical lower stratospheric warming effects of the sulphate aerosols would increase the geoengineering detectability during the first 10 years of implementation on the global, hemispheric and tropical scales. Minimising the upper stratospheric (above 30 hPa) contribution to the vertical temperature structure could further improve the geoengineering detectability in the Tropics by 12.5%. These results suggest the possibility of efficient geoengineering monitoring with optimal fingerprint detection techniques, especially if the vertical temperature structure is included in the detection diagnostic.