Causal links of past climate change in Coupled Model Intercomparison Project Phase 5 climate models

Abstract: The climate system is influenced by various external forcings (e.g. volcanic forcing, solar forcing and change of greenhouse gas concentrations) and its own internal climate variability. While evaluating the causes of past climate change is important to expand our knowledge of the impacts of different external forcings, understanding the coupling between major internal climate drivers is crucial for the prediction of future climate. This thesis uses the Granger causality test to investigate the impacts of external forcings on subdecadal variability of regional near-surface air temperature (SAT) during the past millennium (period 850-1850). In addition, the role of solar forcing on low-frequency variability of the North Atlantic is examined and the two-way causal relationship between the El Niño-Southern Oscillation (ENSO) and the Indian Ocean Dipole (IOD) is evaluated. The investigations are mainly based on the output from the Coupled Model Intercomparison Project Phase 5 (CMIP5) simulations.The results of this thesis support the conclusion for robust impact of volcanic forcing on SAT during preindustrial times of the past 1000 years. Besides, the results indicate that there is solar signal in SAT response of the tropical and subtropical regions whereas the influence of greenhouse gases radiative forcing on regional SAT is weak and uncertain. In addition, solar forcing might have an influence, at multi-centennial time scales, on the North Atlantic winter sea surface temperatures while the connection between solar forcing and the North Atlantic Oscillation is weak in the models. The results for the investigation of the causal impact of ENSO on the IOD show that this impact is weaker in numerous models than the result in reanalysis datasets, whereas the influence of the IOD on ENSO is robust for most models and reanalysis datasets.Overall, the findings in this thesis contribute to the knowledge of the influence of external forcings on regional climate change and help us to better understand the two-way causal relationship between ENSO and the IOD.