Interannual variability and predictability of the Indian Summer Monsoon - El Niño Southern Oscillation system

El Niño-Southern Oscillation (ENSO) and the Indian Summer Monsoon (ISM) are two of the most energetic and influential climatic phenomena on the planet. Although they originate in the tropical Indo-Pacific region, they can extend their reach well beyond, through atmospheric teleconnnections that can affect patterns of climate variability worldwide. This thesis takes part in a global effort to improve our understanding of the potential predictability of ISM rainfall and ENSO, by exploring the large-scale teleconnections associated with the whole monsoon-ENSO system on interannual timescales, as well as the role played by leading modes of coupled variability, particularly in tropical and extratropical parts of the Indo-Pacific region. Based on statistical diagnoses of observations before and after the 1976-77 climate shift, and numerical experiments with the SINTEX-F coupled model, our results highlight the importance of mid-latitudes in the predictability of the ISM-ENSO system. Indeed, significant and robust precursors are identified in the North Pacific and South Indian Oceans during the previous boreal winter/early spring, and have the potential to predict ISM and ENSO events with longer lead-times than their traditional tropical predictors. Besides, this predictability is stronger for the 'late' ISM rainfall during August-September. This second part of the ISM season is also characterized by the occurrence of ocean-atmospheric processes in the Indian Ocean, which compete with the effect of ENSO on the monsoon. The enhancement of these local feedbacks could explain the apparent weakening of the seasonal ENSO-monsoon relationship observed in recent decades.