The warm and cold water route ocean gateways are important oceanographic locations with respect to global climate. By the advection of salinity anomalies to the North Atlantic, changes at these ocean gateways are suggested to affect the strength and geometry of the Atlantic Meridional Overturning Circulation (AMOC). Adjustments of the AMOC can play a crucial role in the Earth‘s climate, and are suggested to be related to climate changes of the Late Pleistocene. However, little is known about the role of the warm and cold water route gateways on past glacial-interglacial and millennial time scales. This thesis documents a study which utilises Earth system modelling, combined with analyses of climate proxy data, to investigate the behaviour of these ocean gateways during changing climates of the past. The development of an adapted Earth System Model shows that it is possible to improve the simulation of climatological mean transport rates through the Indian-Atlantic Ocean Gateway (I-AOG, warm water route gateway) and Drake Passage (cold water route gateway). These ocean gateway transports are often overestimated in contemporary state-of-the-art climate models. The adapted model therefore provides a solid platform for the application of palaeo boundary conditions and the investigation of these ocean gateways during the past. Comparing pre-industrial and Last Glacial Maximum (LGM) climate simulations reveals that the I-AOG transport was only moderately weaker during the LGM, contrasting against the general hypothesis inferred from proxy data. A new hypothesis is developed which can consolidate these results. Supported by proxy data, the modelled Drake Passage throughflow is substantially weaker during the LGM. This might have potential implications for the geometry of the relatively shallow glacial AMOC. In response to freshwater perturbation - mimicking Heinrich events - an increase in Drake Passage throughflow is simulated, simultaneous with only weak changes of the I-AOG transport. The response at both locations is supported by proxy data, and suggests that the Drake Passage might be more important than previously considered with respect to the mechanisms involved in the abrupt resumption of the AMOC to interstadial conditions.