Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.

Subduction is one of the main components of the Earth's plate-tectonic engine. Where two plates converge, the coolest and densest plate will slide below the other and sink into the underlying viscous mantle, taking with it basaltic ocean crust, sediments and fluids. The recycling of fluids and sediments lies at the root of the Earth's most explosive volcanism, and is crucial in the formation of continental crust, and production and concentration of ores. Subduction also produces the world's largest earthquakes. However, while in some zones the two plates appear strongly coupled and motions give rise to mega earthquakes, others converge without shaking. Plate coupling may also determine why along some subduction zones, mountain belts like the Andes and Rockies formed, while behind others, new oceans, like the Philippine Sea and Fiji Basin opened. Volcanic arc positions are controlled by how steeply the plate descends into the mantle. The shape of the downgoing plate in the mantle also affects how easily it can sink into the deep mantle. In spite of the importance of these subduction characteristics, at present, we do not understand what forces govern plate coupling, subducting plate shape or subduction motions. The gravitational pull from the dense sinking plates is generally considered to be the dominant driving force of plate tectonics. However, neither observed motions at subduction zones, nor downgoing plate shape, nor upper plate deformation correlate with the density of the downgoing plate. It has been proposed that forcing by the overriding plate or the strength of the downgoing plate can overrule the effects of downgoing plate density. However, we recently developed a two-dimensional model of purely density driven subduction, which demonstrated that the expressions of downgoing plate density can be counterintuitive. For example, we find that young light plates can subduct faster than old dense plates and what is more they often do (Goes et al., Nature 2008). This discovery illustrates our lack of understanding of subduction forces. Here we propose a comprehensive investigation that combines numerical modelling with observations, to explore how three-dimensional variation in downgoing plate density and strength determines subduction behaviour. In the first part of the project, we will systematically document the sensitivity of plate motions and downgoing plate shape to spatial and temporal variations in plate structure. In the second part, we will run a set of models for the Pacific 'Ring of Fire', location of the world's largest subduction zones. Our models of Pacific subduction will be driven by densities from the best-constrained history of plate ages. Where our modelled subduction behaviour is consistent with observed present-day downgoing plate shape, and the history of plate motions, downgoing plate density exerts the dominant control; elsewhere, additional forces must play a role.