How sphericity combines with the age and width of slabs to dictate the dynamics and evolution of subduction systems on Earth

The dynamics and evolution of subduction systems has been extensively studied through computational modelling. Most existing models have been executed in a Cartesian domain and, as such, the potential role of Earth's spherical geometry in modulating the evolution of subduction systems remains poorly explored. Here, we simulate multi-material free-subduction in a 3-D spherical domain, to investigate the effect of plate age (through thickness and density) and width on the evolution of subduction systems. To isolate effects arising from sphericity, we compare results with equivalent models in a Cartesian domain. Our spherical models yield results that are consistent with existing Cartesian studies: (i) subducting slabs retreat more and subduct at a shallower angle as plate age increases, due to increased slab strength and slab pull, which contribute to higher bending resistance and sinking rate; and (ii) wider slabs develop along-strike variations in trench curvature, trending towards a 'W' shape with increasing width, due to toroidal flow at the slab edges. We find, however, that these along-strike variations are restricted to older slabs that are able to drive trench retreat. When comparing spherical and Cartesian models, we find that: (i) subducting slabs descend faster in a spherical domain, since downwelling material converges to concentrate slab buoyancy; (ii) these faster descent rates reduce the time available for bending at the trench, resulting in effectively stronger slabs; (iii) the curvature of slabs on a sphere increases their effective strength relative to straight slabs in a Cartesian domain; and (iv) the curvature of the transition zone in spherical models enhances slab stagnation at 660 km depth. These differences become more prominent as slab widths increase, with predicted slab morphologies and trench curvatures differing significantly between Cartesian and spherical models for trenches that are wider than 2000 km. Taken together, our results suggest that Cartesian models are suitable for investigating the dynamics of narrow subduction zones. However, spherical models should be utilised when exploring the dynamics and evolution of subduction zones wider than ~2000 km: at such length-scales, the effects of Earth's curvature become too significant to ignore.