Subduction kinematics and dynamics in 3D space: Insight from laboratory and numerical modelling

W. P. Schellart 1 , J. Freeman 1 , D. R. Stegman 2 , L. Moresi 2 & D. May 2

1 Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia
2 School of Mathematical Sciences, Monash University, Melbourne, VIC 3800, Australia

During development of the plate tectonic theory in the 1960's and 1970's it was soon realized that the motion of the tectonic plates, which cover the Earth's surface, is primarily driven by dense subducted slabs located in the mantle. Subducted slabs form along convergent plate boundaries where an oceanic plate is thrust underneath an overriding plate into the Earth's mantle. The slabs are denser than the ambient mantle, forcing the slab to sink and pulling the trailing plate at the surface into the mantle, thus driving plate motion at the surface and flow in the Earth's mantle. The kinematics and dynamics of this sinking process and the flow that is induced by the sinking has been investigated extensively in two-dimensional space, but three-dimensional investigations are still in their infancy . Three-dimensional investigations are important, because subduction zones on Earth are limited in trench-parallel extent (~200-7000 km) and most are curved. We have recently attempted to address this deficiency by doing laboratory and numerical simulations of progressive subduction in 3D space [Schellart, 2004; Stegman et al., 2006]. These works have shown that in self-consistent dynamic models, where plate motions and trench migration are not imposed but evolve naturally, subducting slabs and associated trenches predominantly retreat oceanward, inducing flow in the mantle that is toroidal in nature and located exclusively around the lateral edges of the slab. We observed no poloidal flow around the slab tip, as was suggested by previous workers. The (non-) existence of poloidal flow around the slab tip is important for understanding the relationship between slab width and trench velocity. Current and previous works have shown that there is an inverse relationship between slab width and trench retreat velocity: the wider the slab, the slower the trench retreat velocity [Schellart, 2004; Stegman et al., 2006; work in progress]. Such a relationship is most easily explained if in fact, there is no (or negligible) trench rollback-induced poloidal flow around the slab tip. To address this issue in more detail, new laboratory and numerical models have been run to study the subduction-induced flow patterns in the mantle. Our models confirm previous observations that rollback-induced flow of the mantle occurs in a toroidal fashion and occurs exclusively around the lateral edges of the slab (Figs. 15, 16).

Figure 15. 3D view of numerical simulation of subduction illustrating slab geometry and toroidal flow patterns in the mantle (arrows) at 200 km depth.

Figure 16. Side view of laboratory simulation (composite image of nine photographs) of subduction illustrating slab geometry and poloidal flow patterns in the mantle in the centre of the subduction zone.

 

References: Schellart, W.P., Kinematics of subduction and subduction-induced flow in the upper mantle. Journal of Geophysical Research 109 , B07401, doi:10.1029/2004JB002970, 2004.

Stegman, D., J. Freeman, W.P. Schellart, L. Moresi and D. May, Influence of trench width on subduction hinge retreat rates in 3-D models of slab rollback. Geochemistry Geophysics Geosystems 7 , Q03012, doi:10.1029/2005GC001056, 2006.