Most published subduction modelling and much palaeotectonic speculation incorporate the false assumption that subducting oceanic plates slide down fixed slots. In fact, hinges roll back into oceanic plates and slabs sink more steeply than the inclinations of the Benioff zones which define transient positions of the slabs. The lower parts of overlying mantle wedges sink with the slabs, pulling away from partial-melt zones higher in the wedges. The complex behaviour of arc systems can be comprehended in terms of this mechanism of subduction. The common regime in overriding plates is extensional, and leading edges are crumpled only in collisions. Shear coupling between subducting slabs and overriding plates is limited to shallow depths and varies widely, with corresponding variations in tectonic erosion, accretion, and regurgitation of high- P subducted materials. Arcs can advance, lengthen, change curvature, festoon around obstacles, rotate while deforming, and fold and pinch shut. Two arcs can collide as an intervening oceanic plate is subducted simultaneously beneath both, or they can migrate apart as new lithosphere is formed between them. Subduction cannot occur simultaneously beneath opposite sides of a rigid plate because impossible retrograde subduction would be required beneath one of them. Histories, including inception ages, collisions, polarity reversals and stage of petrological evolution, vary greatly along continuous arc systems. Long-continuing steady-state systems are uncommon. Magmatic arcs are properly viewed as features migrating with sinking lower plates, not as fixed features of upper plates. Hot inclined zones within mantle wedges, midway between sinking slabs and overriding crust, are avenues for replenishment of mantle pulled away with subducting plates and also are sites of generation of arc protomelts as volatiles rise into them from dehydrating slabs. Back-arc basins form by spreading behind migrating arcs; strips of arcs may be abandoned in the spreading systems. An arc can migrate so rapidly that it plates out oceanic lithosphere rather than producing a welt. Exposed sections of the upper mantle and basal crust of arcs show that the Mohorovičić discontinuity is a self-perpetuating density filter and that the already-evolved basaltic and melabasaltic melt that leaves the mantle forms great basal-crust sheets of norite, gabbro and granulite. All more-evolved rock types in these sections are generated in the crust by fractionation, secondary melting and contamination (and this falsifies much petrological modelling).
No takes yet. Share an insight, caveat, or question.
Warren Hamilton (1994) studied this question.
Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context: