Scientific theories of how subduction and plate tectonics began on Earth--and what the tectonic structure of Earth was before this--remain enigmatic and contentious. Understanding viable scenarios for the onset of subduction and plate tectonics is hampered by the fact that subduction initiation processes must have been markedly different before the onset of global plate tectonics because most present-day subduction initiation mechanisms require acting plate forces and existing zones of lithospheric weakness, which are both consequences of plate tectonics. However, plume-induced subduction initiation could have started the first subduction zone without the help of plate tectonics. Here, we test this mechanism using high-resolution three-dimensional numerical thermomechanical modelling. We demonstrate that three key physical factors combine to trigger self-sustained subduction: (1) a strong, negatively buoyant oceanic lithosphere; (2) focused magmatic weakening and thinning of lithosphere above the plume; and (3) lubrication of the slab interface by hydrated crust. We also show that plume-induced subduction could only have been feasible in the hotter early Earth for old oceanic plates. In contrast, younger plates favoured episodic lithospheric drips rather than self-sustained subduction and global plate tectonics.

S U M M A R YA preexisting weakness zone in the lithosphere is required to initiate subduction. Here, we focus on a new type of weakness zone, a Subduction-Transform Edge Propagator (STEP) fault, which is inherited from a tear along the edge of a slab. Using coupled thermal-mechanical models, we show that STEP fault-perpendicular convergence results in a dipping shear zone in any tectonic setting. At a continental margin, this shear zone dips towards the continent, which is an excellent starting condition for ocean-continent subduction. If (far field) convergence persists, STEP faults become new subduction boundaries. The trench moves landward during the earliest stages of convergence. When slab pull becomes a dominant driving force, after ∼80 km convergence, trench roll-back commences. The initial geometry and mechanical properties of the sub-crustal STEP fault zone affect the results; subduction initiation is facilitated by a wide (∼100 km) and low-viscosity weakness zone. Incipient subduction is easier for young oceanic lithosphere due to its lower flexural rigidity and is insensitive to the far field convergence rate. As STEP faults are commonly associated with young oceanic lithosphere, subduction initiation is thus relatively easy along them. Of particular interest are continent-ocean margins where STEP faulting has occurred.

No abstract

Initiation of subduction following the impingement of a hot buoyant mantle plume is one of the few scenarios that allow breaking the lithosphere and recycling a stagnant lid without requiring any preexisting weak zones. Here, we investigate factors controlling the number and shape of retreating subducting slabs formed by plume‐lithosphere interaction. Using 3‐D thermomechanical models we show that the deformation regime, which defines formation of single‐slab or multi‐slab subduction, depends on several parameters such as age of oceanic lithosphere, thickness of the crust and large‐scale lithospheric extension rate. Our model results indicate that on present‐day Earth multi‐slab plume‐induced subduction is initiated only if the oceanic lithosphere is relatively young (<30–40 Myr, but >10 Myr), and the crust has a typical thickness of 8 km. In turn, development of single‐slab subduction is facilitated by older lithosphere and pre‐imposed extensional stresses. In early Earth, plume‐lithosphere interaction could have led to formation of either episodic short‐lived circular subduction when the oceanic lithosphere was young or to multi‐slab subduction when the lithosphere was old.

The classical Wilson Cycle concept, describing repeated opening and closing of ocean basins, hypothesizes spontaneous conversion of passive continental margins into subduction zones. This process, however, is impeded by the high strength of passive margins, and it has never occurred in Cenozoic times. Here using thermomechanical models, we show that additional forcing, provided by mantle flow, which is induced by neighboring subduction zones and midmantle slab remnants, can convert a passive margin into a subduction zone. Models suggest that this is a long‐term process, thus explaining the lack of Cenozoic examples. We speculate that new subduction zones may form in the next few tens of millions of years along the Argentine passive margin and the U.S. East Coast. Mantle suction force can similarly trigger subduction initiation along large oceanic fracture zones. We propose that new subduction zones will preferentially originate where subduction zones were active in the past, thus explaining the remarkable colocation of subduction zones during at least the last 400 Myr.