International audienceSubduction infancy corresponds to the first few million years following subduction initiation, when slabs start their descent into the mantle. It coincides with the transient (yet systematic) transfer of material from the top of the slab to the upper plate, as witnessed by metamorphic soles welded beneath obducted ophiolites. Combining structure–lithology–pressure–temperature–time data from metamorphic soles with flow laws derived from experimental rock mechanics, this study highlights two main successive rheological switches across the subduction interface (mantle wedge vs. basalts, then mantle wedge vs. sediments; at ∼800 °C and ∼600 °C, respectively), during which interplate mechanical coupling is maximized by the existence of transiently similar rheologies across the plate contact. We propose that these rheological switches hinder slab penetration and are responsible for slicing the top of the slab and welding crustal pieces (high- then low-temperature metamorphic soles) to the base of the mantle wedge during subduction infancy. This mechanism has implications for the rheological properties of the crust and mantle (and for transient episodes of accretion/exhumation of HP-LT rocks in mature subduction systems) and highlights the role of fluids in enabling subduction to overcome the early resistance to slab penetration

International audienceMetamorphic soles are tectonic slices welded beneath most large-scale ophiolites. These slivers of oceanic crust metamorphosed up to granulite facies conditions are interpreted as forming during the first million years of intra-oceanic subduction following heat transfer from the incipient mantle wedge towards the top of the subducting plate. This study reappraises the formation of metamorphic soles through detailed field and petrological work on three key sections from the Semail ophiolite (Oman and United Arab Emirates). Based on thermobarometry and thermodynamic modelling, it is shown that metamorphic soles do not record a continuous temperature gradient, as expected from simple heating by the upper plate or by shear heating as proposed in previous studies. The upper, high-temperature metamorphic sole is subdivided in at least two units, testifying to the stepwise formation, detachment and accretion of successive slices from the down-going slab to the mylonitic base of the ophiolite. Estimated peak pressure-temperature conditions through the metamorphic sole, from top to bottom, are 850°C and 1 GPa, 725°C and 0.8 GPa and 530°C and 0.5 GPa. These estimates appear constant within each unit but differing between units by 100 to 200°C and ~0.2 GPa. Despite being separated by hundreds of kilometres below the Semail ophiolite and having contrasting locations with respect to the ridge axis position, metamorphic soles show no evidence for significant petrological variations along strike. These constraints allow us to refine the tectonic–petrological model for the genesis of metamorphic soles, formed via the stepwise stacking of several homogeneous slivers of oceanic crust and its sedimentary cover. Metamorphic soles result not so much from downward heat transfer (ironing effect) as from progressive metamorphism during strain localization and cooling of the plate interface. The successive thrusts originate from rheological contrasts between the sole, initially the top of the subducting slab, and the peridotite above as the plate interface progressively cools. These findings have implications for the thickness, the scale and the coupling state at the plate interface during the early history of subduction/obduction systems