Continental collision and slab break-off: A comparison of 3-D numerical models with observations

Conditions and dynamics of subduction-collision and subsequent 3-D slab break-off and slab tear propagation are quantified, for the first time, using fully dynamic numerical models. Model results indicate that collision after the subduction of old, strong subducting oceanic slab leads to slab break-off at 20-25 Myr after the onset of continental collision, and subsequently a slab tear migrates more or less horizontally through the slab with a propagation speed of 100-150 mm/yr. In contrast, young, weak oceanic slabs show the first break-off already 10 Myr after continental collision, and can experience tear migration rates up to 800 mm/yr. Slab strength plays a more important role in the timing of slab break-off and the speed of a propagating slab tear than (negative) slab buoyancy does. Slab break-off is viable even for slabs that are supported by the viscosity jump and phase change between the upper and lower mantle.

The density of the oceanic slab and the subducting continental block is important for the amount of continental subduction and the depth of slab break-off. A 40-km thick continental crust can be buried to depths greater than 200 km, although this maximum depth is significantly less for younger or very weak slabs, or thicker continental crust. Slab break-off typically starts at a depth of 300 km, mostly independent of mantle rheology, but, like continental crustal burial, can be shallower for young or buoyant plates. Our 3-D models illustrate how, due to the difference in necking in 2-D and 3-D, break-off has an intrinsic small preference to start as a slab window within the slab's interior, rather than as a slab tear at the slab edge. However, any significant asymmetry in the collision setting, e.g. earlier collision at one end of the subduction zone, would override this, and leads to slab tearing starting near one edge of the slab.

These results put important new constraints on the dynamics of the collision and subsequent slab break-off for modern collision belts. For a proposed timing of the initial Arabia-Eurasia collision at 35 Ma, break-off of the ~ 200-Myr-old Neo-Tethys slab is unlikely to have occurred before 15-10 Ma. Furthermore, our results illustrate that shallow, early break-off of weak slabs provides a viable explanation for the absence of blueschists and ultra-high pressure metamorphism in the Precambrian geological record.