# The slab detachment benchmark#

This section was contributed by Cedric Thieulot and Anne Glerum.

Slab detachment (slab break-off) may occur in the final stages of subduction as a consequence of the combination of a buoyant crust and strong slab pull. It is often invoked to explain geophysical and geological observations such as tomographic images of slab remnants and exhumed ultra-high-pressure rocks .

This benchmark is based on the setup by S. Schmalholtz , which was subsequently run with by A. Glerum . The computational domain is a $$1000 \text{ km}\times 660 \text{ km}$$ box. No-slip boundary conditions are imposed on the sides of the system, while free-slip boundary conditions are imposed at the top and bottom.

Two materials are present in the domain: the lithosphere and the mantle as shown in Fig. 186. The gravity acceleration is Earth-like with $$g=9.81 \text{ m}\text{ s}^2$$. The overriding plate is $$80\text{ km}$$ thick and is placed at the top of the domain. The already subducted lithosphere extends vertically into the mantle for $$250 \text{ km}$$. This slab has a density $$\rho_s=3300\text{ kg/m}^3$$ and is characterized by a power-law flow law so that its effective viscosity depends on the square root of the second invariant of the strainrate $$\dot\varepsilon$$:

$\eta_{eff} = \eta_0 \, \dot\varepsilon^{1/n-1}$

with $$n=4$$ and $$\eta_0=4.75\times 10^{11}\text{ Pa . s}$$. The mantle occupies the rest of the domain and has a constant viscosity $$\eta_m=1\times 10^{21}\text{ Pa . s}$$ and a density $$\rho_m=3150\text{ kg/m}^3$$. Viscosity is capped between $$1\times10^{21}\text{ Pa . s}$$ and $$1\times 10^{25} \text{ Pa . s}$$. Fig. 187 shows the various fields and their evolution through time. As shown in Glerum et al. [2018], Schmalholz [2011] the hanging slab necks, helped by the localizing effect of the nonlinear rheology. Model results were shown to compare favorably to the results of Schmalholz [2011] in Glerum et al. [2018], Hillebrand et al. [2014] and the effect of viscosity and material averaging was explored in .