Bending curvatures of subducting plates: old versus young slabs

By combining scaled laboratory experiments and numerical simulations, this study presents a quantitative analysis of the bending radius (R-B) of subducting slabs within the upper mantle, taking into account the effects of age (A). Based on a half-space cooling model, we constrain the density (rho), viscosity (eta) and thickness (h) of slabs as a function of A, and develop representative models to estimate R-B for different A. Laboratory subduction models produce visually contrasting bending curvatures for young (A = 10 Ma), intermediate (A = 70 Ma) and old (A = 120 Ma) slabs. Young slabs undergo rollback, resulting in a small bending radius (scaled up R-B similar to 150 km), whereas old slabs subduct along a uniformly dipping trajectory with large bending radius (R-B similar to 500 km). Equivalent real scale computational fluid dynamic simulations reproduce similar bending patterns of the subducting slabs, and yield R-B versus A relations fairly in agreement with the laboratory results. We balance the buoyancy driven bending, flexural-resistive moments and viscous flow induced suction moment to theoretically evaluate the rate of slab bending. The analytical solution suggests an inverse relation of the bending rate with A, which supports our experimental findings. Finally, slab geometries of selected natural subduction zones, derived from high-resolution seismic tomographic images have been compiled to validate the experimental R-B versus A regression. We also discuss the subduction settings in which this regression no longer remains valid.

Automated summary

This study uses scaled laboratory experiments and numerical simulations to quantitatively analyze the bending radius of subducting slabs in the upper mantle, taking into account the effects of slab age. The experiments and simulations show consistent results and propose theoretical models relating slab bending to age, validated by observational data from natural subduction zones.