Lithosphere tearing at STEP faults: response to edges of subduction zones

Abstract

Slab edges are a relatively common feature in plate tectonics. Two prominent examples are the northern end of the Tonga subduction zone and the southern end of the New Hebrides subduction zone. Near such horizontal terminations of subduction trenches, ongoing tearing of oceanic lithosphere is a geometric consequence. We refer to such kinks in the plate boundary as a Subduction-Transform Edge Propagator, or STEP. Other STEPs are the north and south ends of the Lesser Antilles trench, the north end of the South Sandwich trench, the south end of the Vrancea trench, and both ends of the Calabria trench. Volcanism near STEPs is distinct from typical arc volcanism. In some cases, slab edges appear to coincide with mantle plumes. Using 3D mechanical models, we establish that STEP faults are stable plate tectonic features in most circumstances. In the (probably rare) cases that the resistance to fault propagation is high, slab break-off will occur. Relative motion along the transform segment of the plate boundary often is non-uniform, and the STEP is not a transform plate boundary in the (rigid) plate tectonics sense of the phrase. STEP propagation may result in substantial deformation, rotation, topography and sedimentary basins, with a very specific time-space evolution. Surface velocities are substantially affected by nearby STEPs.

Introduction

Near the majority of horizontal terminations of subduction trenches, continual tearing of the lithosphere is a geometric consequence [1]. With time, the intersection between the subduction fault and the transform segment propagates through the lithosphere, thereby effectively increasing the area of the transform-like fault (Fig. 1). We refer to the tearing transform fragment as a Subduction-Transform Edge Propagator (STEP) fault.

Following the identification of current plate boundaries, plate-tearing configurations were hypothesized (“scissors type faulting” [2], “hinge faulting” [3], [4]). Tearing has been occurring in the following regions in the last couple of million years (Fig. 2); the northern termination of the Tonga trench, the southern end of the New Hebrides trench, the northern and southern ends of the Lesser Antilles trench, the northern end of the South Sandwich trench, the southern end of the Hikurangi trench (North Island, New Zealand), the southern end of the Vrancea trench (Carpathians), on both ends of the Calabria trench (Sicily), both ends of the Hellenic trench, and possibly the Gibraltar arc. This list is not complete in that it does not include collision examples.

From their inventory of seismicity in STEP regions, Bilich et al. [5] find a “conspicuous scarcity” of strike-slip earthquakes. This finding is consistent with the nature of STEP faults, as explained next (Fig. 1a–b). A STEP fault allows subduction to continue. The amount of transform motion is not necessarily uniform along the STEP fault. To illustrate this point, consider a tectonic setting like the Central Mediterranean, where relative motion between continental Europe and Africa has nearly come to a halt and subduction of the Ionian slab is accompanied by slab roll-back. In such “land-locked basins” [6], the overriding plate shows back-arc extension in response to the movement of the trench. Relative horizontal motion across the STEP fault only occurs up to the back-arc domain. STEP propagation acts like a wave traveling through the landscape; at any location along the STEP path, strains and rotations build up until it has passed by. Accurate timing of deformation and paleomagnetic rotation is therefore crucial to detect STEP propagation geologically. In the alternative case that rigid motion of the overriding plate is possible (Fig. 1c), strike-slip motion will occur across the STEP fault that in this case behaves more like a classical transform plate boundary. Whether the overriding plate can rigidly trail the trench during rollback, or even actively overrides the subducting plate without deforming internally, depends on boundary conditions (left side of Fig. 1a–c) and internal strength.

Schellart et al. [7] use analogue models to study the geological consequences of trench retreat for the North Fiji back-arc basin. Implicit in their experimental approach is the assumption that the STEP fault at the southern end of the New Hebrides trench propagates easily. Prescribed saloon door kinematics on the edge of the basin are shown to reproduce the observed structural development within the overriding plate. Return flow around slab edges is addressed in some recent model studies. Kincaid and Griffiths [8] show that return flow depends on velocities of subduction, slab-steepening, and rollback. Vice versa, the kinematics of subduction are significantly affected by the presence of a slab edge, resulting in a geometry of edge-bounded subduction zones that differs from the typical concave shape towards the overriding plate [9], [10], [11].

We first review some of the main characteristics of existing STEP regions. The inferred range of plate tectonic settings in which STEPs occur, inspires end-member dynamic models of STEP regions. These models are intentionally simple, and without a focus on any specific region. We will show that STEP faults are stable in that, once a STEP geometry exists, it will grow upon itself except in relatively extreme cases. Another outcome of the models is that very particular patterns of surface deformation result from tear propagation, which may be geodetically and geologically detectable. Finally, we predict some first order geological imprints from the models.