Search

Subduction interfaces exhibit various slip styles, including slow slip events (SSEs). We use a micromechanics‐based approach to calculate the effective rheology of a shear zone containing ellipsoidal amphibolite clasts deforming by dislocation creep within an interconnected linear‐viscous phyllosilicate‐dominated matrix. Frictional failure occurs if local stress exceeds Mohr‐Coulomb yield strength. At moderate fluid overpressure, mixed‐frictional‐viscous behavior emerges at ∼ ${\sim} $350–560° ${}^{\circ}$C, consistent with a broad zone of mixed fault slip behavior without requiring extreme fluid overpressures. Increasing stress in this transition zone promotes local frictional failure and raises bulk strain rate. If, however, the bulk strain rate increases by more than one order of magnitude, system‐wide frictional sliding becomes preferable. This strain rate increase is insufficient to explain the slip rates observed in geophysically detectable SSEs. Therefore, viscous matrix flow as modeled here cannot explain SSEs without either invoking dynamic weakening within a frictional‐viscous flow or a mechanism switch to dominantly frictional sliding. Plain Language Summary: Subduction plate boundaries are locked near the Earth's surface and will release the stored energy as earthquakes. Subduction zones creep steadily and viscously at deeper depths where temperatures and pressures are high. At the depth of the transition from earthquakes to steady creep, episodic aseismic slip is often observed. Rocks from this region are mixtures of strong, fractured clasts surrounded by a weak matrix. The observations of exhumed rocks suggest that the episodic, aseismic slip may nucleate when local frictional failure occurs in strong clasts, but the surrounding weak matrix stops this failure from generating major earthquakes. However, it is unclear how much the small‐scale rock behavior could be linked to the large‐scale slip. We use a numerical model to simulate the interplay between frictional and viscous creep and calculate the overall behavior of the subduction zone plate boundary. We explore how the slip style changes with depth and determine the transition zone's depth/temperature range. In the transition zone, a small increase in stress or decrease in strength can lead to a change from pure viscous flow to frictional sliding. This study overcomes the scale challenge between the small‐scale features preserved on outcrops and the large‐scale geophysical observations. Key Points: Frictional‐viscous flow in a two‐phase shear zone modeled by a multiscale approach occurs at ∼350–560°C with moderate fluid overpressureStress loading and/or fluid pressure weakening can cause a switch from steady viscous creep to transient frictional slipViscous creep modeled here can accommodate tectonic strain rates but not slow slip events without invoking a switch to frictional sliding [ABSTRACT FROM AUTHOR]

Geophysical observations illustrate that a range of fault slip styles accommodate displacement, ranging from steady creep at plate boundary displacement rates through to episodic earthquake slip at speeds of ~ 1 m/s. Slow slip events (SSEs) represent an intermediate fault slip rate that is faster than plate rate, but too slow to generate observable seismic waves. SSEs are typically observed in or near the transition zone between interseismically locked and fully creeping fault segments. Because SSEs have been observed both up- and downdip of interseismically locked zones, they occur at temperatures ranging from ≤ 100˚C to > 500˚C, and are therefore not controlled by a single, thermally activated mechanism. The physical mechanism behind SSEs, however, remains unclear.Geological observations show that faults can accommodate slip by modes ranging from slip localised on one or more discrete failure planes through to shearing flow distributed within a tabular zone of finite thickness, indicating a large range of possible strain rates in natural faults. Crosscutting relations between brittle and ductile structures demonstrate that strain rate can change both progressively and cyclically within a single fault zone. Cyclic strain rate variations can arise from fluctuations in effective stress, but here we discuss how strain rate transients may also occur in response to internal deformation within a polyphase shear zone.Both geological observations and numerical models have now shown that a range of slip styles may arise from rheological heterogeneity in space. Such heterogeneity is almost ubiquitous in natural faults, where it may be represented by different minerals in a fault gouge, by anastomosing phyllosilicate networks around competent clasts in melanges, and variably metamorphosed and dehydrated rocks within subducting basalts, to name a few examples. In such examples, spatial variations in strain rate are inevitable, and temporal variations may follow. If fault slip style is related to fault zone heterogeneity, then there must be certain conditions that are required to produce SSEs, that can be achieved at a large range of temperatures, but yet are not achieved in fault segments that are fully creeping or locked. To explore these conditions, we use the finite element numerical code 'Underworld' to analyse the behaviour of two-phase shear zones. We suggest there is a minimum viscosity ratio, above which two things happen: 1) stresses are amplified in the stronger component, potentially leading to failure in strong lenses within a flowing, weaker matrix; and 2) strong lenses interact, and episodically form 'force chains' that lead to jamming of the shear zone. The required viscosity contrast (strong/weak) depends on the volume fraction of the weak phase but is typically on the order of 100 – 1000; we suggest that among much evidence for heterogeneity, it is where this level of contrast occurs that slow slip is most likely to arise. [ABSTRACT FROM AUTHOR]