Stability and dynamics of serpentinite layer in subduction zone

Abstract

The hydrous phyllosilicate serpentines have a strong influence on subduction zone dynamics because of their high water content and low strength at shallow and intermediate depths. In the absence of data, Newtonian rheology of serpentinites has been assumed in numerical models yet experimental data show that serpentine rheology is best described by a power law rheology recently determined in subduction zone conditions [Hilairet, N., et al., 2007. High-pressure creep of serpentine, interseismic deformation, and initiation of subduction. Science, 318(5858): 1910–1913]. Using a simple 1D model of a serpentinized channel and – as opposed to previous models – in this power law rheology, we examine the influence of channel thickness, temperature and subduction angle on serpentine flow driven by density contrast (serpentinization degree) with the surroundings. At temperatures of 200–500 °C relevant to intermediate depths a fully serpentinized channel is unlikely to be thicker than 2–3 km. For channel thicknesses of 2 km upward velocities are comparable to those using a constant viscosity of 10 18 Pa s. The velocity profile using power law rheology shows shear zones at the edges of the channel and a low strain rate region at its centre consistent with the frequent observation of weakly deformed HP-rocks. Upward velocities estimated for channels 1 to 3 km thick are comparable to the serpentinization rates for maximum estimates of fluid velocities within shear zones in the literature. Competition between the upward flow and serpentinization may lead to intermittent behavior with alternating growth periods and thinning by exhumation. At shallower levels the thickness allowed for a channel may be up to ~ 8–10 km if the rheology has a higher dependence on stress. We therefore propose that the exhumation of HP oceanic units in serpentinite channels is organized in two levels, the deepest and fastest motion being driven by density contrast with the surrounding mantle and the shallowest circulation being driven by forced return flow. The thicknesses estimated here for serpentinized layers at intermediated depths are similar to the precision of seismic studies. The deepest serpentinite channel may thus be difficult to detect by seismic methods, but it will have a strong influence on the mechanical coupling between the slab and mantle wedge.