Recent observations by GPS have revealed transient but aseismic deformation episodes in shallow subduction zones, thus far in the Pacific Northwest of the US, the Guerrero region along the southwestern coast of Mexico, and the Nankai region in Japan. Those seem to involve episodic slips, accumulating over time windows of a month to a few years, along deeper regions of the subduction interface that would not be expected to nucleate regular earthquakes or to move appreciably during such earthquakes. Those transients pose significant questions as to their origin, and also relative to existing concepts of how locked portions of the fault zone (which will ultimately fail in earthquakes) are loaded. That is, the recent results imply an episodic component of the loading, which has consequences for improved predictability of earthquakes. The research in this project addresses the physical causes of transients. Major questions involve how they start and what controls their migration, in the dip direction and along strike. They occur in regions where there is metamorphic fluid release from the subducted seafloor materials, and thus we focus especially on pore water, at very high pressure, as a possibly essential component of understanding transients. More generally we want to understand how they depend on tectonic parameters characteristic of a particular subduction zone (geometry, convergence rate, thermal structure, pattern of metamorphic reactions). A major component of our work is the development of a 3D numerical model of a subduction fault with temperature-dependent, hence depth-dependent, frictional properties, involving a transition to stable friction downdip from the cooler, locked part of the fault zone. Remarkably, in simulations of long tectonic loading sequences with multiple earthquakes along strike, we found that aseismic transient slip episodes emerged spontaneously, with features like in some of the natural observations. In the planned work, we will incorporate into the modeling a more complete physical description of fluid release, transport and pore pressurization, to develop descriptions of transients in a manner that is also consistent with observations of non-volcanic tremors associated with them in some subduction zones. As a way to check and constrain the assumed rheology of the downdip aseismic fault zone in the model, we plan to investigate the fitting of the model to GPS and InSAR constraints on post-seismic slip from the northern Chile subduction zone.
