Simulations of tremor‐related creep reveal a weak crustal root of the San Andreas Fault

Abstract

Deep aseismic roots of faults play a critical role in transferring tectonic loads to shallower, brittle crustal faults that rupture in large earthquakes. Yet, until the recent discovery of deep tremor and creep, direct inference of the physical properties of lower‐crustal fault roots has remained elusive. Observations of tremor near Parkfield, CA provide the first evidence for present‐day localized slip on the deep extension of the San Andreas Fault and triggered transient creep events. We develop numerical simulations of fault slip to show that the spatiotemporal evolution of triggered tremor near Parkfield is consistent with triggered fault creep governed by laboratory‐derived friction laws between depths of 20–35 km on the fault. Simulated creep and observed tremor northwest of Parkfield nearly ceased for 20–30 days in response to small coseismic stress changes of order 104 Pa from the 2003 M6.5 San Simeon Earthquake. Simulated afterslip and observed tremor following the 2004 M6.0 Parkfield earthquake show a coseismically induced pulse of rapid creep and tremor lasting for 1 day followed by a longer 30 day period of sustained accelerated rates due to propagation of shallow afterslip into the lower crust. These creep responses require very low effective normal stress of ~1 MPa on the deep San Andreas Fault and near‐neutral‐stability frictional properties expected for gabbroic lower‐crustal rock.

A 15 year catalog of more than 1 million low‐frequency earthquakes: Tracking tremor and slip along the deep San Andreas Fault

“…Compared with the 2003 San Simeon earthquake, the 2004 Parkfield earthquake created much larger changes in tremor activity, resulting in elevated LFE rates both NW and SE of Parkfield, with the largest effects generally seen at the shallowest LFE sources closest to the rupture zone. These accelerated rates probably reflect deep postseismic afterslip, which likely extended further into the creeping zone to the NW than beneath the locked zone to the SE [ Shelly and Johnson , ; Johnson et al , ]. Within some LFE families, the rates remain elevated for multiple years after the earthquake, before decaying back near background rates (Figure ).…”

“…This accumulating catalog has already been used in numerous studies. These include characterizations of the effects of the nearby 2003 M 6.5 San Simeon earthquake and the 2004 M 6.0 Parkfield earthquake [ Shelly and Johnson , ; Johnson et al , ], triggering by dynamic stresses of teleseismic and occasionally regional earthquakes [ Peng et al , ; Shelly et al , ; Peng et al , ; Hill et al , ; Peng et al , ], and modulation by tidal forces [ Thomas et al , ; van der Elst et al , ]. Additionally, multiple studies have examined the interactions and observed migration among LFE sources [ Wu et al , , ; Shelly , ; Trugman et al , ].…”

Episodic Tremor and Slip Explained by Fluid‐Enhanced Microfracturing and Sealing

“…In a strike-slip fault context, SSEs remain to be accurately identified (Brenguier et al, 2008;Johnson et al, 2013). However, precise NVT source location on the San Andreas fault system suggests that they occur below the fault plane and extend at least to the base of the crust (Shelly, 2010;Shelly & Hardebeck, 2010).…”

Small interseismic asperities and widespread aseismic creep on the northern Japan subduction interface

“…In both models, slip is computed from boundary element calculations using efficient hierarchical matrix techniques [Bradley, 2014;Johnson et al, 2013]. In both models, slip is computed from boundary element calculations using efficient hierarchical matrix techniques [Bradley, 2014;Johnson et al, 2013].…”