S34A-03
Slow slip events in plate-rate laboratory experiments on samples from shallow regions of subduction megathrusts

Wednesday, 16 December 2015: 16:30
305 (Moscone South)
Matt Ikari1, Yoshihiro Ito2, Kohtaro Ujiie3 and Achim Kopf1, (1)MARUM - University of Bremen, Bremen, Germany, (2)Disaster Prevention Research Institute, Kyoto University, Kyoto, Japan, (3)University of Tsukuba, Tsukuba, Japan
Abstract:
It is now recognized that slow slip events on plate-boundary fault zones occur not only downdip of the seismogenic zone, but also updip within several km from the seafloor. These shallower events are advantageous because the fault material can be sampled by drilling, and in-situ conditions can be replicated in the laboratory. We report here on results of laboratory friction experiments investigating fault zone material sampled during Ocean Drilling Program and International Ocean Drilling Program expeditions to subduction zones in Japan and Costa Rica. We employ laboratory conditions of 7-16 MPa effective stress and ~20 °C simulating ~1-2.5 km burial depth. Additionally, we shear the samples at plate convergence rates of less than 10 cm/yr.

When driven at plate convergence rates, friction perturbations analogous to slow slip events are sometimes observed. In plate-boundary fault zone samples from the Tohoku region at the Japan Trench, these events exhibit stress drops of 3-7% over several hours with a maximum slip rate of 10-25 cm/yr under 7 MPa effective normal stress. Increasing the effective normal stress to 16 kPa results in a corresponding increase in stress drop up to 15% and slip rates up to 57 cm/yr, consistent with critical stiffness theory which predicts that increasing effective normal stress decreases frictional stability. A sample from the plate boundary décollement from the Costa Rica margin exhibited similar slow slip behavior, with stress drops of 3-12% and maximum slip velocities up to 16 cm/yr. This sample is described as a hemipelagic clay, whereas a sample of nannofossil chalk from the same region did not exhibit such behavior. For the Nankai Trough offshore Japan, samples from the megasplay fault zone and décollement did not exhibit the slow slip events observed in the Japan Trench and Costa Rica samples. Analyses of velocity-stepping data from these tests indicate a propensity for velocity-weakening frictional behavior at slow rates. Collectively, our results demonstrate that plate-rate shearing rates in the laboratory can replicate the full range of frictional phenomena on shallow subduction megathrusts. Specifically, slow slip events can be generated and their characteristics depend on sample composition as well as testing conditions.