GSA Annual Meeting in Phoenix, Arizona, USA - 2019

Paper No. 84-10
Presentation Time: 10:45 AM


FISHER, Donald M., Department of Geosciences, Pennsylvania State University, University Park, PA 16802, SMYE, Andrew J., Department of Geosciences, Pennsylvania State University, 407 Deike, University Park, PA 16802, HOOKER, John N., Department of Geosciences, Pennsylvania State University, 305 Deike Building, University Park, PA 16802 and MARONE, Chris, Department of Geosciences, Penn State, Deike Building, University Park, PA 16802

Fault zone asperities, or discrete patches of increased contact area along faults, must fail and restrengthen repeatedly as part of the seismic cycle. The coincidence between epicenters of earthquakes and coseismically defined asperities (i.e., areas of large moment release) indicates that increases in contact area across the plate interface during the interseismic period control the initiation of great earthquakes in subduction zones. Given observed links between smooth thickly sedimented plate interfaces and large magnitude earthquakes, fault rocks exposed on land from thickly sedimented paleo-subduction plate boundaries provide a record of the deformation processes that likely occur during slip along similar active boundaries. Here, we develop a kinetic model for crack-seal deformation in subduction mélange as a mechanism for increasing contact area across the subduction interface through diffusion of silica from scaly fabric in mudstones into cracks within sandstone blocks. This model, which has a driving force related to strength differences within the mélange, can result in maximum healing times of less than a thousand years but is highly dependent on thermal structure along the plate interface, with interface-controlled kinetics for warm subduction zones based on thermal models or in the case of most P-T records from exposed subduction complexes. This mechanism for asperity development along smooth subduction interfaces (i.e., decrease in crack porosity by mineral precipitation) is thus a plausible explanation for the generation of complex spatial variations in coseismic moment release in the seismogenic zone. A 2-D block-slider model that includes thermally activated stochastic nucleation and growth of healed areas along the interface, coupled with a fluid flow model where permeability decreases during crack healing and increases during coseismic slip, produces observed or postulated characteristics of active subduction zones such as Gutenberg-Richter earthquake size distributions and fault-valve fluid flow.