Epistemic uncertainty in California-wide simulations of synthetic seismicity Fred Pollitz, USGS Menlo Park Acknowledgments: David Schwartz, Steve Ward
Ward (2000)
Rundle et al. (2008)
Capabilities of this simulator
Inherently discrete fault zone Inherently discrete fault zone (Ben-Zion and Rice model) (Ben-Zion and Rice model) Homogeneous, planar faults Homogeneous, planar faults Viscoelasticity of lower crust/upper mantle Viscoelasticity of lower crust/upper mantle --> loading through backslip in the --> loading through backslip in the fully relaxed limit fully relaxed limit
Fault network and slip rates courtesy of Steve Ward (June, 2009) Original set of ~4 x 3 km 2 patches are re-grouped and subdivided into set of smoothly-connected, non-overlapping ~1 x 1 km 2 patches fault patches
Stress history on a single fault patch Static frictional stress Dynamic frictional stress Arrest stress Dynamic overshoot parameter: Stress reduction parameter: = s - a D = s - a )/( s - d )
Simulator parameters Dynamic overshoot parameter: Stress reduction parameter: = s - a D = s - a )/( s - d ) Shape of slip-weakening curve Mantle viscosity m
Twenty consecutive M>6.7 ruptures M6.9
M7.2
M7.0
M7.6
M6.7
M7.5
M7.6
M7.2
M6.7
M7.7
M7.3
M7.5
M7.3
M7.2
M6.7
M7.2
M7.3
M7.5
mean recurrence interval (for given magnitude threshold) magnitude-frequency statistics, e.g. b-value coefficient of variation conditional rupture probability Simulator results
Hanks and Bakun (2007 )
Magnitude-frequency statistics
magnitude log 10 [earthquake production rate (yr -1 )] With slip weakening D=1.25 m = 1.2 x Pa s
With slip weakening D=1.25 m = 1.2 x Pa s
Low Mantle Viscosity
High Mantle Viscosity
magnitude log 10 [earthquake production rate (yr -1 )]
Fault system behavior is sensitive to simple parameters Simulations capture Gutenberg-Richter and characteristic-earthquake behavior Comparison of slip maps for the southern SAF suggests that the ‘slip barrier’ nature of a segment boundary depends substantially on mantle viscosity Characteristic magnitude and mean recurrence interval may exhibit a systematic dependence on mantle viscosity Conclusions