Department Logo for Earth, Planetary, and Space Sciences

Blending data and dynamics into equilibrium for the Community Stress Model


Nov. 18, 2015, noon - 12:50 p.m.
Geology 1707

Presented By:
Peter Bird
UCLA

See Event on Google. Subscribe to Calendar

Blending data and dynamics into equilibrium for the Community Stress Model

With 3-D tensor models of stress in the lithosphere in southern California, SCEC could: (a) determine the shear stress on active faults to constrain the physics of slip; (b) predict Coulomb stress changes for operational earthquake forecasting; and (c) test the realism of long-term earthquake sequence simulators. One basis for models is data: stress directions (from focal mechanisms and boreholes), and stress intensity (only from boreholes). But earthquakes rarely occur deeper than 15~20 km, and boreholes rarely deeper than ~6 km. Therefore, data must be supplemented by dynamic models using: laboratory flow laws, a geotherm model, a Moho model, relative plate motions, and locations of active faults. One dynamic model uses code Shells, which solves for 2-D equilibrium of vertically-integrated stresses using 2-D velocity models and 3-D structure. While this model predicts full stress tensors, they are discontinuous and noisy. A newer approach is to model the stress anomaly field as the sum of topographic and tectonic stress anomaly fields. In program FlatMaxwell, the topographic stress is defined as the convolution of topography (and deep density anomalies) with analytic solutions for an elastic half-space. The tectonic stress is modeled by sums of derivatives of a Maxwell vector potential field. The whole stress field is then best-fit (by weighted least squares) to both data and the dynamic model. In practice, FlatMaxwell models are limited in spatial resolution to no more than 6 wavelengths along each side of the model domain. Thus they are quite smooth, and cannot represent stress discontinuities at the Moho predicted by the Shells model. Results to date show a low-amplitude stress anomaly, with peak shear stress of 120 MPa and peak vertically-integrated shear stress of 2.9×1012 N/m. Channeling of deviatoric stress along the strong Peninsular Ranges and Great Valley is seen. In southern California, deviatoric stress and long-term strain-rate are negatively correlated because regions of low heat-flow act as stress guides while deforming very little. In contrast, active faults lie in areas with higher heat-flow, and their low strength keeps deviatoric stresses locally modest. Opportunities for future CSM advances include: [1] collecting more data; [2] tuning the Shells dynamic model; [3] using a different dynamic modeling code; and/or [4] applying a similar Maxwell equilibrium filter to models of the interseismic stress rate.