Models for hydraulic fracturing–induced earthquakes in shales typically ascribe fault activation to elevated pore pressure or increased shear stress; however, these mechanisms are incompatible with experiments and rate-state frictional models, which predict stable sliding (aseismic slip) on faults that penetrate rocks with high clay or total organic carbon. Recent studies further indicate that the earthquakes tend to nucleate over relatively short injection time scales and sufficiently far from the injection zone that triggering by either poroelastic stress changes or pore pressure diffusion is unlikely. Here, we invoke an alternative model based on recent laboratory and in situ experiments, wherein distal, unstable regions of a fault are progressively loaded by aseismic slip on proximal, stable regions stimulated by hydraulic fracturing. This model predicts that dynamic rupture initiates when the creep front impinges on a fault region where rock composition favors dynamic and slip rate weakening behavior.

No abstract

Focal-time analysis is a straightforward data-driven method to obtain robust stratigraphic depth control for microseismicity or induced seismic events. The method eliminates the necessity to build an explicit, calibrated velocity model for hypocenter depth estimation, although it requires multicomponent 3D seismic data that are colocated with surface or near-surface microseismic observations. Event focal depths are initially expressed in terms of zero-offset focal time (two-way P-P reflection time) to facilitate registration and visualization with 3D seismic data. Application of the focal-time method requires (1) high-quality P- and S-wave time picks, which are extrapolated to zero offset and (2) registration of correlative P-P and P-S reflections to provide [Formula: see text] and [Formula: see text] time-depth control. We determine the utility of this method by applying it to a microseismic and induced-seismicity data set recorded with a shallow-borehole monitoring array in Alberta, Canada, combined with high-quality multicomponent surface seismic data. The calculated depth distribution of events is in good agreement with hypocenter locations obtained independently using a nonlinear global-search method. Our results reveal that individual event clusters have distinct depth distributions that can provide important clues about the mechanisms of fault activation.

No abstract

A surface stacking chart (Sheriff and Geldart, 1982; Yilmaz, 1987) is a convenient graphical representation, in source‐receiver space, of the trace geometry for a seismic survey. In this display format, collections of traces corresponding to common‐receiver, common‐source, common‐offset, and common‐midpoint (CMP) gathers are aligned along vertical, horizontal, 45 degree, and 135 degree azimuths (θ), with respect to the receiver axis. Provided that the source spacing (Δs) is an integer multiple of the receiver interval (Δr), traces are exactly aligned in these directions (neglecting skids, detours, etc.). The subsurface multiplicity (fold) along lines of constant CMP is then given by the well known expression.