No abstract

Why do stochastic point-source and finite-fault simulation models not agree on the predicted ground motions for moderate earthquakes at large distances? This question was posed by Ken Campbell, who attempted to reproduce the Atkinson and Boore (2006) ground-motion prediction equations for eastern North America using the stochastic point-source program SMSIM (Boore, 2005) in place of the finitesource stochastic program EXSIM (Motazedian and Atkinson, 2005) that was used by Atkinson and Boore (2006) in their model. His comparisons suggested that a higher stress drop is needed in the context of SMSIM to produce an average match, at larger distances, with the model predictions of Atkinson and Boore (2006) based on EXSIM; this is so even for moderate magnitudes, which should be well-represented by a pointsource model. Why? The answer to this question is rooted in significant differences between pointsource and finite-source stochastic simulation methodologies, specifically as implemented in SMSIM (Boore, 2005) and EXSIM (Motazedian and Atkinson, 2005) to date. Point-source and finite-fault methodologies differ in general in several important ways: (1) the geometry of the source; (2) the definition and application of duration; and (3) the normalization of finite-source subsource summations. Furthermore, the specific implementation of the methods may differ in their details. The purpose of this article is to provide a brief overview of these differences, their origins, and implications. This sets the stage for a more detailed companion article, "Comparing Stochastic Point-Source and Finite-Source Ground-Motion Simulations: SMSIM and EXSIM," in which Boore (2009) provides modifications and improvements in the implementations of both programs that narrow the gap and result in closer agreement. These issues are important because both SMSIM and EXSIM have been widely used in the development of ground-motion prediction equations and in modeling the parameters that control observed ground motions.

No abstract

No abstract

A case study of seismicity induced by hydraulic fracturing operations near Fox Creek, Alberta, is used to evaluate the extent to which the potential for induced seismicity at a site alters the pre-existing hazard from natural seismicity. We find that in low-to-moderate seismicity environments, the hazard from an induced-seismicity source, if one is activated in close proximity to a site, can greatly exceed the hazard from natural background seismicity at most probabilities of engineering interest, over a wide frequency range. The most important parameters in determining the induced-seismicity hazard are the activation probability and the b-value of the initiated sequence. Uncertainty in the value of the key input parameters to a hazard analysis implies large uncertainty (more than an order of magnitude) in the likelihood of strong shaking.