Induced seismicity from anthropogenic sources can be a significant nuisance to a local population and in extreme cases lead to damage to vulnerable structures. One type of induced seismicity of particular recent concern, which, in some cases, can limit development of a potentially important clean energy source, is that associated with geothermal power production. A key requirement for the accurate assessment of seismic hazard (and risk) is a ground-motion prediction equation (GMPE) that predicts the level of earthquake shaking (in terms of, for example, peak ground acceleration) of an earthquake of a certain magnitude at a particular distance. Few such models currently exist in regard to geothermal-related seismicity, and consequently the evaluation of seismic hazard in the vicinity of geothermal power plants is associated with high uncertainty.Various ground-motion datasets of induced and natural seismicity (from Basel, Geysers, Hengill, Roswinkel, Soultz, and Voerendaal) were compiled and processed, and moment magnitudes for all events were recomputed homogeneously. These data are used to show that ground motions from induced and natural earthquakes cannot be statistically distinguished. Empirical GMPEs are derived from these data; and, although they have similar characteristics to recent GMPEs for natural and miningrelated seismicity, the standard deviations are higher. To account for epistemic uncertainties, stochastic models subsequently are developed based on a single corner frequency and with parameters constrained by the available data. Predicted ground motions from these models are fitted with functional forms to obtain easy-to-use GMPEs. These are associated with standard deviations derived from the empirical data to characterize aleatory variability. As an example, we demonstrate the potential use of these models using data from Campi Flegrei.Online Material: Sets of coefficients and standard deviations for various groundmotion models.

No abstract

The growing installation of industrial facilities for subsurface exploration worldwide requires continuous refinements in understanding both the mechanisms by which seismicity is induced by field operations and the related seismic hazard. Particularly in proximity of densely populated areas, induced low-to-moderate magnitude seismicity characterized by high-frequency content can be clearly felt by the surrounding inhabitants and, in some cases, may produce damage. In this respect we propose a technique for time-dependent probabilistic seismic-hazard analysis to be used in geothermal fields as a monitoring tool for the effects of on-going field operations. The technique integrates the observed features of the seismicity induced by fluid injection and extraction with a local ground-motion prediction equation. The result of the analysis is the time-evolving probability of exceedance of peak ground acceleration (PGA), which can be compared with selected critical values to manage field operations. To evaluate the reliability of the proposed technique, we applied it to data collected in The Geysers geothermal field in northern California between 1 September 2007 and 15 November 2010. We show that the period considered the seismic hazard at The Geysers was variable in time and space, which is a consequence of the field operations and the variation of both seismicity rate and b-value. We conclude that, for the exposure period taken into account (i.e., two months), as a conservative limit, PGA values corresponding to the lowest probability of exceedance (e.g., 30%) must not be exceeded to ensure safe field operations. We suggest testing the proposed technique at other geothermal areas or in regions where seismicity is induced, for example, by hydrocarbon exploitation or carbon dioxide storage

No abstract

The accurate modeling of ground motion for induced-seismicity hazard estimation is critically dependent on how amplitudes scale with distance near the hypocenter. A rich database of ground motions from small events recorded at close distances in the Geysers region of California has been used to constrain the near-distance saturation effects that control the maximum observed ground motions and intensities for shallow-induced events. The results of this study support the modeling of these effects using an equivalent point-source concept, in which the effective source depth increases from a value near 1 km at moment magnitude (M) of 2 to a value near 3 km at M 4. This near-distance saturation behavior can be applied to the development of ground-motion models for induced seismicity in any region. BSSA Early Edition / 1

Ground-motion prediction equations (GMPEs) play a crucial role for estimating the seismic hazard in any region using either a deterministic or a probabilistic approach. Indeed, they represent a reliable and fast tool to predict strong ground motion, given source and propagation parameters. In this article, we estimated GMPEs for the South Korea peninsula. GMPEs were computed for peak ground displacement, peak ground velocity, peak ground acceleration, and spectral accelerations (damping at 5%) at 13 different periods from 0.055 to 5 s. We analyzed data from 222 earthquakes recorded at 132 three-component stations of the South Korea Seismic Network, from 2007 to 2012, with local magnitude ranging between 2.0 and 4.9 and epicentral distances varying from 1.4 to ∼600 km. A nonlinear mixed effects technique is used to infer the GMPE coefficients. This technique includes both fixed and random effects and accounts for both inter- and intraevent dependencies in the data.\ud Station-specific corrective coefficients were estimated by a statistical approach and\ud were included in the final ground-motion prediction model. Finally, predictions for\ud peak ground acceleration and spectral acceleration are compared with observations\ud recorded for an ML 5.1 earthquake that occurred in 2014, the data for which were\ud not included in the modeling

Applying the concept of cluster entropy, we demonstrate that the primary P-to-s (Ps) conversions originating from the well-recognized seismological boundaries (Moho, 410 km, and 660 km depth interfaces) in Earth's interior and multiple refl ections (reverberations) between the surface and the Moho-the prominent crust-mantle boundary of Earth-can be unambiguously identifi ed. Their attendant information dimension can act as a discriminant. Documentation of shallow mantle stratifi cation at depths ~150-300 km (L/X layers) using Ps signals is scarce and is sometimes attributed to their weak registration in the seismograms together with interference by reverberations from shallow structure. The slopes of best fi ts to mean cluster information dimension (as function of epicentral distance) of Ps signals from target boundaries such as the Moho, 410 km, and L/X discontinuities are consistently gentle with a slight negative character. In contrast, those related to multiples show a steep positive behavior. This opposite nature can potentially discriminate between reverberations and direct converted waves, even when they tend to arrive in overlapping time windows. Our approach thus enables unequivocal identifi cation of shallow mantle layering from Ps data recorded at diverse tectonic provenances of wide antiquity.

No abstract