Ground motions for induced earthquakes in Oklahoma
Bulletin of the Seismological Society of America 2017
Abstract: (Click to show / hide)We examine ground motions for a sequence of earthquakes in Oklahoma, where assessment of hazard contributions from induced seismicity is of particular interest. We aim to empirically calibrate a model‐driven equation that was derived for central and eastern North America (CENA), so that it will match the observed ground motions of the 2011 Prague, Oklahoma, sequence. We first show that ground motions in Oklahoma decay at a rate similar to the average attenuation observed in the stable continental region of CENA. We then search for any needed adjustments to the CENA source model to match ground motions from the induced seismicity sequence that occurred near Prague in 2011. An interesting feature noted in the ground‐motion analysis is that the stress parameters (Δσ) for the Prague mainshock events (M≥5) are notably higher than those for aftershocks. Moreover, the stress parameter that characterizes the high‐frequency ground‐motion decays in both time and space relative to the three largest events. The largest events in the Prague sequence have similar source parameters to natural CENA events of the same magnitude and focal depth, but their aftershocks have weaker motions.
Constraints on the near-distance saturation of ground-motion amplitudes for small-to-moderate induced earthquakes
Bulletin of the Seismological Society of America 2016
Abstract: (Click to show / hide)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.
Importance of robust attenuation modelling in magnitude-based induced seismicity traffic light protocols - western Alberta
Abstract: (Click to show / hide)Induced-seismicity traffic light protocols with staged magnitude thresholds are implemented to manage the risks associated with induced earthquakes. Reliable modeling of regional attenuation attributes plays an important role in obtaining accurate magnitude estimates, which is directly related to the effectiveness of traffic light protocols due to high costs associated with red-light operation shutdowns. In this study, we demonstrate the impact of region-specific attenuation modeling on the accuracy of event magnitudes for western Alberta case study. We show that a local magnitude (ML) formula that captures the observed regional attenuation attributes results in unbiased and systematically lower magnitudes, in comparison to the standard ML models.
Microseismic event detection and characterization using sparse surface networks
Abstract: (Click to show / hide)In certain shale reservoirs in western Canada and elsewhere in North America, microseismicity associated with hydraulic fracturing can be detected and characterized by sparse surface networks of high-quality three-component instruments. Here we provide case-study examples of sparse surface networks, initially deployed for induced seismic monitoring, that generate rich microseismic data sets in the Duvernay and Montney shale plays. In particular, we show that in environments with favorable surface noise levels, high in-situ stress regime and low anelastic attenuation, such networks can detect microseismic events down to magnitude -1.0 within the hydrofracture-stimulated volume, along with induced seismicity on proximate faults. We utilize three-component detection of compressional and shear waves along with waveform template matching event detection techniques to lower the magnitude of completeness and maximize the recorded event catalog. By locating events using 3D velocity data, grid-search and relative location techniques, the precision and clustering of solutions is optimized for first-pass evaluation of stimulated volume boundaries and delineation of activated geological structures. Furthermore, use of high-quality instruments, including broadband seismometers, allows for unsaturated estimates of event magnitude across the full range of detections, leading to unbiased analysis of magnitude-frequency distributions (b-values) and cumulative radiated seismic energy. For larger magnitude events with sufficient signal to noise ratio, displacement spectral fitting is used to compute source parameters. Larger magnitude events also allow for individual or composite fault plane solutions to be derived and used to perform principal stress axes inversion as well as determine fracture plane orientations. Although surface noise typically interferes with detection of events below magnitude -1.0, spatiotemporal correlations between events above -1.0 and hydrofracture parameters (stage time/location, treatment pressure, slurry rate, proppant volume, etc) can provide operators with metrics for assessing the effectiveness of each frac stage. Consequently, sparse surface networks of high-quality seismic instruments can generate rich and informative data sets that make them a scalable, practical and cost-effective hydraulic fracture imaging option.
Nanometrics Broadband Seismometer
Abstract: (Click to show / hide)
The Duvernay Subscription Array has you covered
Abstract: (Click to show / hide)
Challenges and strategies for monitoring induced seismic activity
AAPG 2015 2015
Abstract: (Click to show / hide)In the past several years, injection induced seismicity has become an important issue. Indeed, more stringent regulations have already been established in Ohio within the past year. However what form these regulations will take remains unclear. Magnitude will play a central role in these policies, but where, if at all, will properties such as location, depth, b-value, and seismicity rate feature? All of these properties have some measure of subjectivity associated with data quality, processing methodology, and a priori knowledge that will inexorably be passed on to the regulations themselves. We address the challenges associated with implementing and meeting regulations that are effective and fair to both the public and the industry. How can we ensure that a seismic monitoring network meets the criteria set out by the regulations? How do we ensure that our attenuation model is giving an unbiased magnitude estimate? How do we incorporate the effect of complex local geology into event location algorithms? We discuss methods and monitoring strategies for industry to overcome these obstacles and to meet new regulations with minimal cost and effort. Finally, we investigate strategies to reduce the subjectivity of the regulations associated with the inherent uncertainty in earthquake properties.
Development of an empirical local magnitude formula for northern Oklahoma
AGU Poster 2015
Abstract: (Click to show / hide)In this paper, we focus on empirically deriving a local magnitude formula for northern Oklahoma. We empirically constrain the attenuation properties within the region of interest based on the amplitude of observed seismograms. We find that existing California-based models show a bias with respect to hypocentral distance and present two different approaches for local magnitude scale calibration. For both methods, the new formula results in magnitudes systematically lower than previous values computed with Eaton’s (1992) and Hutton & Boore (1987) models.
Discrimination of earthquake and blast seismicity in western Alberta
AGU Poster 2015
Abstract: (Click to show / hide)This preliminary study on earthquake versus blast discrimination in western Alberta uses normalized cumulative Arias intensities recorded at regional distances (>50km). Normalized intensities are evaluated after the expected shear-wave arrival to highlight the differences in ground motions following the P coda. Normalized intensities show maximum discrimination between blasts and earthquakes at near-regional distances (50-100km), for wide ground-motion windows. Using area under curve (AUC) analysis of average intensity profiles, confirmed and suspected blast clusters can be discriminated from confirmed earthquakes.
Abstract: (Click to show / hide)
Brochures & Datasheets