ALBERT

Description: On 2001 May 7, following unintentional water injection, a moderate size induced earthquake struck the Ekofisk oil field, North Sea. Despite of its relatively moderate magnitude, clear low-frequency waveforms could be recorded up to more than 2000 km epicentral distance, suggesting a slow rupture at very shallow depth and wave propagation through low-velocity shallow structures. The event poses a rare opportunity to constrain rupture velocity, duration and rise time of a superficial M 〉 4 event occurring on a horizontal plane in soft, water-saturated sediments. Two previous studies discussed the earthquake point source finding vertical dip-slip focal mechanisms with opposite senses of P and T axes. A further investigation was thus required to provide a basis for a deeper discussion of the failure dynamics. We significantly improve the used data set, test different earth models and derive a point source as well as a kinematic rupture model. We carefully discuss parameter uncertainties and effects related to shallow sources and wave propagation through different crustal structures to resolve the previous controversy. We additionally provide a kinematic rupture model, based on apparent source times derived from Rayleigh and Love waves. The waveforms resolve a predominant unilateral rupture along a horizontal plane at about 2 km depth. We derive an unusually slow rupture, consequence of a slow rupture velocity of about 500 m s –1 and a long rise time of about 7 s. An independent modelling of GPS- based static displacements allows to confirm the focal mechanism polarity and to locate the centroid at the eastern side of the field, resulting in a much larger seismic moment in comparison with dynamic seismic moment. The rupture directivity is confirmed by the relative location of the centroid with respect to the epicentre, which is set at the site of water injection.

Description: This interactive webpage contains supplementary information for the publication by Kühn et al. 2020: "Probabilistic moment tensor inversion for hydrocarbon-induced seismicity in the Groningen gas field, the Netherlands, part 1: testing". It allows for an easy comparison between the various tests of inversion parameters and velocity models described for the analysis of the 11th of March 2017 Zeerijp ML 2.1 earthquake on the event induced in the Groningen gas field (Netherlands). Inversion runs collected here comprise the parameters employed for inversion (Problem Config), the inversion results and error estimates (Parameter Results) as well as a multitude of figures.

Description: Due to its remoteness, the CO2 Lab close to the town of Longyearbyen on Svalbard presents a unique opportunity to demonstrate the entire CO2 value chain based on its closed energy system. The formation considered as potential CO2 storage unit consists of mixed sandstone and shale beds, presenting itself as a fractured, low-permeability reservoir. A geophone network surrounding the injection well has been installed to locate microseismic events during injection tests and to estimate background seismicity. During the first water injection in 2010, a microseismic event (M ∼ 1) was recorded and located close to the injection well, followed by a series of aftershocks. Later injection tests did not generate any detectable microseismic events; nevertheless, pressure and flow rate showed a pattern characteristic for fracture opening potentially indicating “aseismic” fracture propagation. Records of ambient seismic noise are analysed by a cross-correlation method in order to reconstruct the impulse functions between sensors. The daily cross-correlations are dominated by tube wave signals originating from the bottom of the well showing a sudden increase of activity. We also demonstrate a noise cancellation method exhibiting great potential towards cancellation of electromagnetic and cultural noise. Albeit several difficulties that were approached at the CO2 Lab, new knowledge and guidelines for best practice containment monitoring using seismic methods in the Arctic could be developed.

Description: In this paper, a new probabilistic seismic hazard assessment (PSHA) is presented for Peninsular India. The PSHA has been performed using three different recurrence models: a classical seismic zonation model, a fault model, and a grid model. The development of a grid model based on a non-parameterized recurrence model using an adaptation of the Kernel-based method that has not been applied to this region before. The results obtained from the three models have been combined in a logic tree structure in order to investigate the impact of different weights of the models. Three suitable attenuation relations have been considered in terms of spectral acceleration for the stable continental crust as well as for the active crust within the Gujarat region. While Peninsular India has experienced large earthquakes, e.g., Latur and Jabalpur, it represents in general a stable continental region with little earthquake activity, as also confirmed in our hazard results. On the other hand, our study demonstrates that both the Gujarat and the Koyna regions are exposed to a high seismic hazard. The peak ground acceleration for 10 % exceedance in 50 years observed in Koyna is 0.4 g and in the Kutch region of Gujarat up to 0.3 g. With respect to spectral acceleration at 1 Hz, estimated ground motion amplitudes are higher in Gujarat than in the Koyna region due to the higher frequency of occurrence of larger earthquakes. We discuss the higher PGA levels for Koyna compared Gujarat and do not accept them uncritically.

Description: We present results from microseismic monitoring and geomechanical analysis obtained at the industrial-scale CO2 sequestration site at the In Salah gas development project in Algeria. More than 5000 microseismic events have been detected at a pilot monitoring well using a master event cross-correlation method. The microseismic activity occurs in four distinct clusters and thereof three clearly correlate with injection rates and wellhead pressures. These event clusters are consistent with a location within the reservoir interval. However, due to insufficient network geometry there are large uncertainties on event location. We estimate a fracture pressure of 155 bar (at the wellhead) from the comparison of injection pressure and injection rate and conclude that reservoir fracture pressure of the injection horizon has most likely been exceeded occasionally, accompanied by increased microseismic activity. Our analysis of 3-D ray tracing for direct and converted phases suggests that one of the event clusters is located at a shallower depth than the reservoir injection interval. However, this event cluster is most likely unrelated to changes in the injection activity at a single well, as the event times do not correlate with the wellhead pressures. Furthermore, this event cluster shows b-values close to one, indicating re-activated natural or tectonic seismicity on pre-existing weakness zones rather than injection induced seismicity. Analysis of event azimuths and significant shear wave splitting of up to 5 per cent provide further valuable insight into fluid migration and fracture orientation at the reservoir level. Although only one geophone was available during the critical injection period, the microseismic monitoring of CO2 injection at In Salah is capable of addressing some of the most relevant questions about fluid migration and reservoir integrity. An improved monitoring array with larger aperture and higher sensitivity is highly recommended, as it could greatly enhance the value of this technique. As such, real-time microseismic monitoring can be used to guide the injection pressure below fracture pressure, thus providing a tool to mitigate the risk of inducing felt seismicity and compromising seal integrity.

Description: The determination of source parameters and full moment tensors in mines is a difficult task, especially, when the velocity model of the mining environment is complex and strongly heterogeneous. The heterogeneities in the velocity model are usually caused by the presence of ore bodies, host rocks, tunnel systems and large excavations due to mining activity. The mined-out cavities introduce strong velocity contrasts in the model, cause multiple scattering of waves and result in a complex wave field with long coda waves. We have analysed five blasts and five induced microseismic events recorded at the Pyhäsalmi ore mine, Finland, and suggest a strategy of successfully inverting for the seismic moment tensors. We compute accurate locations using an eikonal solver and perform the time-domain moment tensor inversion from full waveforms using a generalized linear inversion. Green's functions are computed using a 3-D finite difference visco-elastic code capable of reproducing complex interactions of waves and structures. To suppress the sensitivity of the inversion to inaccuracies of locations and the velocity model, we analyse the data in the frequency range from 30 to 80 Hz. The analysis of blasts and microseismic events proves that the moment tensor inversion is successful. The moment tensors of blasts display a high percentage of positive isotropic components. However, the presence of minor shear faulting triggered during blasting cannot be excluded. On the other hand, the moment tensors of microseismic events display significant negative isotropic and compensated linear vector dipole components. This indicates that the predominant mechanism of the events is probably related to the collapse of rock due to mining activity.

Description: We study acoustic emissions (AEs) associated with shear and tensile failures around a horizontal borehole in a sandstone sample subjected to triaxial stress. The aim is to relate the AE event rate to macroscopic observations of sample deformation and the percentage of isotropic and deviatoric components of the seismic moment tensors to the expected failure mechanisms. The horizontal hole interferes with the applied load and forms a strongly spatially dependent anisotropic stress field, focusing the crack initiation into both shear and tensile failures. The recorded AEs follows reasonably well existing damage models, but the elastic solution of hoop stress does not represent the onset of failure around the borehole. The focal mechanisms correlate with the orientation of macroscopic fractures in the sample. Events close to the borehole show a higher fraction of isotropic percentage in moment tensors compared to events occurring in the macroscopic fracture featuring higher double-couple percentages. The inhomogeneous stress field due to the borehole and the stress induced damage is strongly affecting the axial and radial velocities which in turn affect the waveforms of the recorded AEs and the resulting moment tensors. The VP/VS ratio obtained from the ratio of isotropic to compensated linear vector dipole components of the moment tensors is close to that obtained from ultrasonic velocity measurements.

Description: Continuous monitoring of fracturing processes in mine environments and the consequent characterization of the damage induced during mining exploitation is of primary interest both for mining engineering and civil protection. The development of improved monitoring and imaging methods is of great importance for salt mines as potential reservoirs for CO2 sequestration. Imaging tools able to handle continuous data streams and providing fast reliable information about stress perturbations and fracturing state will offer important new information to support local authorities in decision-making processes. The monitoring framework will manage continuous datasets, including acoustic, seismic, deformation and thermal data, give access to different inversion and modelling techniques. Continuous data acquisition and storage and automated routines for data analysis will be implemented to image the time evolution of 3D structures at a very local scale. Acoustic and seismic data, whose routinely use in mining survey is typically limited to estimate location and magnitude, will be more widely analysed thanks to full waveform analysis, learning from seismological applications at larger scales. Automated processing will include triggering, source location, moment tensor and extended source parameters inversion. This knowledge can be subsequently used to derive local stress perturbations. A set of different tools, based on statistical analysis of spatiotemporal crack distribution, will be applied to identify rupture clusters and fracturing processes. Local earthquake tomography, which couples source location and seismic velocity inversions, will be applied and interpreted to image the velocity structure. The inclusion of data from cavity deformation, thermal and chemical monitoring will complement acoustic and seismic information, providing a multidisciplinary dataset. The coupling of different data and joint interpretation/inversion methods within a common surveying framework will finally provide high-resolution 3D and 4D tomographic images of the mining area and a continuous monitoring of fracturing processes.