The Spatio-Temporal Evolution of Induced Seismicity Clusters in Oklahoma and Southern Kansas, Earthquake Science Center Seminars, U.S. Geological Survey, Menlo Park, 2017-06-29. [ Video ]
Evolution of Stress and Induced Seismicity at the Enhanced Geothermal System in Soultz-sous-Foręts, France, Earthquake Science Center Seminars, U.S. Geological Survey, Menlo Park, 2014-08-06. [ Video ]
> Selected Publications
In situ continuous monitoring of borehole displacements induced by stimulated hydrofracture growth
We provide direct observations of the three-dimensional displacements of a fluid-driven fracture during water injections in a borehole at 1.5 km depth in the crystalline rock at the Sanford Underground Research Facility (USA). Micro-shearing of the borehole initiates on a foliation plane at 61% of the minimum principal stress sigma3 . As the fluid pressure increases further up to 112% of sigma3, borehole axial and radial displacements increase with injection time highlighting the opening and sliding of a new hydrofracture growing 10 m away from the borehole, in accordance with the ambient normal stress regime and in alignment with the microseismicity. Our study reveals how important it is to consider monitoring zonal deformation in conjunction with pressure and flow to better manage the complex hydromechanical evolution of the growing fracture.
Using a deep neural network and transfer learning to bridge scales for seismic phase picking
The important task of tracking seismic activity requires both sensitive detection and accurate earthquake location. Approximate earthquake locations can be estimated promptly and automatically; however, accurate locations depend on precise seismic phase picking, which is a laborious and time-consuming task. We adapted a deep neural network (DNN) phase picker trained on local seismic data to mesoscale hydraulic fracturing experiments. We designed a novel workflow, transfer learning-aided double-difference tomography, to overcome the 3 orders of magnitude difference in both spatial and temporal scales between our data and data used to train the original DNN. Only 3,500 seismograms (0.45% of the original DNN data) were needed to retrain the original DNN model successfully. The phase picks obtained with transfer-learned model are at least as accurate as the analyst's and lead to improved event locations. Moreover, the effort required for picking once the DNN is trained is a small fraction of the analyst's.
Creation of a mixed-mode fracture network at meso-scale through hydraulic fracturing and shear stimulation
Enhanced Geothermal Systems could provide a substantial contribution to the global energy demand if their implementation could overcome inherent challenges. Examples are insufficient created permeability, early thermal breakthrough, and unacceptable induced seismicity. Here we report on the seismic response of a meso-scale hydraulic fracturing experiment performed at 1.5 km depth at the Sanford Underground Research Facility. We have measured the seismic activity by utilizing a novel 100 kHz, continuous seismic monitoring system deployed in six 60 m-length monitoring boreholes surrounding the experimental domain in 3-D. The achieved location uncertainty was on the order of 1 m, and limited by the signal-to-noise ratio of detected events. These uncertainties were corroborated by detections of fracture intersections at the monitoring boreholes. Three intervals of the dedicated injection borehole were hydraulically stimulated by water injection at pressures up to 33 MPa and flow rates up to 5 L/min. We located 1933 seismic events during several injection periods. The recorded seismicity delineates a complex fracture network comprised of multi-strand hydraulic fractures and shear-reactivated, pre-existing planes of weakness that grew unilaterally from the point of initiation. We find that heterogeneity of stress dictates the outcome of hydraulic stimulations, even when relying on theoretically well-behaved hydraulic fractures. Once hydraulic fractures intersected boreholes, the boreholes acted as a pressure relief and fracture propagation ceased. In order to create an efficient sub-surface heat exchanger, production boreholes should not be drilled before the end of hydraulic stimulations.
A Stochastic Study of Flow Anisotropy and Channelling in Open Rough Fractures
The quantification of fluid flow in rough fractures is of high interest for reservoir engineering, especially for deep geothermal applications. Herein, rough self-affine fractures are stochastically generated with incremental shear displacement and geometrically described by two aperture definitions, the vertical aperture avert and the effective aperture aeff. In order to compare their effect on fracture flow, such as anisotropy and channelling, Local Cubic Law (LCL) model-based 2D fluid flow is simulated. The particularity of this approach is the combination of a stochastic generation of self-affine fractures with a statistical analysis (560 individual realizations) of the impact of the LCL's aperture constraint on fracture flow. The results show that aperture definition affects the quantitative interpretation of flow anisotropy and channeling as well as the aperture distribution of the fractures with shearing. Higher values of mean aperture for a given fracture are found using avert, whereas the aperture standard deviation is larger with aeff. In addition, flow anisotropy is significantly sensitive to aperture definition for small shear displacements and shows a relative higher dispersion with aeff. Thus, LCL prediction models based on avert are expected to lead to higher dispersion of anisotropy results with a higher uncertainty (factor ~2). Realizations based on avert lead to an enhanced clustering of high flow rates for higher shearing displacements. This channeling development results in higher total flow rates for these simulations. These findings support the direct calibration of pre-existing LCL anisotropy simulations based on avert towards more representative results using aeff.
Fluid-induced seismicity may occur in a variety of industrial uses of the underground and in various geomechanical settings. Here I compare the clustering behavior of induced seismicity observed at various settings and analyze what earthquake clustering tells us about the physical processes at play.
Microseismic monitoring of meso-scale stimulations for the DOE EGS Collab project at the Sanford Underground Research Facility
The U.S. Department of Energy's Enhanced Geothermal System (EGS) Collab project aims to improve our understanding of hydraulic stimulations in crystalline rock for enhanced geothermal energy production through execution of intensely monitored meso-scale experiments. The first experiment is being performed at the 4850 ft level of the Sanford Underground Research Facility (SURF), approximately 1.5 km below the surface at Lead, South Dakota. Here we report on microseismic monitoring of repeated stimulation experiments and subsequent flow tests between two boreholes in the Poorman Formation. Stimulations were performed at several locations in the designated injection borehole at flow rates from 0.1 to 5 L/min over temporal durations from minutes to hours. Microseismic monitoring was performed using a dense 3D sensor array including two cemented hydrophone strings with 12 sensors at 1.75 m spacing accompanied by 18 3-C accelerometers, deployed in 6 monitoring boreholes, completely surrounding the stimulation region. Continuous records were obtained over a two-month period using a novel dual recording system consisting of a conventional 96 channel exploration seismograph and a high-performance 64 channel digitizer sampling sensors at 4 and 100 kHz respectively. Using a standard STA/LTA triggering algorithm, we detected thousands of microseismic events with recorded energy in a frequency range generally above 3 kHz and up to 40 kHz. The locations of these events are consistent with creation of a hydraulic fracture and additional reactivation of pre-existing structures. Using manual pick refinement and double-difference relocation we are able to track the fracture growth to high precision. We estimate the times and locations of the fracture intersecting a monitoring and the production borehole using microseismic events. They are in excellent agreement with independent measurements using distributed temperature sensing, in-situ strain observations and measurements of conductivity changes.
How faults wake up: the Guthrie-Langston, Oklahoma earthquakes
Large-scale wastewater disposal has led to a fast-paced reawakening of faults in the Oklahoma/Kansas region. High resolution earthquake relocations show that the inventory of ancient basement faults in the study region differs from results of seismic surveys and geologic mapping focused on the sedimentary cover. We analyze the evolution of seismic activity in the Guthrie-Langston sequence in central Oklahoma in greater detail. Here, seismic activity has reactivated a network of at least 12 sub-vertical faults in an area less than 10 km across. Recorded activity began in late 2013 and peaked about 6 months later and includes two M4 earthquakes. These earthquakes characteristically occur at about 4 km depth below the top of the basement and do not reach the sedimentary cover. The sequence shows a radial growth pattern despite being no closer than 10 km to significant wastewater disposal activity. Hydrologic modeling suggests that activity initiated with a time lag of several years relative to early injection activity. Once initiated, earthquake interactions contribute to the propagation of seismicity along the reactivated faults. As a result, the spatio-temporal evolution of the seismicity mimics a diffusive pattern that is typically thought to be associated with injection activity. Analysis of the Fault Slip Potential shows that most faults are critically stressed in the contemporary stress field. Activity on some faults, for which we find low slip probability, suggest a significant contribution of geomechanical heterogeneities to the reawakening of these ancient basement faults.
Irregular Focal Mechanisms Observed at Salton Sea Geothermal Field: Possible Influences of Anthropogenic Stress Perturbation
At the Salton Sea Geothermal Field (SSGF), strain accumulation is released through seismic slip and aseismic deformation. Earthquake activity at the SSGF often occurs in swarm-like clusters, some with clear migration patterns. We have identified an earthquake sequence composed entirely of focal mechanisms representing an ambiguous style of faulting, where strikes are similar but deformation occurs due to steeply-dipping normal faults with varied stress states. In order to more accurately determine the style of faulting for these events, we revisit the original waveforms and refine estimates of P and S wave arrival times and displacement amplitudes. We calculate the acceptable focal plane solutions using P-wave polarities and S/P amplitude ratios, and determine the preferred fault plane. Without constraints on local variations in stress, found by inverting the full earthquake catalog, it is difficult to explain the occurrence of such events using standard fault-mechanics and friction. Comparing these variations with the expected poroelastic effects from local production and injection of geothermal fluids suggests that anthropogenic activity could affect the style of faulting.
A systematic assessment of the spatio-temporal evolution of fault activation through induced seismicity in Oklahoma and southern Kansas
Much of Oklahoma and Southern Kansas has seen widespread seismic activity in the last decade that is attributed to large-scale wastewater disposal into the Arbuckle Group. Using a waveform-relocated earthquake catalog, we perform a systematic study of the activity on several hundreds of identified faults. We use 93 sequences with at least 30 events for a detailed analysis of their spatio-temporal evolution. For most awakened faults, seismicity tends to initiate at shallower depth and migrates deeper along the faults as the sequence proceeds. No major sequence starts with the largest earthquake and many sequences initiate months before they rise to peak activity. We study temporal clustering as a means to quantify earthquake interactions. Some sequences show no temporal clustering similar to Poissonian background seismicity but at much higher rate than the natural background. Other sequences exhibit strong temporal clustering akin to mainshock-aftershock sequences. We conclude that once initiated by anthropogenic forcing, portions of the activated faults in the Oklahoma/Kansas area are close enough to failure to continue to fail through earthquake interaction. We show that the largest earthquakes occur as a result of continued anthropogenic forcing. In many sequences seismicity continues to populate the previously activated region, rather than growing the activated area. Therefore, improved monitoring has the potential to detect minor activity as it initiates failure on specific faults and thus provide warning time to take actions to mitigate the occurrence of potentially damaging earthquakes.
Waveform relocated earthquake catalog for Oklahoma and Southern Kansas illuminates the regional fault network
For much of Oklahoma, augmentation of the seismic network with new public stations in the activated areas has followed rather than preceded the spread of seismicity across the state, and consequently the network geometry is often unfavorable for resolving the underlying fault structures. With this study, we reanalyze the existing earthquake catalog with additional data from two industry operated networks for the period May 2013 to November 2016. These networks include 40 seismic stations and cover seismically active north-central Oklahoma with a station spacing on the order of 25 km. Relative locations obtained from waveform cross-correlation reveal a striking pattern of seismicity illuminating many previously unmapped faults. Absolute depths are usually well constrained to within 1 km. Relative locations provide about one order of magnitude better precision for resolving the structure of seismicity clusters. Relocated epicenters tend to cluster in linear trends of less than 1 km to more than 20 km in length. In areas with stations closer than about 10 km, we can resolve fault planes by strike and dip. These are generally in agreement with surface wave-derived moment tensor solutions.
Quantifying the heterogeneity of the stress field using borehole data
The heterogeneity of the tectonic stress field is a fundamental property which influences earthquake dynamics and subsurface engineering. Self-similar scaling of stress heterogeneities is frequently assumed to explain characteristics of earthquakes such as the magnitude-frequency relation. However, observational evidence for such scaling of the stress field heterogeneity is scarce.
We analyze the local stress orientations using image logs of two closely spaced boreholes in the Coso Geothermal Field with sub-vertical and deviated trajectories, respectively, each spanning about 2 km in depth. Both the mean and the standard deviation of stress orientation indicators (borehole breakouts, drilling-induced fractures and petal-centerline fractures) determined from each borehole agree to the limit of the resolution of our method although measurements at specific depths may not. We find that the standard deviation in these boreholes strongly depends on the interval length analyzed, generally increasing up to a wellbore log length of about 600 m and constant for longer intervals. We find the same behavior in global data from the World Stress Map. This suggests that the standard deviation of stress indicators characterizes the heterogeneity of the tectonic stress field rather than the quality of the stress measurement. A large standard deviation of a stress measurement might be an expression of strong crustal heterogeneity rather than of an unreliable stress determination. Robust characterization of stress heterogeneity requires logs that sample stress indicators along a representative sample volume of at least 1 km.
Analysis of borehole breakout development using continuum damage mechanics
Damage distribution and evolution have a significant effect on borehole stress concentrations. To model the complex fracturing process and inelastic deformation in the development of the borehole breakout, we implement a continuum damage mechanics (CDM) concept that takes tensile and compressive failure mechanisms into account. The proposed approach explicitly models the dissipative behavior of the material due to cracking and its evolution, which leads to an inhomogeneous redistribution of material properties and stresses in the vicinity of the borehole wall. We apply a constitutive plastic model for Berea sandstone and compare our numerical results to laboratory experiments performed on Tablerock sandstone. We are able to reproduce several characteristics of the failure process during the breakout development as observed in experimental tests, e.g. localized crack distribution in the vicinity of the borehole wall, damage evolution, which exhibits a widening process in the beginning followed by subsequent growth in depth, and shear fracturing-dominated breakout growth in sandstone. A comparison of our results with laboratory experiments performed on a range of stress conditions shows a good agreement of the size of borehole breakouts. The importance of the constitutive damage law in defining the failure mechanisms of the damaging processes is discussed. We show that the depth and the width of breakouts are not independent of each other and no single linear relation can be found between the size of breakouts and the magnitude of the applied stress. Consequently, only one far field principal stress component can be estimated from breakout geometry, if the other two principal stresses are known and sufficient data on the plastic parameters are available.
Geodetic slip model of the MW 5.8 3 September, 2016 Pawnee, Oklahoma, earthquake: Evidence for fault zone collapse
The 3 September 2016 Mw 5.8 Pawnee earthquake in northern Oklahoma is the largest earthquake ever recorded in Oklahoma. The coseismic deformation was measured with both Interferometric Synthetic Aperture Radar and Global Positioning System (GPS), with measureable signals of order 1 cm and 1 mm, respectively. We derive a coseismic slip model from Sentinel-1A and Radarsat 2 interferograms and GPS static offsets, dominated by distributed left-lateral strike slip on a primary west-northwest-east-southeast-trending subvertical plane, whereas strike slip is concentrated near the hypocenter (5.6 km depth), with maximum slip of ~1 m located slightly east and down-dip of the hypocenter. Based on systematic misfits of observed interferogram line-of-sight (LoS) displacements, with LoS based on shear-dislocation models, a few decimeters of fault-zone collapse are inferred in the hypocentral region where coseismic slip was the largest. This may represent the postseismic migration of large volumes of fluid away from the high-slip areas, made possible by the creation of a temporary high-permeability damage zone around the fault.
Analysis and interpretation of stress indicators in deviated wells of the Coso Geothermal Field
Characterizing the tectonic stress field is an integral part of the development of hydrothermal systems and especially for enhanced geothermal systems (EGS). With a well characterized stress field the propensity of fault slip on faults with known location and orientation can be identified. Faults that are critically oriented for faulting with respect to the stress field are known to provide natural fluid pathways. A high slip tendency makes a fault a likely candidate for reactivation during the creation of an EGS. Similarly, the stress state provides insight for the potential of larger, damaging earthquakes should extensive portions of well-oriented, larger faults be reactivated.
The analysis of stress indicators such as drilling-induced fractures and borehole breakouts is the main tool to infer information on the stress state of a geothermal reservoir. The standard procedure is applicable to sub-vertical wellbore sections and highly deviated sections have to be discarded. However, in order to save costs and reduce the environmental impact most recent wells are directionally drilled with deviations that require appropriate consideration of the deviated trajectory. Here we present an analysis scheme applicable to arbitrary well trajectories or a combination of wells to infer the stress state. Through the sampling of the stress tensor along several directions additional information on the stress regime and even relative stress magnitudes can be obtained.
We apply this method on image logs from the pair of wells 58-10 and 58A-10 that were drilled from the same well pad. Both wells have image logs of about 2 km of their trajectories that are separated by less than 300 m. For both wells we obtain a mean orientation of SHmax of N23° with large standard deviations of locations of stress indicators of 24° and 26°, respectively. While the local stress direction is highly variable along both wells with dominant wavelengths from around 50 to 500 m, the mean directions are very consistent and also agree with previous stress estimates in the eastern part of the Coso Geothermal Field. In order to obtain a reliable estimation of the stress orientation in this setting, it is necessary to sample the stress field on an interval long to capture several of the dominant wavelengths.
Geologic Setting of the West Flank: A FORGE Site Adjacent to the Coso Geothermal System
The West Flank FORGE (Frontier Observatory for Research in Geothermal Energy) site is located immediately west and outside of the Coso geothermal field, eastern California. Coso is a fluid-dominated, high temperature geothermal system that has been producing power continuously since 1987. The reservoir is composed of highly faulted, fractured and hydrothermally altered Cretaceous and Jurassic plutonic basement rocks. The heat source is a shallow, silicic magma chamber associated with overlying, Quaternary rhyolites and basalts of the Coso volcanic field. Over 30 years of development drilling and associated investigations demonstrates that several well-defined boundaries exist at Coso beyond which there is no commercial, hydrothermal geothermal potential. The West Flank is just west of one of these boundaries and it meets all required FORGE temperature (175-225°C), host rock (crystalline rock), and depth (1.5-4.0 km) criteria.
It has been hypothesized that the Coso volcanic field exists within a right-releasing step-over between two NW-striking, dextral fault systems. Two distinct fault populations yield high permeability drilling targets in the geothermal field. The first population contains WNW-trending with antithetical, NE-trending strike-slip faults and the second includes N- to NNE-trending normal faults. The West Flank appears to be separated from the geothermal field by one such northerly trending fault bounded by a felsic dike swarm. The West Flank's 83-11 well suggests that the West Flank is comprised of weakly altered, leucogranite and diorites typical of the Jurassic to Cretaceous basement. The stress field in the West Flank has been rotated to 81° in contrast to the minimum principal stress orientations within the geothermal field which range from 103° to 108°. Limited drilling, geophysics and other data sets are being assessed and synthesized to develop a working, 3-D conceptual geologic model of the West Flank for FORGE.
Induced seismicity in geothermal reservoirs: A review of forecasting approaches
In order to reach Europe's 2020 and 2050 targets on greenhouse gas emissions, geothermal resources have to contribute substantially to carbon-free energy needs. However, public opinion can be a major show stopper for future large scale application of deep geothermal power plants because induced seismicity is often perceived as an unsolicited and uncontrollable side effect of geothermal development. In the last decade, significant advances have been made in the development of models to forecast induced seismicity, which are either based on catalogues of induced seismicity, on the underlying physical processes, or on a hybrid philosophy. In this paper, we provide a comprehensive overview of the existing approaches applied to geothermal contexts. It serves to understand the merits and pitfalls of the different approaches, to identify the gaps in our understanding, and to describe the needs in geothermal related observations. In a large majority, the forecasting approaches focus on the stimulation phase of enhanced geothermal systems which are the most prone to generate seismic events. Besides statistical models which should better perform for real-time applications during reservoir stimulation, physics-based models have the advantage to consider subsurface characteristics and to estimate the impact of the fluid circulation on the reservoir. Hence, to mitigate induced seismicity during major hydraulic stimulations, the application of hybrid methods in a decision support system seems the best available solution. So far, little attention has been paid to geochemical effects on the failure process and to production periods. The quantitative modelling of induced seismicity is still a challenging and complex task. Suitable resources need to be invested to allow the scientific community continue research and development to successfully forecast induced seismicity in geothermal fields. This is a prerequisite to make this renewable energy resource sustainable and accessible world wide.
Differentiating induced and natural seismicity using space-time-magnitude statistics applied to the Coso Geothermal Field
A remarkable characteristic of earthquakes is their clustering in time and space, displaying their self-similarity. It remains to be tested if natural and induced earthquakes share the same behavior. We study natural and induced earthquakes comparatively in the same tectonic setting at the Coso Geothermal Field. Covering the pre-and co-production periods from 1981 to 2013, we analyze inter-event times, spatial dimension and frequency-size distributions for natural and induced earthquakes. Individually, these distributions are statistically indistinguishable. Determining the distribution of nearest-neighbor distances in a combined space-time-magnitude metric lets us identify clear differences between both kinds of seismicity. Compared to natural earthquakes, induced earthquakes feature a larger population of background seismicity and nearest-neighbors at large magnitude-rescaled times and small magnitude-rescaled distances. Local stress perturbations induced by field operations appear to be strong enough to drive local faults through several seismic cycles and reactivate them after time-periods on the order of a year.
Induced seismicity in geothermal reservoirs: Physical processes and key parameters
In order to reach Europe's 2020 and 2050 targets on greenhouse gas emissions, geothermal resources have to contribute substantially to carbon-free energy needs. Deep geothermal developments, however, are often accompanied by induced seismicity due to stimulation. The induced seismicity can be a threat for the development of future large scale application of deep geothermal power plants. Therefore, understanding the physical processes at the origin of the seismicity induced by forced fluid circulation in geothermal fields is essential, and this paper reviews the current knowledge in connection with field cases.
The driving force of a seismic event is a change of the stress state in the crust. To asses this quantitatively, one needs to know the initial stress state, the spatio-temporal stress changes, the failure criterion, and the rupture dynamics that describes how a seismic event is produced. Several existing geomechanical-numerical models are, in theory, capable of predicting the spatio-temporal changes of the stress state and few of their effects on induced seismicity. They consider coupling between geomechanical, fluid flow, and heat transport processes, with different levels of complexity. The characteristics of the recorded seismicity induced in geothermal fields play a major role to calibrate and to assess the models.
The Soultz-sous-Foręts enhanced geothermal system in France, where thousands of seismic events were induced during stimulations, is an example representative for fields developed in deep crystalline rocks. In such formations, induced seismicity mainly occurs on a network of pre-existing faults and fractures oriented in accordance with the stress field, and shearing on these structures is apparently the dominating failure mechanism. Several physics-based models have been tested on this well-documented field to reproduce the observations. By contrast, the hydraulic fracturing operations carried out in Groß Schönebeck (Germany) geothermal reservoir, which is located in sedimentary formations at a depth similar to Soultz-sous-Foręts, induced very few and weak seismic events. This behavior is consistent with a less seismogenic tensile fracture opening as being the dominant failure mechanism. Interestingly, the few recorded and located seismic events at this site likely occurred on a pre-existing fault.
To characterize a geothermal field with regards to the expected induced seismicity, the following key factors are proposed: natural seismicity at a field scale, stress field, structural fracture/fault characterization, rock type, history of pressure and injection/circulation rates, and past induced seismicity. To mitigate induced seismicity during major hydraulic stimulations, and to prevent large-magnitude event occurrence once injection stopped, early-warning systems and decision support systems are required. To feed these systems, we advocate the application of hybrid methods and the development of fast models. Hybrid methods would combine the best a priori knowledge of the expected behavior of the field underground, inherited from geomechanical-numerical modeling, with a statistical approach based on the real-time observation of the induced seismicity. Fast models would capture the essential physics while minimizing computing time. The quantitative understanding of induced seismicity, however, remains a challenging and complex matter. Only an integration of all current research and development efforts, in the fields of modeling, measuring, monitoring, and matching, will make a chance on success.
Analysis of stress heterogeneities in fractured crystalline reservoirs
The present day state of stress has a key influence on fluid flow through fractured geo-reservoirs. Wellbore breakouts are commonly used as the principal indicator of stress direction. However, variation of breakout orientation with depth, especially in the vicinity of fracture zones, is frequently observed. This study describes a systematic analysis of breakout occurrence, variation of breakout orientation and its modeling. Numerical modeling which takes into account the elastic property changes as a result of fracturing and fracture filling is developed to better quantify the very local scale breakout orientation heterogeneity observed in this study. Two different mechanisms for the breakout rotation are proposed. Anomalies of breakout orientation in the vicinity of fracture zones reflect the large-scale stress heterogeneity which might be caused by the previous slip acting on the fault plane. The local breakout orientation anomalies around minor fractures might be the effect of the material heterogeneities around borehole due the intersection between the borehole with the fracture. The results of this study provide a better understanding of stress-induced borehole elongations in fractured rocks. Borehole breakout heterogeneities do not seem to be related to the principal stress heterogeneity only, but also to the effect of mechanical heterogeneities like weak zones with different elastic moduli, rock strength and fracture patterns. Consequently, care has to be taken when inferring the principal stress orientation from borehole breakout data observed in fractured rock.
Natural or induced: Identifying natural and induced swarms from pre-production and co-production microseismic catalogs at the Coso Geothermal Field
Increased levels of seismicity coinciding with injection of reservoir fluids have prompted interest in methods to distinguish induced from natural seismicity. Discrimination between induced and natural seismicity is especially difficult in areas that have high levels of natural seismicity, such as the geothermal fields at the Salton Sea and Coso, both in California. Both areas show swarm-like sequences that could be related to natural, deep fluid migration as part of the natural hydrothermal system. Therefore, swarms often have spatio-temporal patterns that resemble fluid-induced seismicity, and might possibly share other characteristics.
The Coso Geothermal Field and its surroundings is one of the most seismically active areas in California with a large proportion of its activity occurring as seismic swarms. Here we analyze clustered seismicity in and surrounding the currently produced reservoir comparatively for pre-production and co-production periods. We perform a cluster analysis, based on the inter-event distance in a space-time-energy domain to identify notable earthquake sequences. For each event j, the closest previous event i is identified and their relationship categorized. If this nearest neighbor's distance is below a threshold based on the local minimum of the bimodal distribution of nearest neighbor distances, then the event j is included in the cluster as a child to this parent event i. If it is above the threshold, event j begins a new cluster. This process identifies subsets of events whose nearest neighbor distances and relative timing qualify as a cluster as well as a characterizing the parent-child relationships among events in the cluster.
We apply this method to three different catalogs: (1) a two-year microseismic survey of the Coso geothermal area that was acquired before exploration drilling in the area began; (2) the HYS_catalog_2013 that contains 52,000 double-difference relocated events and covers the years 1981 to 2013; and (3) a catalog of 57,000 events with absolute locations from the local network recorded between 2002 and 2007. Using this method we identify 10 clusters of more than 20 events each in the pre-production survey and more than 200 distinct seismicity clusters that each contain at least 20 and up to more than 1000 earthquakes in the more extensive catalogs.
The cluster identification method used yields a hierarchy of links between multiple generations of parent and offspring events. We analyze different topological parameters of this hierarchy to better characterize and thus differentiate natural swarms from induced clustered seismicity and also to identify aftershock sequences of notable mainshocks. We find that the branching characteristic given by the average number of child events per parent event is significantly different for clusters below than for clusters around the produced field.
Impact of fracture networks on borehole breakout heterogeneities in crystalline rock
Breakouts are commonly used as principal indicator of stress orientation. However, variation of breakout orientation with depth, especially in the vicinity of fracture zones, is frequently observed. This study describes a systematic analysis of breakout occurrence, variation of breakout orientation and fracture characteristics. We infer the impact of fracture networks on the development of breakouts from detailed analysis of 1221 borehole elongation pairs in the vicinity of 1871 natural fractures observed in the crystalline section of the GPK4 well of the Soultz-sous-Foręts geothermal field (France). Breakout orientation anomalies are found to concentrate in the immediate vicinity of fault cores and to decrease with distance to the fault core. Patterns of breakout orientation in the vicinity of natural fractures suggest that the breakout rotation, relative to the mean Shmin direction, is strongly influenced by the fracture orientation. Even a direct relationship between fracture and breakout orientations is found in some depth intervals. In highly fractured zones, with different fracture families present, breakout orientations are especially heterogeneous, resulting from the overlapping effects of the fracture network. Additionally, breakouts are typically found to be asymmetrical in zones with high fracture density. Borehole breakout heterogeneities do not seem to be related to the principal stress heterogeneity only, but also to the effect of mechanical heterogeneities like weak zones with different elastic moduli, rock strength and fracture patterns. Consequently, care has to be taken when inferring the principal stress orientation from borehole breakout data observed in fractured rock.
Large magnitude events during injections in geothermal reservoirs and hydraulic energy: A heuristic approach
The occurrence of induced seismicity during reservoir stimulation requires robust real-time monitoring and forecasting methods for risk mitigation. We propose to derive an estimation of Mmax (here defined as the largest single seismic event occurring during or after reservoir stimulation) using hydraulic energy as a proxy to forecast the total induced seismic moment and to model the transient evolution of the seismic moment distribution (based on the Gutenberg-Richter relation). The study is applied to the vast dataset assembled at the European pilot research project at Soultz-sous-Foręts ( Alsace, France), where four major hydraulic stimulations were conducted at 5 km depth. Although the model could reproduce the transient evolution trend of Mmax for every dataset, detailed results show different agreement with the observations from well to well. This might reveal the importance of mechanical and geological conditions that may show strong local variations in the same EGS.
Forward modelling of seismicity rate changes in georeservoirs with a hybrid geomechanical-statistical prototype model
A key challenge for the development of Enhanced Geothermal Systems (EGS) is to forecast the probability of occurrence of seismic events that have the potential to damage man-made structures. Induced seismicity results from man-made time-dependent stress changes, e.g., due to fluid injection, shut-in and fluid or steam production. To accomplish a classical Probabilistic Seismic Hazard Assessment (PSHA) a catalogue of induced seismicity is required. In addition, PSHA does not return any practical recommendation for how to treat the reservoir geomechanically in order to lower the probability of occurrence of induced seismicity. Thus, we propose to link the simulated stress changes from forward geomechanical numerical reservoir models with the statistical rate-and-state approach ofDieterich (1994). Using this link we translate the modelled time-dependent stress changes into time-dependent changes of seismicity rates. This approach is general and independent of the incorporated geomechanical numerical model used. We exemplify our hybrid model approach using a geomechanical model that describes the stimulation of the well GPK4 at the EGS site in Soultz-sous-Foręts (France) including the shut-in phase. By changing the injection rate in the geomechanical model we generate various synthetic injection scenarios. With these scenarios we can study the effect on the seismicity rate and provide a recommendation for which injection experiment results in the least increase of seismicity rate. The results indicate an explicit coupling between the time-depending stress changes and the induced seismicity rate for each scenario. Even though the hybrid model cannot be used in general to derive absolute values of the rate of induced seismicity a priori (this is only possible if the geomechanical model can be calibrated against observed induced events), it serves as a tool to test the effect of stress changes on the induced seismicity rate. The approach described here is a prototype model illustrating the general workflow. In particular the geomechanical model can be replaced by any other type of reservoir description.
Time-dependent brittle creep as mechanism for time-delayed wellbore failure
By drilling a wellbore cavity, high stresses arise at the wellbore wall, leading to the formation of breakouts which enlarge the hole to an elongated shape oriented along the direction of the minimum principal stress. If formation of breakouts is delayed, rock debris falls on the drill bit which may lead to stuck pipe problems or even abandonment of the drill string. Reasons for such time-delayed failure of the wellbore may be due to chemical fluid-rock interaction, especially in swelling clays. However, such delayed instabilities have also been observed e.g. in gneiss formations at the KTB borehole (Germany) that are not known to exhibit a swelling behavior. We propose to explain observations of delayed wellbore failure by time-dependent brittle creep, which has been observed for many types of rocks. Following this approach, rock fails under loads less than their short-time strength but after a long enough time span. This time is in exponential relation to the load applied to the rock. We implement a model developed for the creation of shear bands on the basis of time-dependent brittle creep by Amitrano and Helmstetter. Here, progressive damage of the formation is captured by a damage parameter D and the time-to-failure TTF. Young's modulus E is decreased by a factor every time TTF is expired, i.e. when failure is reached. Subsequently, stresses are redistributed according to the new distribution of E in the formation. Using this approach, we obtain closure of the well with primary and secondary creep phases. Wellbore breakouts are formed progressively with deepening and widening of the initially damaged zone. After a certain time the formation of breakouts comes to an end with an Omori-like decay of failure approaching a steady-state.
Change of stress regime during geothermal reservoir stimulation
Earthquakes are induced by man-made changes of the stress field by injection or withdrawal of fluids in hydrocarbon production, geothermal exploitation, and wastewater disposal. However, the actual perturbation of the stress field and stress release by injection-induced seismicity remains largely unknown. We provide evidence for currently not understood hydromechanical processes after shut-in of the well. We invert earthquake focal mechanisms from a massive stimulation to invert for stress resolved in time and depth to obtain changes of the stress orientation and magnitude. Prior information about fracture orientations from well logs is taken into account. Comparison with independent stress measures reveals that stresses obtained from inversion of fluid-induced seismicity are highly perturbed and not representative of the initial stress field. The horizontal stresses change by tens of megapascals, turning the stress regime from transitional normal faulting/strike-slip faulting to pure normal faulting. The observed stress changes are attributed to large-scale aseismic deformation.
Economic evaluation of geothermal reservoir performance through modeling the complexity of an operating EGS
The EGS pilot project in Soultz-sous-Foręts, is now operated by an industry consortium, heading for optimal reservoir management. A 3D thermo-hydraulic numerical model, based on a complex geological model of the reservoir is presented with the goal to determine input parameter for an economic analysis, comparing reservoir management based on levelized cost of energy. Over the projected life time of 30 years no major thermal breakthrough is predicted, small temperature decline affects net energy output only negligibly. The results highlight the benefits of multi-well systems, offering a larger heat exchanger surface and higher flexibility for reservoir management.
A better understanding of induced seismicity occurring during the creation of an enhanced geothermal system is vital for future large scale application of geothermal power. Especially the occurrence of large magnitude events (MW>2.0) after shut-in is lacking a comprehensive understanding. We analyze the stimulation of well GPK2 at the Soultz-sous-Foręts (Alsace, France) pilot site with emphasis on the shut-in. We observe a sudden change in spatio-temporal evolution of seismicity starting with shut-in that cannot be explained be currently available approaches to explain the occurrence of post-shut-in seismicity. We relate these observations to structural geological features of the reservoir surrounding well GPK2 such as large faults and the transition between two granite facies.
Flow anisotropy in sheared fractures with self-affine surfaces
The hydraulic behavior of fractures is important for understanding geothermal reservoirs. Thus experimental and numerical analysis of fractures is important. The roughness of the surfaces of natural fractures has an effect on their permeability.
Analytical methods fail to predict this effect in different scales. Here the hydraulic properties of rough fractures are investigated with numerical methods. The fracture model consists of two identical self-affine surfaces, sheared element wise against each other. The heterogeneous aperture is represented by hydraulic conductivity in the numerical model. We use the finite element method in 2D for simulating the fluid flow through the fracture model. Fluid flow is forced with Dirichlet boundary conditions parallel and perpendicular to the shearing direction.
The calculated mean flow through all investigated fractures show an anisotropic behavior, dependent on the orientation of the pressure gradient. The flow parallel (perpendicular) to the shearing direction is increased (enhanced). The effect of channeling perpendicular to the direction is the reason for the anisotropy in fluid flow.
The role of triggering by static stress transfer during geothermal reservoir stimulation
During creation of an Enhanced Geothermal System, massive fluid injections are conducted to induce fracture shear which generates reservoir permeability. In this study we analyze coseismic static stress transfer caused by induced seismic events during a stimulation at the European research project at Soultz-sous-Foręts (Alsace, France). For this purpose we developed an efficient method to calculate coseismic static stress changes from an elliptical slip distribution on a circular fracture using superposition of rectangular sources, which enables us to apply an analytical solution for fast computation. This method is applied on a data set of 715 focal mechanisms derived from seismic recordings of the stimulation of the well GPK2 to calculate temporal evolution of static stress transfer. We find that the resulting structure of coseismic stress changes can be divided into three parts: a quiet zone where no spreading of seismicity occurs, an active zone within the created reservoir with ongoing fracturing and a process zone where the growth of the reservoir occurs. Static stress changes in the active zone are of the order of 1 MPa, both positive and negative, but may exceed this value considerably on a local scale. Analysis of stress changes from a cluster of events that occurred after shut-in lets us conclude that triggering by coseismic static stress changes is possible for some events. Our analysis shows that triggering by static stress transfer plays a minor role for injection induced seismicity in a volumetric reservoir, whereas it can be quite effective for rupture propagation along single large fault zones.
Fluid-induced microseismicity in pre-stressed rock masses
We model microseismicity triggered by fluid injection on the basis of the theory of poroelasticity accounting for the external stress field. Consideration of the fully coupled poroelastic field equations enables us to apply a Coulomb failure criterion using pore fluid pressure and stress tensor as well as the coefficient of friction. The poroelastic fields are calculated with the finite-element method simulating fluid injection with constant injection rate into a 2-D domain. The influence of diffusivity, injection rate and stress field on the occurrence of microseismicity is analysed and compared to simulations based on pore fluid pressure diffusion only. We show that an anisotropic initial stress field causes elongated microseismic clouds. These clouds are indistinguishable from those generated in poroelastic solids under isotropic stress but exhibiting anisotropic hydraulic diffusivity. This similarity shows that microseismicity distributions dependent on both, the hydraulic properties and the coupling of pore fluid pressure to the stress field. In particular, neglecting the influence of the external stress field may lead to overestimation of the anisotropy of diffusivity tensor components. Furthermore, the results of our numerical simulations are strongly sensitive to changes of fluid injection rate.
New galactic open cluster candidates from DSS and 2MASS imagery
An inspection of the DSS and 2MASS images of selected Milky Way regions has led to the discovery of 66 stellar groupings whose morphologies, color-magnitude diagrams, and stellar density distributions suggest that these objects are possible open clusters that do not yet appear to be listed in any catalogue. For 24 of these groupings, which we consider to be the most likely to be candidates, we provide extensive descriptions on the basis of 2MASS photometry and their visual impression on DSS and 2MASS. Of these cluster candidates, 9 have fundamental parameters determined by fitting the color-magnitude diagrams with solar metallicity Padova isochrones. An additional 10 cluster candidates have distance moduli and reddenings derived from K magnitudes and (J-K) color indices of helium-burning red clump stars. As an addendum, we also provide a list of a number of apparently unknown galactic and extragalactic objects that were also discovered during the survey.