Long-period Long-duration Research Articles

Evaluating proxies for the drivers of natural gas productivity using machine-learning models

The extensive development of unconventional reservoirs using horizontal drilling and multistage hydraulic fracturing has generated large volumes of reservoir characterization and production data. The analysis of this abundant data using statistical methods and advanced machine-learning (ML) techniques can provide data-driven insights into well performance. Most predictive modeling studies have focused on the impact that different well completion and stimulation strategies have on well production but have not fully exploited the available in situ rock property data to determine its role in reservoir productivity. We have used machine-learning techniques to rank rock mechanical properties, microseismic attributes, and stimulation parameters in the order of their significance for predicting natural gas production from an unconventional reservoir. The data for this study came from a hydraulically fractured well in the Marcellus Shale in Monongalia County, West Virginia. The data classes included measurements aggregated by well completion stage that included (1) gas production, (2) well-log-derived measurements including bulk density, elastic moduli, shear impedance, compressional impedance, brittleness, and gamma measurements, (3) microseismic attributes, (4) long-period long-duration (LPLD) event counts, (5) fracture counts, and (6) stimulation parameters that included the fluid injection volume and average pumping pressure. To identify observable proxies for the drivers of gas production, we evaluated five commonly used ML approaches including multivariate adaptive regression spline, Gaussian mixture model, random forest, gradient boosting, and neural network. We selected five variables including LPLD event count, seismogenic b-value, hydraulic diffusivity, cumulative moment, and fluid volume as the features most likely to impact gas productivity at the stage level in the study area. The data-driven selection of these parameters for their importance in determining gas production can help reservoir engineers design more effective hydraulic-fracture treatments in the Marcellus Shale and other similar unconventional reservoirs. Plain language summary: We use machine-learning methods and data-driven selection of reservoir parameters to rank and better understand their importance in determining gas production, which can help reservoir engineers design more effective hydraulic-fracture treatments in the Marcellus Shale and other similar unconventional reservoirs.

Integrating distributed acoustic sensing, borehole 3C geophone array, and surface seismic array data to identify long-period long-duration seismic events during stimulation of a Marcellus Shale gas reservoir

Microseismic monitoring by downhole geophones, surface seismic, fiber-optic distributed acoustic sensing (DAS), and distributed temperature sensing (DTS) observations were made during the hydraulic fracture stimulation of the MIP-3H well in the Marcellus Shale in northern West Virginia. DAS and DTS data measure the fiber strain and temperature, respectively, along a fiber-optic cable cemented behind the casing of the well. The presence of long-period long-duration (LPLD) events is evaluated in the borehole geophones, DAS data, and surface seismic data of one of the MIP-3H stimulated stages. LPLD events are generally overlooked during the conventional processing of microseismic data, but they represent significant nonbrittle deformation produced during hydraulic fracture stimulation. In a single stage that was examined, 160 preexisting fractures and two faults of suboptimal orientation are noted in the image logs. We identified two low-frequency ([Formula: see text]) events of large temporal duration (tens of seconds) by comparing the surface seismic data, borehole geophone data, and DAS amplitude spectra of one of the MIP-3H stages. Spectrograms of DAS traces in time and depth reveal that the first low-frequency event might be an injection noise that has footprints on all DAS channels above the stimulated stage. However, the surface seismic array indicates an LPLD event concurrent with the first low-frequency event on DAS. The second LPLD event on DAS data and surface seismic data is related to a local deformation and does not have footprints on all DAS channels. The interpreted events have duration less than 100 s with frequencies concentrated below 10 Hz, and are accompanied by microseismic events.

Long-Period, Long-Duration Seismic Events and Their Probable Role in Reservoir Stimulation and Stage Productivity

SummaryHydraulic fracturing is a well-established technique to extract gas or liquid hydrocarbons from low-permeability formations such as shale and tight gas reservoirs. Diffusion of hydrofracturing fluid outward from the stimulated fractures into the target formation produces slip across pre-existing fractures and other discontinuities in the rock. Microseismic events recorded by downhole seismic-monitoring arrays are a manifestation of associated deformation. Recent investigations suggest that the total cumulative seismic moment of microearthquakes during hydraulic fracturing is only a small portion of the total seismic-energy release expected for the fluid volume injected into the formation. These observations suggest that other sources of energy release (such as inelastic deformation), contemporaneous with microseismicity, should be considered relevant to the hydraulic-fracturing process. Recent observations on long-period, long-duration (LPLD) seismic events suggest that slow slip emission along weaknesses that are misaligned with respect to the present-day stress field is likely an important mechanism of deformation and should be better understood and quantified in reservoir stimulations. In Morgantown, West Virginia, we conducted seismic monitoring of hydraulic-fracturing activity using an array of five broadband, three-component (3C) surface seismometers. Using this network, we identified 89 high-amplitude, impulsive events and 436 LPLD events, with highly emergent waveform characteristics. In these interpreted LPLD events, we observed a significant concentration of energy in the 0.8- to 3-Hz frequency range. During hydraulic fracturing, LPLD events were found to occur most frequently when the pumping pressure and rate were at or near maximum values. Because the main purpose of hydraulic fracturing is to stimulate oil and gas production from the less-permeable reservoir, we compared the relative production contributions/stage to the frequency of the occurrence of suspected LPLD events. We found a positive correlation between the frequency of LPLD events and stage-by-stage gas production, highlighting the potential contribution of slow deformation processes and its effectiveness in the reservoir stimulation.

Long‐Period Long‐Duration Events Detected by the IRIS Community Wavefield Demonstration Experiment in Oklahoma: Tremor or Train Signals?

In this study, we apply a recently developed local similarity method to detect long‐period long‐duration (LPLD) seismic events using 1‐month continuous waveforms recorded by a nodal array from the Incorporated Research Institutions for Seismology (IRIS) community wavefield experiment in Oklahoma. The local similarity method detects seismic events by correlating waveforms at each station with its neighboring stations. Using this method, we visually identify 21 potential tremor‐like LPLD events lasting more than 300 s during the 1‐month period. However, with beamforming analysis, we find that the source locations and waveform characteristics of these events are consistent with train‐generated seismic signals from the nearby Union Pacific railway. Additional evidence includes amplitude decay away from the railway track, and similarities in frequency–time contents with other confirmed train‐generated seismic signals. This case study highlights the need of dense‐array observations to distinguish between natural and anthropogenic LPLD seismic signals.

Toward the Origin of Long‐Period Long‐Duration Seismic Events during Hydraulic Fracturing Treatment: A Case Study in the Shale Play of Sichuan Basin, China

Long‐period long‐duration (LPLD) seismic events have been recently observed over several unconventional reservoirs, and are postulated to play a significant role in accommodating deformation within the stimulation zone. However, the origin of LPLD events is still under debate. In this study, we present observations of tremor‐like events during the multistage hydraulic fracturing treatment of a pilot horizontal well within the Weiyuan Shale play in southwestern Sichuan basin, China. These tremor‐like signals recorded by the deep downhole monitoring array (15 Hz geophones) resemble the previously reported LPLD examples in terms of both duration and frequency content. In this particular case, with the concurrent records from the surface broadband array, we however confirm that these signals are located well outside the treatment area, and most probably originate from regional earthquakes that occurred during the treatment course within 250 km radius of the treatment well, instead of local sources directly related to the hydraulic stimulation. This study implies that the attenuated signals of regional seismicity may constitute a particularly deceptive pitfall for identifying and misinterpreting LPLD events during hydraulic fracturing, especially in seismically active regions. Our results also highlight the importance of surface broadband array in monitoring hydraulic fracturing activities.

Long‐period long‐duration seismic events during hydraulic fracturing: Implications for tensile fracture development

AbstractLong‐period long‐duration (LPLD) seismic events are observed from a microseismic data set acquired by surface receivers in the Eagle Ford Shale. These events are characterized by low frequencies of 10–60 Hz and long durations of 30–60 s. The seismograms are dominated with P waves, and the frequency spectra have peaks at several isolated frequencies, similar to volcanic tremors. The LPLD events are located close to the horizontal hydraulic treatment well and migrate away from the well with time. These observations suggest that the LPLD events are related to hydraulic fracturing and are possibly caused by the jerky opening and resonance of fluid‐filled cracks. Imaging this type of LPLD events can potentially map fluid flow and tensile fracture development during hydraulic fracturing.

Long-period, long-duration seismicity observed during hydraulic fracturing of the Marcellus Shale in Greene County, Pennsylvania

Hydraulic fracturing is a well-completion activity that enables the economic extraction of hydrocarbons from unconventional reservoirs such as shale, which are of a naturally low permeability. During hydraulic fracturing, the movement of water into the reservoir from newly created hydraulic fractures reactivates preexisting faults and fractures, initiating shear failure on a complex network of planar features. Although the energy released during hydraulic fracturing is small, the seismic emissions arising from shear failure can be detected using surface and borehole seismometers, and these emissions are an important diagnostic tool for evaluating the effectiveness of reservoir stimulation. To investigate seismic activity during hydraulic fracturing of six horizontal Marcellus Shale wells in Greene County, Pennsylvania, we deployed a surface seismometer within the footprint of the horizontal wells. We recorded 53 high-amplitude, impulsive events and 144 long-period, long-duration (LPLD) events. LPLD events identified in this study show a low-frequency, low-amplitude precursor followed by a high-frequency, high-amplitude primary S-wave signal and are similar to long-duration events (30–60 s) identified in previous studies. After a thorough investigation, we found no temporal correlation with seismic events reported in regional earthquake catalogs and data from USArray stations, suggesting that these observed LPLDs are not attenuated signals from regional earthquakes. Spectral analysis of LPLD events reveals concentrated energy between 1 and 30 Hz. LPLD events were found to occur most frequently when the pumping pressure and rate were at maximum values. Recent findings suggest that “slow-slip emission” along discontinuities that are unfavorably oriented in the ambient stress field is likely a dominant and vital mechanism of deformation during reservoir stimulation. We compared the radiated microseismic energy plus the theoretical fracture energy to the total hydraulic input energy and found an approximately 75% deficit in the energy budget. We propose that LPLD deformation accounts for some portion of this energy deficit.

On the nature of long-period long-duration seismic events detected during hydraulic fracturing

Long-period long-duration (LPLD) seismic events are low-amplitude tremor-like seismic signals that have been observed in some microseismic monitoring data sets acquired during hydraulic fracturing operations. The LPLD events have been interpreted to be associated with slow slip along preexisting fractures presumed to either have high clay content or be misaligned with respect to the current-day principal stress directions. However, a recent study indicates that regional earthquakes, when recorded on vertical downhole monitoring arrays, have similar signal characteristics to LPLD events and that care must be taken when analyzing and interpreting such signals. Using data from a hydraulic fracturing microseismic data set in which LPLD events have previously been identified and well documented, together with data from the EarthScope Transportable USArray, we have investigated the hypothesis that the documented LPLD events were regional earthquakes. We have determined that the LPLD events corresponded with signals recorded on the USArray at distances of up to 350 km away from the injection well, although they were not listed in any regional earthquake catalog. The spatial coverage of the USArray allows the sources of many of the LPLD events to be relocated outside of the treatment well area and thus suggests that they are regional earthquakes of magnitude smaller than M2.5 rather than locally sourced events related to the hydraulic fracturing stimulation process.

Ambient Seismic Imaging Throughout the Unconventional Field’s Life Cycle

Technology Update Newly developed ambient seismic imaging methods provide valuable information throughout the life cycle of an unconventional field. Methods developed by Global Geophysical Services enable new insights for exploration, development planning, fracture and refracture design, reservoir management, production forecasting, and reserves estimation. Ambient seismic imaging has also been used to monitor conventional reservoirs, waterfloods, CO2 sequestration, mine stability, and water influx into mines. Additional applications include geothermal systems and induced seismicity monitoring. Many assume that the term “ambient seismic” (imaging artificial or natural seismic emissions) is synonymous with “microseismic” (imaging earthquakes with a magnitude less than 0, i.e., microearthquakes or MEQs). However, advances in the understanding of natural and induced seismic emissions and new imaging and analytical methods have differentiated these terms. Microseismic methods typically locate MEQs by use of classical seismological methods: polarization angles and arrival times of P and S waves. Thus, the term “microseismic” implies methods relying only on MEQs, the “dots-in-a-box” long familiar to fracturing engineers and development geologists. In recent years, research has shown that the majority of the seismic energy radiated during a fracture treatment is in seismic phenomena other than MEQs. Other emission sources include longperiod, long-duration (LPLD) activity (Das and Zoback 2013a and 2013b), long-duration signals (LDS) (Sicking et al. 2014), and fluid resonance in the fracture system (Tary and van der Baan 2012). These types of emissions cannot be analyzed with classical methods because they do not have distinct, noticeable first arrivals or recognizable P/S wave separation. Such emissions can persist or fade in and out for minutes to hours, rather than milliseconds, and therefore cannot be classified as distinct events such as MEQs. Furthermore, there may be other seismic phenomena that have yet to be identified and categorized. Seismic emission tomography (SET) is an alternative to locating earthquakes with P and S wave arrivals and polarization angles. It is in wide use for surface monitoring and most recently for downhole monitoring. SET focuses total trace energy by means of ray tracing through the velocity model in short time steps (on the order of 100 ms), thus imaging energy from many types of seismic emissions, including MEQs, LPLD, LDS, fluid resonance, and possibly unknown phenomena. Each computed SET time step produces a voxel-by-voxel depth image of seismic emissions in the same way that a 3D surface reflection seismic survey produces a 3D time image. MEQs located in individual time steps are termed “proxy hypocenters” because they represent brief, high-energy events. However, they are not identified uniquely as MEQs by classical methods.

Detection of the Long Period Long Duration (LPLD) Events in Time-and Frequency-Domain

Long Period Long Duration (LPLD) signals are unusual seismic events that can be observed during hydraulic fracturing. These events are very similar in appearance to tectonic tremors sequences, which were first observed in subduction zones. Their nature is not well known. LPLD might be related to the productivity of the reservoir. Different methods of the LPLD events’ detection recorded during hydraulic fracturing are presented. The author applied two methods for LPLD detection–Butterworth filtering and Continuous Wavelet Transform (CWT). Additionally, a new approach to LPLD events detection–instantaneous seismic attributes–was used, common in a classical seismic interpretation but not in microseismic monitoring.

Regional seismicity: A potential pitfall for identification of long-period long-duration events

Long-period long-duration (LPLD) events are tremorlike signals that have been observed during monitoring of hydraulic-fracture treatment programs. LPLD events have been interpreted to reflect slow deformation processes on fractures or faults that are misoriented for reactivation with respect to the present-day stress field. Regional earthquakes could easily be mistaken for LPLD events because both are characterized by similar frequency content ([Formula: see text]) and duration ([Formula: see text]). Using data from a 10.5-month continuous downhole deployment of a 15-Hz geophone array in a tight-sand gas field in western Canada, we compared recordings of small earthquakes with previously published LPLD events. We determined that regional earthquakes can show similar waveform characteristics to LPLD events, underscoring the importance of distinguishing regional earthquake signals from LPLD events to ensure robust interpretation of reservoir deformation processes.

Long-period long-duration seismic events during hydraulic stimulation of shale and tight-gas reservoirs — Part 2: Location and mechanisms

Long-period long-duration (LPLD) seismic events that have been observed during hydraulic stimulation of shale-gas and tight-gas reservoirs appear to represent slow shear slip on relatively large faults. Within the limitations of the recording geometry, we determine the areas in the reservoirs where the events are located in two case studies in the Barnett shale. In one data set, LPLD events appear to occur in the region where the density of natural fractures as well as the fluid pressure during pumping were highest. In the other data set, the LPLD events are observed to occur between two wells and seem to establish a hydraulic connection between them. In both data sets, the LPLD events occur in areas with very few located microearthquakes. A combination of factors such as high fluid pressure and/or high clay content is potentially responsible for the slowly slipping faults. The LPLD events appear to be occurring only on faults large enough to produce a sequence of slow slip events. We suggest that these slowly slipping faults contribute appreciably to the stimulation of these extremely low-permeability reservoirs and hence mapping the distribution of faults and fractures and areas with rock properties that favor slow, sustained slip, can help in optimizing production.

Long-period, long-duration seismic events during hydraulic stimulation of shale and tight-gas reservoirs — Part 1: Waveform characteristics

Long-period long-duration (LPLD) seismic events are relatively low-amplitude signals that have been observed during hydraulic fracturing in several shale-gas and tight-gas reservoirs. These events are similar in appearance to tectonic tremor sequences observed in subduction zones and transform fault boundaries. LPLD events are predominantly composed of S-waves, but weaker P-waves have also been identified. In some cases, microearthquakes are observed during the events. Based on the similarity with tectonic tremors and our observations of several impulsive S-wave arrivals within the LPLD events, we interpret the LPLD events as resulting from the superposition of slow shear-slip events on relatively large faults. Most large LPLD waveforms appear to start as a relatively slower, low-amplitude precursor, lacking clear impulsive arrivals. We estimate the energy carried by the larger LPLD events to be [Formula: see text] times greater than a [Formula: see text] microseismic event that is typical of the events that occur during hydraulic stimulation. Over the course of the entire stimulation activity of five wells in the Barnett formation (each hydraulically fractured ten times), the LPLD events were found to cumulatively release over an order of magnitude higher energy than microearthquakes. The large size of these LPLD events, compared to microearthquakes, suggests that they represent slip on relatively large faults during stimulation of these extremely low-permeability reservoirs. Moreover, they imply that the accompanying slow slip on faults, probably mostly undetected, is a significant deformation process during multistage hydraulic fracturing.