No abstract

The Pawnee M5.8 earthquake is the largest event in Oklahoma instrument recorded history. It occurred near the edge of active seismic zones, similar to other M5+ earthquakes since 2011. It ruptured a previously unmapped fault and triggered aftershocks along a complex conjugate fault system. With a high-resolution earthquake catalog, we observe propagating foreshocks leading to the mainshock within 0.5 km distance, suggesting existence of precursory aseismic slip. At approximately 100 days before the mainshock, two M ≥ 3.5 earthquakes occurred along a mapped fault that is conjugate to the mainshock fault. At about 40 days before, two earthquakes clusters started, with one M3 earthquake occurred two days before the mainshock. The three M ≥ 3 foreshocks all produced positive Coulomb stress at the mainshock hypocenter. These foreshock activities within the conjugate fault system are near-instantaneously responding to variations in injection rates at 95% confidence. The short time delay between injection and seismicity differs from both the hypothetical expected time scale of diffusion process and the long time delay observed in this region prior to 2016, suggesting a possible role of elastic stress transfer and critical stress state of the fault. Our results suggest that the Pawnee earthquake is a result of interplay among injection, tectonic faults, and foreshocks.

The rapidly increased earthquake rate in the central United States has been linked with wastewater injection. While the overall understanding appears clear at large scales, the interaction between injection and faulting at smaller scales within individual sequences is still not clear. For an earthquake sequence in central Oklahoma, we conduct finer‐scale analysis of the spatiotemporal evolution of seismicity and pore pressure modeling. The pore pressure modeling suggests that nearby wells show much stronger correlation with earthquake sequence evolution. Detailed temporal analysis found correlation between earthquake rate, seismic moment, and injection rates from wells in close proximity. However, the observed maximum magnitude (Mmax) is about 1 order of magnitude smaller than expected based on a theoretical relationship between Mmax and cumulative volume. This discrepancy may point toward additional parameters, such as fault size and stress, which influence Mmax. The lower Mmax is consistent with the truncated Gutenberg‐Richter distribution observed from matched filter detected catalog. Overall, the detailed observations suggest that it is possible to resolve relationships between individual disposal wells and induced earthquake sequences.

Many previous studies have suggested that wastewater disposal is the most probable factor affecting increased seismicity in Oklahoma since 2009. While this relationship is clear at the state scale, a systematic quantitative analysis of the spatiotemporal relationships between injection and seismicity is needed. We first apply multiscale analyses to assess the temporal correlation between injection rate and seismicity rate at a range of different grid sizes, which demonstrate clear temporal correlations within the two main seismic regions at variable time delays. The time delay variability decreases with larger grid sizes, whereby the average time delay ranges from 150 to 220 days. The average time delay at large scales is consistent with inferred large‐scale diffusive migration away from areas of high injection rates with diffusivities of 0.5 to 2.0 m2/s. The inferred large‐scale diffusivities are consistent with an expected range of diffusivity within the Arbuckle Group where wastewater disposals are occurring. However, individual earthquake clusters have diffusivities that are about one to two orders lower than the large‐scale models. We interpret this as a manifestation of a two‐layered diffusion model with high diffusivity within the injection layer above basement, which facilitates stress transfer at a larger spatial footprint, triggering seismic slip at multiple seismogenic faults within the crystalline basement with low diffusivity, similar to fluid‐driven clusters in other tectonic regions.

Fault location and geometry are critical considerations in the reactivation of preexisting faults. Here, we combine relocated earthquake catalogs and focal mechanisms to delineate seismogenic faults in Oklahoma and southern Kansas and analyze their stress state. We first identify and map seismogenic faults based on earthquake clustering. We then obtain an improved stress map using 2,047 high‐quality focal mechanisms. The regional stress map shows a gradual transition from oblique normal faulting in western Oklahoma to strike‐slip faulting in central and eastern Oklahoma. Stress amplitude ratio shows a strong correlation with pore pressure from hydrogeologic models, suggesting that pore pressure exhibits a measurable influence on stress patterns. Finally, we assess fault stress state via 3‐D Mohr circles; a parameter understress is used to quantify the level of fault criticality (with 0 meaning critically stressed faults and 1 meaning faults with no applied shear stress). Our results indicate that most active faults have near vertical planes (planarity >0.8 and dip >70°), and there is a strong correlation between fault length and maximum magnitude on each fault. The fault trends show prominent conjugate sets that strike [55–75°] and [105–125°]. A comparison with mapped sedimentary faults and basement fractures reveals common tectonic control. Based on 3‐D Mohr circles, we find that 78% of the faults are critically stressed (understress ≤0.2), while several seismogenic faults are misoriented with high understress (>0.4). Fault geometry and local stress fields may be used to evaluate potential seismic hazard, as the largest earthquakes tend to occur on long, critically stressed faults.

In development of the Bakken/Three Forks play, it is crucial to obtain a strong understanding of not just the hydraulic fracture geometry, but also what portion of those hydraulic fractures are conductive. If both parameters and their interactions are not fully understood, then development of the play could be severely compromised due to unoptimized well spacing and completion design. This study represents a two-pronged approach to better understand this interaction. The first step was to perform a Sealed Wellbore Pressure Monitoring (SWPM) test to gain an understanding of hydraulic half-length (Haustveit. et al. 2020). Then, a conductive interference test was performed to utilize Chow Pressure Group (CPG) to understand the conductive half-length (Chu et al. 2018). This paper will address the results from these two tests and how they can be coupled together to optimize the unique relationship between well spacing and completion design to maximize the value in development of the Bakken/Three Forks play or any play both new and mature. The SWPM test was successfully completed on a nine well zipper frac operation consisting of two pads (four well pad/five well pad) where four Middle Bakken and five Three Forks wells were stimulated. The SWPM results provided insight into the hydraulic fracture geometry of the stimulation in multiple scenarios of vertical and lateral separation, as well as various amounts of offsetting depletion. The next step in the analysis was performing a CPG interference test on the five well zipper pad. The CPG results provided insight into not just the initial conductive geometry, but a three month follow up test also showed how the conductivity of the fractures rapidly degrade over time. By coupling the SWPM and CPG analysis together, an operator can learn where hydraulic fractures are growing and what portion of those fractures are conductive. This project design of coupled SWPM and CPG provided multiple learnings: Hydraulic fractures for a well in either the Middle Bakken or Three Forks grow through the Lower Bakken Shale and create large geometries in both the landing and staggered zone Hydraulic growth is faster and geometry larger growing towards modern completion parents versus vintage completion parents A relatively small portion of the hydraulic geometry is conductive, and although early time wells communicate through the Lower Bakken Shale, a 3-month interference test shows closure between the Three Forks and Middle Bakken. From these learnings, an optimized development is being developed for the Bakken/Three Forks play and a similar workflow can be applied to any play both new or mature to maximize value and returns for operators.

Until recently, microseismic has been the primary diagnostic for estimating "bulk" or stage-level fracture geometry, including asymmetry due to parent-child interactions, for modern multi-cluster plug-and-perf completions. However, microseismic cannot provide details on individual fractures or cluster-level measurements. With the continued advances in fiber optic technologies, we can now measure cluster level fracture behavior at the wellbore and in the far-field. Characterizing the relationship between wellbore and far-field fracture geometry, referred to as fracture morphology, is important when simultaneously optimizing completion design and well spacing. Microseismic and fiber optics are very robust, but expensive, technologies and this limits the frequency of their application. Recently developed low-cost pressure-based technologies enable high-volume data acquisition but may not provide the same level of detail compared to microseismic and fiber optic measurements. This paper presents a case history that details the application of deployable fiber optics to characterize fracture geometry and morphology using microseismic and strain data. The paper also presents results from Sealed Wellbore Pressure Monitoring (SWPM) (Haustveit et al. 2020), comparing the lower-cost SWPM technology to the higher-cost deployable fiber. Wireline-fiber was deployed in the inner two wells, one Middle Bakken (MB) and one Three Forks (TF), of a four-well pad. Surface pressures were recorded on all wells on the pad and nearby parent wells. The outer two wells, one MB and one TF, were completed first, using zipper operations. Fiber-based microseismic and strain measurements were used to characterize fracture geometry and morphology, and parent-child interactions. Pressure measurements on the two inner wells were used for SWPM, providing estimates of completion effectiveness and fracture geometry using Volume to First Response (VFR) measurements. The microseismic data showed asymmetric growth from the eastern well to the parent well pad, with fractures covering the entire parent well pad. More symmetric fracture growth was measured for the western well, as the parent well pad was farther away. The microseismic data provided fracture geometry measurements consistent with previous measurements in the same area using a geophone array. The SWPM results compared favorably to the fiber measurements using the high confidence data. However, there were data acquisition complexities with both technologies that will be detailed in the paper. Fiber strain measurements provided detailed information on fracture morphology, showing significant decreases in the number of far-field hydraulics as distance increases from the completion well. The advancements in Low Frequency Distributed Acoustic Sensing (Ugueto et al. 2019) provides the ability to monitor hydraulic fractures approaching, passing above/under, and intersecting the monitoring location. Both fiber and SWPM showed much faster fracture growth within the same formation compared to fracture growth between formations. The integration of the fiber optic measurements and SWPM results have provided important insights into fracture geometry and morphology, leading to improved hydraulic fracture models.

No abstract