Search

Seismic source locations are fundamental to many fields of Earth and planetary sciences, such as seismology, volcanology and tectonics. However, seismic source detection and location are challenging when events cluster closely in space and time with signals tangling together at observing stations, such as they often do in major aftershock sequences. Though emerging algorithms and artificial intelligence (AI) models have made processing high volumes of seismic data easier, their performance is still limited, especially for complex aftershock sequences. In this study, we propose a novel approach that utilizes three‐dimensional image segmentation—a computer vision technique—to detect and locate seismic sources, and develop this into a complete workflow, Source Untangler Guided by Artificial intelligence image Recognition (SUGAR). In our synthetic and real data tests, SUGAR can handle complex, energetic earthquake sequences in near real time better than skillful analysts and other AI and non‐AI based algorithms. We apply SUGAR to the 2016 Kaikōura, New Zealand sequence and obtain five times more events than the analyst‐based GeoNet catalog. The improved aftershock distribution illuminates a continuous fault system with extensive fracture zones beneath the segmented, discontinuous surface ruptures. Our method has broader applicability to non‐earthquake sources and other time series image data sets. Plain Language Summary: Detecting and locating earthquakes is fundamental to seismology, volcanology, and tectonics. A number of emerging algorithms, including some based upon artificial intelligence (AI), have made processing large volumes of seismic data much easier. However, their performance is still limited, especially in clustered aftershock sequences whose signals overlap at observing seismographs. We propose a new, AI computer vision‐based approach to this problem, and develop it into a complete earthquake detection and location workflow, named SUGAR. Tests on synthetic and real earthquake data sets show that SUGAR characterizes complex earthquake sequences better than other AI and non‐AI algorithms or professional analysts. We apply SUGAR to the complex aftershock sequence of the 2016 Mw 7.8 Kaikōura, New Zealand earthquake, detecting five times more events than the analyst‐based GeoNet catalog. Whereas surface breaks of the Kaikōura earthquake are highly discontinuous, our improved aftershock distribution supports a continuous fault system surrounded by extensive fracture zones at depth. Our method has broader potential for other types of seismic sources and image series. Key Points: We propose a new seismic source detection and location approach based on the source‐scanning algorithm and 3D image segmentationThis approach outperforms human analysts and popular artificial intelligence (AI) and non‐AI based methods in characterizing intense aftershock sequencesThe resulting catalog of the 2016 Kaikōura earthquake sequence suggests a continuous fault system surrounded by extensive fracturing [ABSTRACT FROM AUTHOR]

We investigate the origin of a long‐lived earthquake cluster in the Fars arc of the Zagros Simply Folded Belt that is colocated with the major Shanul natural gas field. The cluster emerged in January 2019 and initially comprised small events of Mn ∼ 3–4. It culminated on 9 June 2020 with a pair of Mw 5.4 and 5.7 earthquakes, which was followed by >100 aftershocks. We assess the spatiotemporal evolution of the earthquake sequence using multiple event hypocenter relocations, waveform inversions, and Sentinel‐1 Interferometric Synthetic Aperture Radar (InSAR) measurements and models. We find that the early part of the sequence is spatially distinct from the 9 June 2020 earthquakes and their aftershocks. Moment tensors, centroid depths, and source parameter uncertainties of 15 of the largest (Mn ≥ 4.0) events show that the sequence is dominated by reverse faulting at shallow depths (mostly ≤4 km) within the sedimentary cover. InSAR modeling shows that the Mw 5.7 mainshock occurred at depths of 2–8 km with a rupture length and maximum slip of ∼20 km and ∼0.5 m, respectively. Our results suggest that the 2019–2020 Khalili earthquake sequence was likely influenced by operation of the Shanul field, though elevated natural seismicity in the Zagros makes the association difficult to prove. Understanding how to distinguish man‐made from natural seismicity is helpful for hazard and risk assessment, notably in the Zagros, which is both seismically active and rich in oil and gas reserves. Plain Language Summary: Earthquakes caused by human activities have been documented in a growing number of regions worldwide, but recognizing these events in areas of naturally elevated seismicity remains challenging. We investigate the origin of a long‐lived earthquake cluster in the Fars arc of the Zagros Simply Folded Belt—one of the world's most seismically active mountain belts—that is colocated with a major natural gas field. The seismicity led to public concern and speculation that nearby natural gas extraction was responsible for the seismicity. We assess the spatiotemporal evolution of the earthquake sequence using satellite geodesy and seismology measurements and models. Our results support these being anthropogenic earthquakes, most probably related to operations in the gas field. We suggest that the exploitation of the reservoirs in Iran should be preceded by risk assessment studies and accompanied by the implementation of dedicated, sophisticated monitoring, which would allow seismicity to be detected early and tracked more closely. Key Points: Discrimination of anthropogenic earthquakes in areas of naturally elevated seismicity is challengingThe 2019–2020 Khalili earthquake sequence might be an example of induced seismicity linked to gas extraction in the ZagrosUnderstanding anthropogenic and natural seismicity is important in Iran, which is both seismically active and rich in hydrocarbon reserves [ABSTRACT FROM AUTHOR]

We use calibrated earthquake relocations to reassess the distribution and kinematics of faulting in the Zagros range, southwestern Iran. This is among the most seismically active fold‐and‐thrust belts globally, but knowledge of its active faulting is hampered by large errors in reported epicenters and controversy over earthquake depths. Mapped coseismic surface faulting is extremely rare, with most seismicity occurring on blind reverse faults buried beneath or within a thick, folded sedimentary cover. Therefore, the distribution of earthquakes provides vital information about the location of active faulting at depth. Using an advanced multievent relocation technique, we relocate ∼2,500 earthquakes across the Zagros mountains spanning the ∼70‐year instrumental record. Relocated events have epicentral uncertainties of 2–5 km; for ∼1,100 of them we also constrain origin time and focal depth, often to better than 5 km. Much of the apparently diffuse catalog seismicity now collapses into discrete trends highlighting major active faults. This reveals several zones of unmapped faulting, including possible conjugate left‐lateral faults in the central Zagros. It also confirms the activity of faults mapped previously on the basis of geomorphology, including oblique (dextral‐normal) faulting in the NW Zagros. We observe a primary difference between the Lurestan arc, where seismicity is focused close to the topographic range front, and the Fars arc, where out‐of‐sequence thrusting is evident over a width of ∼100–200 km. We establish a focal depth range of 4–25 km, confirming earlier suggestions that earthquakes are restricted to the upper crust but nucleate both within and beneath the sedimentary cover. Key Points: We present high precision earthquake locations for ∼70 years of instrumental seismicity in the Zagros fold‐and‐thrust belt of southern IranWe relocate ∼2,500 earthquakes with epicentral uncertainties of <5 km, highlighting the existence of numerous unmapped faultsFocal depth distribution (4–25 km) throughout the Zagros range implies earthquakes nucleate both within and beneath the sedimentary cover [ABSTRACT FROM AUTHOR]

We propose a new method, named Seismicity‐Scanning based on Navigated Automatic Phase‐picking (S‐SNAP), that is capable of delineating complex spatiotemporal distributions of seismicity. This novel algorithm takes a cocktail approach that combines source scanning, kurtosis‐based phasepicking, and the maximum intersection location technique into a single integrated workflow. This method is automated, detecting and locating earthquakes efficiently, comprehensively, and accurately. We apply S‐SNAP to a data set recorded by a dense local seismic array during a hydraulic fracturing operation to test this novel approach and to demonstrate its effectiveness in relation to existing methods. Overall, S‐SNAP found about 3.5 times as many high‐quality events as a template matching‐based catalogue. All events in the previous catalogue are identified with similar epicenters, depths, and magnitudes, while no false detections are found by visual inspection. Key Points: We take a cocktail approach combining merits of three different techniques to delineate complex spatiotemporal distribution of seismicityThis new method, named Seismicity‐Scanning based on Navigated Automatic Phase‐picking (S‐SNAP), is comprehensive, accurate, and automatedIn case study using dense array data, it located about 3 times more events than a template matching method without any false detections [ABSTRACT FROM AUTHOR]

The 12 November 2017 Mw 7.3 Ezgeleh‐Sarpolzahab earthquake is the largest instrumentally recorded earthquake in the Zagros Simply Folded Belt by a factor of ∼10 in seismic moment. Exploiting local, regional, and teleseismic data and synthetic aperture radar interferometry imagery, we characterize the rupture, its aftershock sequence, background seismicity, and regional tectonics. The mainshock ruptured slowly (∼2 km/s), unilaterally southward, for ∼40 km along an oblique (dextral‐thrust) fault that dips ∼14°E beneath the northwestern Lurestan arc. Slip is confined to basement depths of ∼12–18 km, resolvably beneath the sedimentary cover which is ∼8 km thick in this area. The gentle dip angle and basement location allow for a broad slip area, explaining the large magnitude relative to earthquakes in the main Fars arc of the Zagros, where shallower, steeper faults are limited in rupture extent by weak sedimentary layers. Early aftershocks concentrate around the southern and western edges of the mainshock slip area and therefore cluster in the direction of rupture propagation, implying a contribution from dynamic triggering. A cluster of events ∼100 km to the south near Mandali (Iraq) reactivated the ∼50° dipping Zagros Foredeep Fault. The basement fault responsible for the Ezgeleh‐Sarpolzahab earthquake probably accounts for the ∼1 km elevation contrast between the Lurestan arc and the Kirkuk embayment but is distinct from sections of the Mountain Front Fault that define frontal escarpments elsewhere in the Zagros. It may be related to a seismic interface underlying the central and southern Lurestan arc, and a key concern is whether or not the more extensive regional structure is also seismogenic. Key Points: The Mw 7.3 Ezgeleh‐Sarpolzahab earthquake ruptured a gently dipping oblique fault in the Lurestan arc basementThis faulting may be pervasive across the Lurestan arc, though it is unlikely seismogenic everywhereRelocated aftershocks correlate strongly with directivity implying a component of dynamic triggering [ABSTRACT FROM AUTHOR]

We investigate the 4 April 2010 Mw 7.2 El Mayor‐Cucapah (Mexico) earthquake using three‐dimensional surface deformation computed from preevent and postevent airborne lidar topography. By profiling the E‐W, N‐S, and vertical displacement fields at densely sampled (∼300 m) intervals along the multisegment rupture and computing fault offsets in each component, we map the slip vector along strike. Because the computed slip vectors must lie on the plane of the fault, whose local strike is known, we calculate how fault dip changes along the rupture. A principal goal is to resolve the discrepancy between field‐based inferences of widespread low‐angle (<30°) oblique‐normal slip beneath the Sierra Cucapah, and geodetic and/or seismological models which support steeper (50°–75°) faulting in this area. Our results confirm that low‐angle slip occurred along a short (∼2 km) stretch of the Paso Superior fault—where the three‐dimensional rupture trace is also best fit by gently inclined planes—as well as along shorter (∼1 km) section of the Paso Inferior fault. We also characterize an ∼8‐km fault crossing the Puerta accommodation zone as dipping ∼60°NE with slip of ∼2 m. These results indicate that within the northern Sierra Cucapah, deep‐seated rupture of steep faults (resolved by coarse geodetic models) transfers at shallower depths onto low‐angle structures. We also observe a statistically significant positive correlation between fault dip and slip, with slip pronounced along steep sections of fault and inhibited along low‐angle sections. This highlights the important role of local structural fabric in controlling the surface expression of large earthquakes. Key Points: We use differential lidar to map the slip vector distribution and shallow fault geometry for the 2010 El Mayor‐Cucapah earthquake ruptureWe observe localized subsurface low‐angle (<30°) oblique‐normal slipWe observe a statistical correlation between dip and slip magnitude along the low‐angle structures [ABSTRACT FROM AUTHOR]

The 1959 Mw ∼7.2 Hebgen Lake earthquake is among the largest continental normal faulting events recorded, as well as one of the earliest associated with a multifault rupture. Multimeter vertical slip was observed on three main, morphologically distinct strands: the Hebgen fault and southeastern section of the Red Canyon fault, which both follow sharp topographic rangefronts, and the Red Canyon fault Kirkwood Ridge section, which cuts steep topography in the footwall of the Hebgen fault. We augment early field, seismological, and geodetic studies by investigating the modern surface rupture using newly acquired airborne lidar topography. By estimating throw from scarp profiling of the ∼36.5 km primary surface rupture, we show both that peak 1959 slip occurred at a structurally mature part of the fault and that many 1959 slip minima are associated with clear structural complexities. Vertical slip often substantially exceeds throw measured at the fault free face immediately after the earthquake; the scarps do not conclusively express beveled forms characteristic of repeated slip and degredation, yet must in places capture both the 1959 earthquake (for which we estimate an average throw of 2.64 m) and one or two preceding latest Pleistocene–Holocene events known from trenching. This has wider, cautionary implications for interpreting paleo‐earthquake chronologies and deriving magnitudes from morphologically simple scarps. By comparing 1959‐only and multievent vertical displacement populations, and considering preliminary paleoseismic data, we suggest that large surface‐rupturing earthquakes on the Hebgen and Red Canyon faults involve highly variable slip distributions. Key Points: Multistranded 1959 Hebgen Lake earthquake rupture reassessed with newly acquired airborne lidar topography dataMorphologically simple scarp captures 1959 event and, in places, one or more Holocene paleo‐earthquakes known from trenchingThe Hebgen‐Red Canyon fault system ruptures with variable slip‐per‐event at a site [ABSTRACT FROM AUTHOR]

Abstract: Three‐dimensional near‐fault coseismic deformation fields from high‐resolution differential topography provide new information on the behavior of the shallow fault zone in large surface‐rupturing earthquakes. Our work focuses on the 16 April 2016 Mw 7.0 Kumamoto, Japan, earthquake, which ruptured ~40 km of the Futagawa‐Hinagu Fault Zone on Kyushu Island with an oblique strike‐slip mechanism and surface offset exceeding 2 m. Our differential lidar analysis constrains the structural style of strain accommodation along the primary fault trace and the surrounding damage zone. We show that 36 ± 29% and 62 ± 32% of the horizontal and vertical deformation, respectively, was accommodated off the principal fault trace. The horizontal strains of up to 0.03 suggest that the approximate elastic strain limit was exceeded over a ~250 m width in many locations along the rupture. The inelastic deformation of the fault volume produced the observed distributed deformation at the Earth's surface. We demonstrate a novel approach for calculating 3‐D displacement uncertainties, indicating errors of centimeters to a few decimeters for displacements computed over 50 m horizontal windows. Errors correlate with land cover and relief, with flatter agricultural land associated with the highest displacement uncertainty. These advances provide a framework for future analyses of shallow earthquake behavior using differential topography. [ABSTRACT FROM AUTHOR]