| Citation: |
Li ZF (2021). Recent advances in earthquake monitoring I: Ongoing revolution of seismic instrumentation. Earthq Sci 34(2): 177–188,. DOI: 10.29382/eqs-2021-0011
|
Seismology is a data-driven science, in which major advancements typically result from improvements in our observation ability (Shearer, 2009). Earthquake monitoring, at its most fundamental form, is the extraction of basic earthquake information (e.g., occurrence time, location, and magnitude) from continuous seismic waveforms. It targets not only large earthquakes but also small shocks with vibration amplitudes beyond human perception limit. Earthquake catalogs, which are major products of global and regional seismic network operations, are essential for understanding earthquake phenomena and probing the Earth’s subsurface structures. In source studies, they are critical for understanding earthquake mechanism and potential seismic hazards, such as background seismicity (Hutton et al., 2010), earthquake sequence evolution (Peng ZG and Zhao P, 2009; Trugman and Ross, 2019), earthquake triggering (Ross et al., 2019), and fault zone mapping (Hauksson, 2005; Ross et al., 2019). Meanwhile, in structural studies, earthquake catalogs are typically used as a basis for seismic tomography. Thus, earthquake monitoring is the basis of observational seismology.
Earthquake monitoring can be generalized as the detection of natural and human-induced seismic events that have trackable seismic footprints. This typically includes, but is not limited to, volcanic unrest (Buurman and West, 2010), landslides (Ekström and Stark, 2013; Lin CH et al., 2010), nuclear tests (Stump, 1991), chemical explosions (Ma X et al., 2020), traffic flow (Wang X et al., 2020), and building collapses (Kim et al., 2001). In these fields, seismological approaches are employed to analyze non-conventional targets, which then develop into new interdisciplinary branches, such as forensic (Meng HR and Ben‐Zion, 2018), environmental (Ekström and Stark, 2013), and urban seismology (Lindsey et al., 2020a; Riahi and Gerstoft, 2015). Owing to the development of local and regional seismic arrays, these new branches have received increasing attention from seismological research communities. Therefore, this review focuses on earthquake seismology and discusses other relevant interests to the broad seismological research community.
Although earthquake monitoring has been prevalent since the birth of modern seismology in the late nineteenth century, the pace of improvement has accelerated significantly in the past decade. Numerous seismological findings, including the discoveries of tectonic tremors, understanding foreshocks and aftershocks, earthquake triggers, and improved fault mapping, have benefited from the improvements in monitoring ability. These improvements primarily exist in two areas: seismic instrumentation (hardware side) and data processing techniques (software side), owing to the rapid development of the manufacturing industry and computer science. This review will focus on the advances in seismic instrumentation, emphasizing on nodes and fiber-optic sensing. The advances in data processing techniques will be reviewed in the second paper.
First, the present status of traditional seismic arrays of broadband and other types of seismometers is briefly introduced, and their strengths and limitations are discussed. Subsequently, a discussion is provided on ultra-dense seismic arrays, also known as large-N arrays, including nodes, micro-electromechanical systems (MEMS), and distributed acoustic sensing (DAS), which typically have hundreds to thousands of sensors with spacings ranging from meters to kilometers. Focus will be given to emerging fiber-optic sensing for submarine monitoring. These novel technologies represent important advances in observational seismology and have the potential to transform our way of conducting seismology research. Notably, there are several review papers discussing the history and current status of specific seismic networks (Peterson and Hutt, 2014; Hutton et al., 2010) and the applications of emerging seismic instruments (e.g. Zhan ZW, 2020). However, unlike previous reviews, this review emphasizes the overall trends in seismic instrumentation and compares their advantages and disadvantages with respect to earthquake monitoring.
Seismic networks have been in existence for approximately a century, and in 1929, seven Wood-Anderson seismometers first formed the Southern California Seismic Network (SCSN). Modern seismic networks typically consist of broadband and strong-motion seismometers. Broadband seismometers have wide recording ranges from hundreds of seconds to hundreds of hertz and are sensitive to ground motions of up to a minimum of 10–10 m/s2 (Clinton and Heaton, 2002). Such broad ranges and high sensitivities render them the optimal option as permanent seismic networks. Thus, they are vital to global seismological studies and are indispensable in cases that require long-period and high-frequency signals (e.g., studies of remote earthquake triggering). Contrastingly, strong-motion seismometers measure ground accelerations up to 2 g, complementing the near-field monitoring ability in large earthquakes, wherein broadband seismometers tend to saturate. A comprehensive discussion regarding the abilities of broadband and strong-motion seismometers can be found in Clinton and Heaton (2002).
The SCSN, as an exemplary seismic network, has developed from seven seismometers in 1929 to more than 600 seismometers in 2021 (Figure 1a; Hauksson et al., 2020). In 1929, each station consisted of a horizontal Wood-Anderson component and a high-gain short-period vertical component (Hutton et al., 2010). Presently, each station is typically equipped with co-located three-component broadband and strong-motion seismometers. These different sensors are employed to cover the entire range of frequency content and dynamic amplitude, thereby providing great expediency for diverse seismological and engineering research.
The SCSN has been producing local earthquake catalogs for a magnitude of completeness of M3.25 from 1932, presently reaching as low as M1.7. The present routine detection method is short-term-average/long-term average (STA/LTA) (Allen, 1982) combined with manual review. STA/LTA only tracks amplitude changes without completely utilizing the complex waveform features, thereby typically failing at low signal-to-noise ratios. Comparatively, using a more sophisticated method, such as template matching (Gibbons and Ringdal, 2006), the magnitude of completeness can reach M0.3 (Ross et al., 2019). The errors in absolute locations in southern California estimated using HypoInverse (Klein, 2002) or SIMULPS (Thurber, 1983) are <0.75 km horizontally and 1.25 km vertically. Moreover, the errors in relative locations using waveform cross-correlation are one order of magnitude smaller (Hauksson et al., 2012). Microearthquakes relocated using double difference methods (Trugman and Shearer, 2017; Waldhauser and Ellsworth, 2000) reveal intricate details about numerous small faults in southern California (Hauksson, 2005; Hauksson et al., 2012; Ross et al., 2019) and enable high-resolution travel-time regional tomography (e.g. Allam and Ben-Zion, 2012).
A similar design has been adopted for seismic networks in Japan. After the 1995 Kobe earthquake, the Japanese government initiated the Headquarters for Earthquake Research Promotion project aiming to construct a national shaking map and promote understanding of the long-term possibility of earthquake occurrences (Okada et al., 2004). This project led to the construction of three nationwide seismic networks, which include high-sensitivity seismographs (Hi-net, Figure 1b), broadband seismographs (F-net), and strong-motion seismographs (K-net). The Hi-net stations involve boreholes with a depth of 100–200 m to reduce the impact of anthropogenic noise. Each station is equipped with short-period velocity seismometers, accelerometers, and tiltmeters. To date, the total number of high-sensitivity seismographs is >1000 with an interstation spacing of 20 km across Japan. The magnitude of completeness is approximately M1.9 (Nanjo et al., 2010), comparable to that in southern California.
Dense networks enable various scientific discoveries and novel seismological techniques. For example, Obara (2002) discovered non-volcanic tremors in the subduction zone of southwestern Japan. Although non-volcanic tremors appear spuriously noisy with sparse network, they emerge remarkably well on a dense network. Using Hi-net data, Obara (2002) located the tremors at the subduction interface and confirmed that they are associated with subduction dynamics. Furthermore, Ishii et al. (2005) developed the back projection method, which utilize the dense nature of the Hi-net to illuminate the rupture process of the 2004 Sumatra-Andaman earthquake. This technique is now standard for imaging the rupture of large earthquakes. There are more examples of seismological advancements alike that benefited from highly dense and high-quality seismic networks.
The major trend in permanent seismic networks is the accelerated increase in the number of stations to improve density and coverage. Another key feature is using multiple types of seismometers such as broadband, short periods, strong-motion, strainmeter, and tiltmeter. Specifically, co-located instruments at the same site can help obtaining a comprehensive measurement of the wavefields. The trend of adding more and diverse sensors is presently one of the driving forces for the accumulation of large seismic data.
Although traditional seismic networks are the backbone of global and regional earthquake monitoring, they exhibit various limitations. First, the networks of sophisticated instruments are expensive, and they are unlikely to cover a wide area with sub-kilometer interstation spacing. Thus, temporary dense arrays are required on several occasions. For example, following a large earthquake, seismologists typically deploy additional sensors surrounding the aftershock zone to obtain high resolution. However, the shipment and deployment of broadband seismometers can take days to weeks, thereby missing critical measurements immediately following the mainshock.
Another limitation is that most seismic networks are located on continents and not seafloors. There are only scattered stations on islands and scattered ocean-bottom seismometers on seafloors globally. This not only leads to unknowns regarding the seismicity and structure of the oceanic lithosphere but also unbalanced imaging of the Earth’s deep interior. Efforts have been made to improve this biased coverage, particularly in risky near-trench regions of subduction zones, using real-time seafloor monitoring networks, such as the Japanese S-net (Uehira et al., 2018), and floating hydrophones, such as the Mermaid array (Simons et al., 2006). However, these networks are expensive in terms of infrastructure construction and routine maintenance. Overall, limited by the current observational techniques, the oceanic area, which accounts for 70.9% of the Earth’s surface, is largely untapped in seismic monitoring.
Finally, traditional seismometers have inherently limited sensing ranges. For example, static offset cannot be measured using either broadband or strong-motion seismometers. In this regard, global navigation satellite systems are indispensable (Bock et al., 2000; Larson et al., 2003). These systems are increasingly being used to study large earthquakes, such as for rupture imaging (Yue H and Lay, 2011), and magnitude estimation in earthquake early warning (Ruhl et al., 2017), which is a rapidly growing field called seismogeodesy. In addition, traditional seismometers measure translational motions and ignore rotational motions. Some studies suggest that the rotational component could play a role in seismic hazards and could be useful to image earthquake processes and the Earth’s subsurface structures (Lin CJ et al., 2011; Reinwald et al., 2016). Although there have been long-term efforts for designing rotational seismometers (Evans et al., 2016; Schreiber et al., 2009), they are far from being common in routine seismological research.
A large gap existed in the resolution limits between earthquake and exploration seismology is partly because of the distinct spatial scales of interest in the two fields. However, some critical zones of interest in earthquake seismology, such as fault zones and induced seismicity regimes, are similar to those in oil and gas sites. In these areas, seismologists require similar high resolutions and search for shortcuts to ultra-dense seismic networks at an affordable cost. Since 2011, there have been rapid developments in this technology, including short-period standalone nodes (e.g. Ben-Zion et al., 2015; Brenguier et al., 2015; Hansen and Schmandt, 2015; Lin FC et al., 2013), MEMS and smartphones (e.g. Kong et al., 2016; Wu YM et al., 2016), and the recently emerging optic fiber sensing (e.g. Lindsey et al., 2017; Wang HF et al., 2018; Yu CQ et al., 2019).
Nodes are seismic systems that integrate a geophone, digitizer, and battery into a standalone unit. They are lightweight and compact, thereby allowing quick deployment. A pioneering example of such arrays in passive source seismology can be traced back to the Long Beach array in 2011 (Lin FC et al., 2013; Schmandt and Clayton, 2013). The Long Beach array contained 5200 nodes with an average spacing of 100 m and was operated for six months (Figure 2a). By applying ambient noise tomography to this array, Lin FC et al. (2013) obtained high-resolution 3D near-surface structures including Rayleigh waves up to 4 Hz. Schmandt and Clayton (2013) analyzed the patterns of teleseismic travel times to estimate the upper mantle depth. Using downward continued array data, Inbal et al. (2015; 2016) detected numerous small earthquakes in the lower crust and upper mantle beneath the Long Beach area. However, their conclusions were challenged by Li ZF et al. (2018) and Yang L et al. (2021), who detected earthquakes without downward continuation and found no evidence of deep events. Despite the discrepancy, these studies validated the strong detection capability of dense arrays.
Inspired by the success of the Long Beach array, the applications of nodal arrays in passive source studies have increased rapidly. Nodal arrays have been used to monitor induced seismicity [e.g., Sweetwater, Texas (Sumy et al., 2015)], volcanoes [Yellowstone (Huang HH et al., 2015), Mount St. Helens (Hansen and Schmandt, 2015), and Piton de la Fournaise volcano in La Reunion (Brenguier et al., 2015)], and fault zone imaging [the San Jacinto Fault in California, (Ben-Zion et al., 2015)]. Nodal sensors have also become a part of rapid response systems in aftershock monitoring to obtain better coverage around rupture zones (Beskardes et al., 2019; Catchings et al., 2020; Pankow et al., 2021). A focus section in Seismological Research Letters (Karplus and Schmandt, 2018) demonstrated a collection of large-N applications.
Numerous sensors provided by the ultra-dense arrays not only allow noise reduction via stacking but also warrant new analysis approaches are inapplicable to sparse arrays. For example, Li ZF et al. (2018) proposed a function called neighboring station similarity for earthquake detection, which utilizes the highly similar earthquake waveforms but different local noise of neighboring stations. This approach exhibited promising detection results for the Long Beach array (Li ZF et al., 2018) and the Incorporated Research Institute for Seismology (IRIS) wavefield demonstration experiment (Li CY et al., 2018). Large-N arrays also provide high resolution in imaging the source properties of earthquakes within or near the array. For example, Fan WY and McGuire (2018) used 350-sensor arrays in Oklahoma to analyze an M2 event. They demonstrated that the dense coverage tightly constrains various properties of such a small event, including the rupture velocity, rupture length and width, and stress drop. These results demonstrate the potential of large-N arrays for monitoring microearthquakes.
In addition to earthquakes, nodal arrays have been used to monitor anthropogenic seismic signals. Specifically, Riahi and Gerstoft (2015) utilized the Long Beach array to analyze metro train activity, arrival and departure of aircrafts, and city highway traffic, thereby demonstrating the applicability of the array in monitoring human activities within a typical city block. Meng HR and Ben‐Zion (2018) identified air traffic events on a 1108-sensor array and managed to invert the flying track. Li ZF et al. (2018) applied the local similarity method to the IRIS wavefield experiment in Oklahoma and detected long-duration signals produced from nearby rail train movement. They used these signals to invert the near-surface attenuation beneath the array. These results not only demonstrate the potential of large-N arrays for smart-city monitoring but also reveal that anthropogenic signals could be a useful source for subsurface imaging, which may contribute to city seismic hazard mapping.
MEMS, which are common in smartphones and laptops, are another type of inexpensive sensor that can be used to construct ultra-dense arrays. The Quake Catcher Network attempts connecting external or internal (smartphones, tablets, and laptops) MEMS accelerometers forming a seismic network mainly for seismological studies (Cochran et al., 2009). Raspberry Shake, another popular low-cost seismic sensor with MEMS, offers both one- and three-component types and can not only be used for individual scientific education purposes (Walter et al., 2019) but also for scientific monitoring in challenging working environments (Winter et al., 2021).
Kong et al. (2016) developed a smartphone application called MyShake, which crowdsources real-time MEMS recordings from the smartphones of the users. They used a machine-learning model to distinguish earthquakes from everyday human motions and demonstrated that the magnitude and location of earthquakes can be reasonably derived from MEMS waveforms. However, practical difficulties to ensure continuous internet connection and power supply remain in the real-time transmission of voluminous data from private users. In comparison, Wu YM et al. (2013) initiated a P-alert project in the Taiwan region to deploy a network of MEMS sensors with plug-in power (Figure 2b), which developed to 757 stations in May 2021. This network exhibited reliable early warning capability for the 2016 M6.4 Meinong earthquake in southwestern Taiwan (Wu YM et al., 2016).
However, smartphone MEMS sensors typically have high self-noise, thereby limiting their applicability in small earthquake detection. Although some high-end MEMS sensors (e.g., HP MEMS) have self-noise levels comparable to those of broad seismometers, they are too expensive for widespread use in seismology (Kong et al., 2016). Inbal et al. (2019) investigated the feasibility of using the MyShake array to detect and locate local small earthquakes. They revealed that the S wave spectra stacked from 100 smartphones exceeded the ambient nighttime noise level for local earthquakes of 1.5 < M < 2 with an epicentral distance <5 km. However, these constraints are not feasible for high-resolution earthquake monitoring, as the primary monitoring targets are M < 2 events in most scenarios. Therefore, MEMS and crowdsourced smartphone data are most suitable for the detection and characterization of moderate to large earthquakes.
Since 2017, DAS has emerged as a novel technology to obtain numerous seismic sensors at a relatively low cost. Essentially, DAS transforms a long optical fiber into thousands of vibration sensors with a spacing as small as 1 m. This is achieved by sending light into the fiber and measuring the backscatter from the inherent impurities in fiberglass [additional technical details can be found in Masoudi and Newson (2016)]. Unlike inertial seismometers that measure particle velocity or acceleration, DAS measures the dynamic strain or strain rate longitudinally along the fiber.
The concept of DAS was proposed in the 1990s followed by applications in various fields. However, its applicability in earthquake seismology has only begun recently. Lindsey et al. (2017) revealed a high correlation between the amplitude and phase of DAS waveforms with a co-located inertial seismometer. Using a fiber in telecommunication conduits, they demonstrated that the detection of teleseismic waves requires little cable-to-soil coupling. In addition, Jousset et al. (2018) showed that DAS can record earthquake signals and microseisms with quality comparable to that of geophones. They also applied an Akaike information criteria phase picker to obtain P and S arrivals and located an earthquake within the array. Wang HF et al. (2018) scrutinized an M4.3 event recorded by co-located DAS and geophone arrays at Brady Hot Springs, Nevada (Figure 2c). They revealed that the differences in the signal-to-noise ratios of the DAS and geophones are generally within a factor of five. They also found good correspondence among the DAS waveforms, finite differences in the geophone waveforms, and the synthetics. Li ZF and Zhan ZW (2018) applied template matching to the Brady array and detected 20 times more earthquakes than those detected from a regional catalog. They demonstrated that, although below the noise level, half of the new events could be verified via thousands of channel-to-channel correlations. Although still in the early stages, these preliminary results demonstrate that DAS can record high-fidelity wavefields of local and regional small earthquakes (Figure 3).
Regarding urban seismology, DAS exhibits more potential applicability than nodal arrays because optical fibers are pervasive in modern cities. Using a DAS array with Pasadena City telecommunication fibers, Wang X et al. (2020) observed the broadband seismic signatures of various celebration activities during the Rose Parade, an annual event in Pasadena that celebrates the New Year. Lindsey et al. (2020a) used a local optical fiber DAS array to monitor traffic changes due to the COVID-19 pandemic. Their results showed that DAS could become an important component of smart-city sensing systems.
Moreover, DAS has a unique ability to operate on seafloor fibers with the measurement unit on land. Lindsey et al. (2019) used a DAS array in Monterey Bay, California, to delineate submarine faults that have not been identified previously. They observed various interesting seismic signatures associated with ocean dynamics, such as the generation of secondary microseisms and sediment transport induced by storms. Similarly, Williams et al. (2019) and Sladen et al. (2019) reported the observations of microseisms and surface gravity waves from submarine DAS arrays in Belgium and France, respectively. These results demonstrate the capability of DAS in monitoring seafloor seismicity and delineating unmapped faults and ocean dynamics. Hence, DAS could possibly become a new paradigm of submarine seismic networks.
Although increasing applications of DAS are expected to fill the monitoring gap in continental shelves and near trench areas, the length of DAS (< 100 km) appears too short to cover wide oceanic areas. Two different optical sensing technologies have recently been proposed to mitigate this problem. Marra et al. (2018) proposed a frequency metrology technique that injects a laser signal to the fiber and measures the output from the other end. The seismic perturbation can then be extracted by comparing the phase difference between the input and output signals. As the measured perturbation is the integration of the entire fiber, identifying the event location requires two or more corroborating arrays. Alternatively, Zhan ZW et al. (2021) analyzed the light polarization changes in daily telecommunication signals and revealed that the polarization changes correspond well to nearby ocean-floor earthquakes. This new method avoids the requirement of injecting an extra laser to the fiber and is highly attractive for earthquake monitoring because the infrastructure and signal sources already exist. They applied this method to a 10,000-km long Curie cable, offshore from southern California to central Chile Figure 3a, and detected multiple moderate to large earthquakes along the route. However, both these optic sensing techniques encounter the same problem of lacking spatial resolution and thus do not provide convenient ways to locate events. Despite this limitation, they can work over a considerably longer distance than DAS and utilize the current submarine fiber networks. Thus, their potential of widespread applications to existing submarine cables worldwide is promising.
The applicability of low-cost ultra-dense arrays has achieved numerous important results and has greatly supplemented the monitoring gap of traditional seismic networks at regional and local scales. Their enhanced spatial sampling and azimuth coverage have increased the resolution of earthquake seismology to approximately that of exploration seismology. However, there are certain limitations. Specifically, the sensitivity and bandwidth of low-cost sensors cannot match with that of sophisticated broadband seismometers. In addition, some fundamental characteristics of the new instruments are not yet clear, thereby hindering the maximum exploitation of the data. Therefore, the strengths and limitations of each type of instrument must be considered while designing earthquake monitoring networks (Table 1).
| Feature | Broadband | Nodes | Micro-electromechanical systems (MEMS) | Distributed acoustic sensing (DAS) | Other optic sensing (M2018, Z2021*) |
| Typical spacing | 1–100 km | 100 m–10 km | 1–10 km | 1–10 m | N/A |
| # of sensors | 10–100 | 100–1000 | 100–1000 | 1000–10000 | N/A |
| Array Aperture | Flexible | Flexible | Flexible | < 100 km | Depend on fiber length |
| Sensitivity | High | Intermediate | Low | Intermediate | Intermediate |
| Power | Cable/solar | Battery | Cable/battery | Cable | Cable |
| Real-time telemetry | Yes | Mostly no | Yes/No | Yes | Yes |
| Submarine | Ocean Bottom Seismometer | No | No | Yes | Yes |
| Note: M2018 and Z2021 represent Marra et al. (2018) and Zhan ZW et al., (2021), respectively. | |||||
DownLoad:
CSV
Nodes produce measurements similar to traditional seismometers that are familiar to seismologists. The array aperture is sufficiently flexible to fit different study scales. However, presently, nodes can operate autonomously for approximately one month, and most lack real-time data telemetry because of limited battery capacity and data storage. Hence, nodes are mostly used as temporary networks. An emerging type of node utilizes a 4G telecommunication network for real-time telemetry, such as UGL-3C from the University of Science and Technology of China, and IMU-3C from SmartSolo (personal communications with Junlun Li, and Hongwei Xu). Although data transmission typically relies on battery performance, this feature makes them particularly useful for rapid responses to large earthquakes to monitor ongoing aftershocks.
In comparison, DAS has several advantages. First, it provides denser spatial sampling than nodes, enabling an unaliased recording of high-frequency wavefields. Second, it can be operated in real time as its power and data storage units are in the interrogator end, which is an indispensable feature for permanent networks. Third, it can avoid the cost of infrastructure construction through utilizing existing telecommunication fiber optics (without interfering with their routine functionality). As optical fibers are widespread globally, it is reasonable to visualize a world wherein every city converts its existing telecommunication fiber optics to DAS seismic networks, which could drastically improve earthquake monitoring and elucidate seismic hazards in the most populous metropolitan areas.
Despite its advantages, it is still challenging to build a purely DAS-based earthquake catalog from scratch. Although many studies have shown that DAS is broadband (Lindsey et al., 2020b; Wang X et al., 2020; Yu CQ et al., 2019), the amount of self-noise, exact amplitude and phase responses, and their dependence on cable coupling, laser stability, and fiber types are largely unknown. Without this knowledge, magnitude calibration and other amplitude-based calculations cannot be performed. Similar problems exist in other optical sensing techniques, including the frequency metrology (Marra et al., 2018) and polarization methods (Zhan ZW et al., 2021). Second, the channel locations of DAS cannot be accurately mapped. Although the channel distance from the laser port can be calculated using the flight time, the cable can have sags or loops. Therefore, tap tests are practically required to locate specific channels. Convenient methods for mapping of all the channels are still lacking. Third, DAS only records the longitudinal direction and lacks three-component information. For a curved fiber, this direction changes with fiber strike, which is different from the consistent east or north components of inertial seismometers. Finally, limited by the fading light power, current DAS arrays can only be <100 km long, which is shorter than the rupture zone of a typical M7 earthquake, unlike the flexible aperture of nodal arrays. In very large earthquakes, multiple DAS arrays are required.
To further leverage ultra-dense instrumentation for improving earthquake monitoring ability, four directions are proposed herein:
1) Understanding data characteristics of new instruments (e.g., self-noise, instrument response, bandwidth, and impact of cable coupling). This will require a series of experiments under well-controlled conditions, as well as a systematic comparison with different types of co-located instruments (e.g., DAS, nodes, and broadband seismometers).
2) Exploring the best practices of instrument combinations for different application scenarios. For example, to record the early aftershocks in aftershock monitoring, DAS could be employed because of its easy shipment and deployment. Afterward, nodes and broadband seismometers can be used to obtain a broader aperture and better azimuthal coverage. In submarine environments, optic sensing can be combined with scattered ocean-bottom seismometers for developing a comprehensive system with detection, location, and magnitude calibration abilities.
3) Developing data storage and transmission, processing, and visualization solutions for large-N datasets. An individual experiment can easily produce terabytes of data because of the numerous sensors. This is a grand challenge in data sharing for typical seismological data centers and for traditional processing and visualization tools (e.g., Seismic Analysis Code (Goldstein et al., 2003)). Thus, next-generation software infrastructures are urgently needed to fulfill the drastic changes in the size of modern seismic data.
4) Developing new earthquake detection methodologies that maximize the use of ultra-dense sampling. Traditional methods, such as STA/LTA (Allen, 1982), were invented during the period of sparse seismic data and did not fully exploit new forms of data. For example, the observations from numerous channels naturally form 2D images, in which earthquake signals could be more efficiently and accurately detected. Future research directions should consider the development of advanced machine learning and image processing techniques for detection and phase picking. The overarching goal is establishing a workflow for the construction of a DAS-based earthquake catalog from scratch.
Seismology has witnessed great improvements in permanent seismic networks over the past decade. The increasingly dense coverage, enhanced data quality, and instrumental diversity have significantly improved our earthquake monitoring abilities. Ultra-dense seismic instruments, including nodes and fiber-optic sensing approaches, provide remarkable dense coverage in regional and local earthquake monitoring. Nodes offer simple installation and flexible apertures and produce data similar to traditional seismometers. However, they are limited in power efficiency and data storage and are thus only suitable for temporary networks. In comparison, DAS can be operated in real time using an in-house power supply. Real-time data telemetry is also possible through internet. Thus, DAS has the potential to become a next-generation permanent network. DAS and other fiber sensing techniques can fill the observation gap in submarine environments and could have a profound impact on the comprehensive branches of observational seismology.
However, owing to the limitations of different types of instruments, it is unrealistic to expect one all-encompassing solution. The designing of monitoring systems depends on the different application scenarios, for example, on continents or ocean floor, regional or local scale, temporary or permanent, and frequency/sensitivity ranges of interest. Moreover, we need to understand the data characteristics of the emerging instruments, explore the best practice of instrument combinations in different application scenarios, develop new software infrastructures suitable for large data sizes, and design new methods to maximize the use of numerous channels. Future progress in these directions will facilitate the improvement of modern earthquake monitoring and eventually contribute to an improved understanding of earthquake phenomena and the Earth’s interior.
The study is supported by the USTC Research Funds of the Double First-Class Initiative (No. YD2080002006). I would like to thank the two anonymous reviewers for their constructive comments, which helped improve this manuscript. I would also like to thank Xin Cui for helping with the figures. The network data of the SCSN, Hi-Net, and P-Alert were accessed from the Southern California Seismic Data Center (https://scedc.caltech.edu/, doi:10.7909/C3WD3xH1, last accessed March 25, 2021), National Research Institute for Earth Science and Disaster Resilience (https://www.hinet.bosai.go.jp/, last accessed March 25, 2021), and the P-Alert Team (https://palert.earth.sinica.edu.tw/, last accessed March 25, 2021), respectively. The data of the world’s submarine cables were accessed from TeleGeography (https://www.submarinecablemap.com, last accessed March 25, 2021).
|
Allam AA, and Ben-Zion Y (2012). Seismic velocity structures in the southern California plate-boundary environment from double-difference tomography. Geophys J Int 190(2): 1181 –1196. doi: 10.1111/j.1365-246x.2012.05544.x
|
|
Allen R (1982). Automatic phase pickers: Their present use and future prospects. Bull Seismol Soc Amer 72(6B): S225–S242.
|
|
Ben-Zion Y, Vernon FL, Ozakin Y, Zigone D, Ross ZE, Meng HR, White M, Reyes J, Hollis D, and Barklage M (2015). Basic data features and results from a spatially dense seismic array on the San Jacinto fault zone. Geophys J Int 202(1): 370–380. doi: 10.1093/gji/ggv142
|
|
Beskardes GD, Wu Q, Hole JA, Chapman MC, Davenport KK, Brown LD, and Quiros DA (2019). Aftershock sequence of the 2011 Virginia Earthquake derived from the dense AIDA array and backprojection. Bull Seismol Soc Amer 109(1): 19–33. doi: 10.1785/0120180107
|
|
Bock Y, Nikolaidis RM, Jonge PJ de, and Bevis M (2000). Instantaneous geodetic positioning at medium distances with the Global Positioning System. J Geophys Res: Solid Earth 105(B12): 28223 – 28253 . doi: 10.1029/2000jb900268
|
|
Brenguier F, Kowalski P, Ackerley N, Nakata N, Boué P, Campillo M, Larose E, Ramboud S, Pequegnat C, Lecocq T, Roux P, Ferrazzini V, Villeneuve N, Shapiro NM, and Chaput J (2015). Toward 4D noise-based seismic probing of volcanoes: Perspectives from a large-N experiment on Piton de la Fournaise Volcano. Seismol Res Lett 87(1): 15–25. doi: 10.1785/0220150173
|
|
Buurman H, and West ME (2010). Seismic precursors to volcanic explosions during the 2006 eruption of Augustine Volcano: Chapter 2. USGS Numbered Series No. 1769–2, p. 4157
|
|
Catchings RD, Goldman MR, Steidl JH, Chan JH, Allam AA, Criley CJ, Ma ZN, Langermann DS, Huddleston GJ, McEvilly AT, Mongovin DD, Berg EM, and Ben-Zion Y (2020). Nodal seismograph recordings of the 2019 Ridgecrest earthquake sequence. Seismol Res Lett 91(6): 3 622−3 633
|
|
Clinton JF, and Heaton TH (2002). Potential advantages of a strong-motion velocity meter over a strong-motion accelerometer. Seismol Res Lett 73(3): 332–342. doi: 10.1785/gssrl.73.3.332
|
|
Cochran ES, Lawrence JF, Christensen C, and Jakka RS (2009). The Quake-Catcher Network: Citizen science expanding seismic horizons. Seismol Res Lett 80(1): 26–30. doi: 10.1785/gssrl.80.1.26
|
|
Ekström G, and Stark CP (2013). Simple scaling of catastrophic landslide dynamics. Science 339(6126): 1416 – 1419 . doi: 10.1126/science.1232887
|
|
Evans JR, Kozák JT, and Jedlička P (2016). Developments in new fluid rotational seismometers: Instrument performance and future directions. Bull Seismol Soc Amer 106(6): 2865 –2876. doi: 10.1785/0120150265
|
|
Fan WY, and McGuire JJ (2018). Investigating microearthquake finite source attributes with IRIS Community wavefield demonstration experiment in Oklahoma. Geophys J Int 214(2): 1 072–1 087. doi: 10.1093/gji/ggy203
|
|
Gibbons SJ, and Ringdal F (2006). The detection of low magnitude seismic events using array-based waveform correlation. Geophys J Int 165(1): 149–166. doi: 10.1111/j.1365-246x.2006.02865.x
|
|
Goldstein P, Dodge D, Firpo M (2003). SAC2000: Signal processing and analysis tools for seismologists and engineers. In: Lee WHK, Kanamori H, Jennings PC, and Kisslinger C eds. IASPEI International Handbook of Earthquake and Engineering Seismology. Academic Press, London
|
|
Hansen SM, and Schmandt B (2015). Automated detection and location of microseismicity at Mount St. Helens with a large-N geophone array. Geophys Res Lett 42(18): 7390 – 7397 . doi: 10.1002/2015gl064848
|
|
Hauksson E (2005). Southern California hypocenter relocation with waveform cross-correlation. Part 1: Results using the double-difference method. Bull Seismol Soc Amer 95(3): 896–903. doi: 10.1785/0120040167
|
|
Hauksson E, Yang W, and Shearer PM (2012). Waveform relocated earthquake catalog for Southern California (1981 to June 2011). Bull Seismol Soc Amer 102(5): 2239 – 2244 . doi: 10.1785/0120120010
|
|
Hauksson E, Yoon C, Yu E, Andrews JR, Alvarez M, Bhadha R, and Thomas V (2020). Caltech/USGS Southern California Seismic Network (SCSN) and Southern California Earthquake Data Center (SCEDC): Data availability for the 2019 Ridgecrest sequence. Seismol Res Lett 91(4): 1961 –1970. doi: 10.1785/0220190290
|
|
Huang HH, Lin FC, Schmandt B, Farrell J, Smith RB, and Tsai VC (2015). The Yellowstone magmatic system from the mantle plume to the upper crust. Science 348(6236): 773–776. doi: 10.1126/science.aaa5648
|
|
Hutton K, Woessner J, and Hauksson E (2010). Earthquake monitoring in Southern California for seventy-seven Years (1932-2008). Bull Seismol Soc Amer 100(2): 423–446. doi: 10.1785/0120090130
|
|
Inbal A, Ampuero JP, and Clayton RW (2016). Localized seismic deformation in the upper mantle revealed by dense seismic arrays. Science 354(6308): 88–92. doi: 10.1126/science.aaf1370
|
|
Inbal A, Clayton RW, and Ampuero JP (2015). Imaging widespread seismicity at midlower crustal depths beneath Long Beach, CA, with a dense seismic array: Evidence for a depth-dependent earthquake size distribution. Geophys Res Lett 42(15): 6 314–6 323. doi: 10.1002/2015gl064942
|
|
Inbal A, Kong Q, Savran W, and Allen RM (2019). On the feasibility of using the dense MyShake smartphone array for earthquake location. Seismol Res Lett 90(3): 1209 – 1218 . doi: 10.1785/0220180349
|
|
Ishii M, Shearer PM, Houston H, and Vidale JE (2005). Extent, duration and speed of the 2004 Sumatra-Andaman earthquake imaged by the Hi-Net array. Nature 435(7044): 933–936. doi: 10.1038/nature03675
|
|
Jousset P, Reinsch T, Ryberg T, Blanck H, Clarke A, Aghayev R, Hersir GP, Henninges J, Weber M, and Krawczyk CM (2018). Dynamic strain determination using fibre-optic cables allows imaging of seismological and structural features. Nature Communications 9(1): 1–11. doi: 10.1038/s41467-018-04860-y
|
|
Karplus M, and Schmandt B (2018). Preface to the focus section on geophone array seismology. Seismol Res Lett 89(5): 1597 –1600. doi: 10.1785/0220180212
|
|
Kim WY, Sykes LR, Armitage JH, Xie JK, Jacob KH, Richards PG, West M, Waldhauser F, Armbruster J, Seeber L, Du WX, Lerner-Lam A (2001). Seismic waves generated by aircraft impacts and building collapses at World Trade Center, New York City. Eos Trans AGU 82(47): 565–571. doi: 10.1029/01eo00330
|
|
Klein FW (2002). User’s guide to HYPOINVERSE-2000, A Fortran program to solve for earthquake locations and magnitudes. U SGS, Vol. 2002–171, p. 123
|
|
Kong QK, Allen RM, Schreier L, and Kwon YW (2016). MyShake: A smartphone seismic network for earthquake early warning and beyond. Science Advances 2(2): e1501055. doi: 10.1126/sciadv.1501055
|
|
Larson KM, Bodin P, and Gomberg J (2003). Using 1-Hz GPS data to measure deformations caused by the denali fault earthquake. Science 300(5624): 1421 – 1424 . doi: 10.1126/science.1084531
|
|
Li CY, Li ZF, Peng ZG, Zhang CY, Nakata N, and Sickbert T (2018). Long-period long-duration events detected by the IRIS Community wavefield demonstration experiment in Oklahoma: Tremor or train signals. Seismol Res Lett 89(5): 1 652–1 659. doi: 10.1785/0220180081
|
|
Li ZF, and Zhan ZW (2018). Pushing the limit of earthquake detection with distributed acoustic sensing and template matching: A case study at the Brady geothermal field. Geophys J Int 215(3): 1 583–1 593. doi: 10.1093/gji/ggy359
|
|
Li ZF, Peng ZG, Hollis D, Zhu LJ, and McClellan J (2018). High-resolution seismic event detection using local similarity for Large-N arrays. Scientific Reports 8(1): 1646 . doi: 10.1038/s41598-018-19728-w
|
|
Li ZF, Shen ZC, Yang Y, Williams E, Wang X, and Zhan ZW (2021). Rapid response to the 2019 Ridgecrest earthquake with distributed acoustic sensing. AGU Advances 2(2): e2021AV000395. doi: 10.1029/2021AV000395
|
|
Lin CH, Kumagai H, Ando M, and Shin TC (2010). Detection of landslides and submarine slumps using broadband seismic networks. Geophys Res Lett 37(22): L22309. doi: 10.1029/2010gl044685
|
|
Lin CJ, Huang, HP, Pham ND, Liu CC, Chi WC, and Lee WHK (2011). Rotational motions for teleseismic surface waves. Geophys Res Lett 38(15): L15301. doi: 10.1029/2011gl047959
|
|
Lin FC, Li D, Clayton RW, and Hollis D (2013). High-resolution 3D shallow crustal structure in Long Beach, California: Application of ambient noise tomography on a dense seismic array. Geophysics 78(4): Q45–Q56. doi: 10.1190/geo2012-0453.1
|
|
Lindsey NJ, Martin ER, Dreger DS, Freifeld B, Cole S, James SR, Biondi BL, Ajo-Franklin JB (2017). Fiber-Optic network observations of earthquake wavefields. Geophys Res Lett 44(23): 11 792–11 799. doi: 10.1002/2017gl075722
|
|
Lindsey NJ, Dawe TC, and Ajo-Franklin JB (2019). Illuminating seafloor faults and ocean dynamics with dark fiber distributed acoustic sensing. Science 366(6469): 1 103–1 107. doi: 10.1126/science.aay5881
|
|
Lindsey NJ, Yuan SY, Lellouch A, Gualtieri L, Lecocq T, and Biondi BL (2020a). City-scale dark Fiber DAS measurements of infrastructure use during the COVID-19 Pandemic. Geophys Res Lett 47(16): e2020GL089931. doi: 10.1029/2020gl089931
|
|
Lindsey NJ, Rademacher H, and Ajo-Franklin JB (2020b). On the broadband instrument response of fiber-optic DAS arrays. J Geophys Res: Solid Earth 125(2): e2019JB018145
|
|
Ma X, Zhao LF, Xie XB, He X, and Yao ZX (2020). Regional seismic characteristics of chemical explosions on the eastern margin of the Junggar Basin, Northwest China, and of historical semipalatinsk nuclear tests. Bull Seismol Soc Amer 111(1): 606–620. doi: 10.1785/0120200151
|
|
Marra G, Clivati C, Luckett R, Tampellini A, Kronjäger J, Wright L, Mura A, Levi F, Robinson S, Xuereb A, Baptie B, and Calonico D (2018). Ultrastable laser interferometry for earthquake detection with terrestrial and submarine cables. Science 361(6401): 486–490. doi: 10.1126/science.aat4458
|
|
Masoudi A, and Newson TP (2016). Contributed review: Distributed optical fibre dynamic strain sensing. Review of Scientific Instruments 87(1): 011501 . doi: 10.1063/1.4939482
|
|
Meng HR, and Ben-Zion Y (2018). Characteristics of airplanes and helicopters recorded by a dense seismic array near Anza California. J Geophys Res: Solid Earth 123(6): 4783 – 4797 . doi: 10.1029/2017jb015240
|
|
Nanjo KZ, Ishibe T, Tsuruoka H, Schorlemmer D, Ishigaki Y, and Hirata N (2010). Analysis of the completeness magnitude and seismic network coverage of Japan. Bull Seismol Soc Amer 100(6): 3 261–3 268. doi: 10.1785/0120100077
|
|
Obara K (2002). Nonvolcanic deep tremor associated with subduction in Southwest Japan. Science 296(5573): 1 679–1 681. doi: 10.1126/science.1070378
|
|
Okada Y, Kasahara K, Hori S, Obara K, Sekiguchi S, Fujiwara H, and Yamamoto A (2004). Recent progress of seismic observation networks in Japan —Hi-net, F-net, K-NET and KiK-net—. Earth Planet Space 56(8): xv–xxviii. doi: 10.1186/bf03353076
|
|
Pankow KL, Rusho J, Pechmann JC, Hale JM, Whidden K, Sumsion R, Holt J, Mesimeri M, Wells D, and Koper KD (2021). Responding to the 2020 Magna, Utah, earthquake sequence during the COVID-19 Pandemic Shutdown. Seismol Res Lett 92(1): 6–16. doi: 10.1785/0220200265
|
|
Peng ZG, and Zhao P (2009). Migration of early aftershocks following the 2004 Parkfield earthquake. Nature Geoscience 2(12): 877–881. doi: 10.1038/ngeo697
|
|
Peterson JR, and Hutt CR (2014). World-Wide Standardized Seismograph Network: A Data Users Guide. U.S. Geological Survey Open-File Report 2014-1218, p. 74. doi: 10.3133/ofr20141218
|
|
Reinwald M, Bernauer M, Igel H, and Donner S (2016). Improved finite-source inversion through joint measurements of rotational and translational ground motions: A numerical study. Solid Earth 7(5): 1 467–1 477. doi: 10.5194/se-7-1467-2016
|
|
Riahi N, and Gerstoft P (2015). The seismic traffic footprint: Tracking trains, aircraft, and cars seismically. Geophys Res Lett 42(8): 2 674–2 681. doi: 10.1002/2015gl063558
|
|
Ross ZE, Trugman DT, Hauksson E, and Shearer PM (2019). Searching for hidden earthquakes in Southern California. Science 364(6442): 767–771. doi: 10.1126/science.aaw6888
|
|
Ruhl CJ, Melgar D, Grapenthin R, and Allen RM (2017). The value of real-time GNSS to earthquake early warning. Geophys Res Lett 44(16): 8 311–8 319. doi: 10.1002/2017gl074502
|
|
Schmandt B, and Clayton RW (2013). Analysis of teleseismic P waves with a 5200-station array in Long Beach, California: Evidence for an abrupt boundary to Inner Borderland rifting. J Geophys Res: Solid Earth 118(10): 5 320–5 338. doi: 10.1002/jgrb.50370
|
|
Schreiber KU, Hautmann JN, Velikoseltsev A, Wassermann J, Igel H, Otero J, Vernon F, and Wells JPR (2009). Ring Laser measurements of ground rotations for seismology. Bull Seismol Soc Amer 99(2B): 1 190–1 198. doi: 10.1785/0120080171
|
|
Shearer PM (2009). Introduction to Seismology. The 2nd ed. Cambridge University Press, Cambridge, pp. 1−412
|
|
Simons FJ, Nolet G, Babcock JM, Davis RE, and Orcutt JA (2006). A future for drifting seismic networks. Eos Trans AGU 87(31): 305–307. doi: 10.1029/2006eo310002
|
|
Sladen A, Rivet D, Ampuero JP, De Barros L, Hello Y, Calbris G, and Lamare P (2019). Distributed sensing of earthquakes and ocean-solid Earth interactions on seafloor telecom cables. Nature Communications 10(1): 5 777. doi: 10.1038/s41467-019-13793-z
|
|
Stump BW (1991). Nuclear explosion seismology: Verification, source theory, wave propagation and politics. Reviews of Geophysics 29(S2): 734–742. doi: 10.1002/rog.1991.29.s2.734
|
|
Sumy DF, Woodward R, Barklage M, Hollis D, Spriggs N, Gridley JM, and Parker T (2015). Sweetwater, Texas Large N Experiment. AGU Fall Meeting Abstracts, San Francisco, CA, 14-18 Dec, S41A–2703.
|
|
Thurber CH (1983). Earthquake locations and three-dimensional crustal structure in the Coyote Lake Area, central California. J Geophys Res: Solid Earth 88(B10): 8 226–8 236. doi: 10.1029/jb088ib10p08226
|
|
Trugman DT, and Ross ZE (2019). Pervasive foreshock activity across Southern California. Geophys Res Lett 46(15): 8 772–8 781. doi: 10.1029/2019gl083725
|
|
Trugman DT, and Shearer PM (2017). GrowClust: A hierarchical clustering algorithm for relative earthquake relocation, with application to the Spanish Springs and Sheldon, Nevada, earthquake sequences. Seismol Res Lett 88(2A): 379–391. doi: 10.1785/0220160188
|
|
Uehira K, Mochizuki M, Kanazawa T, Shiomi K, Kunugi T, Aoi S, Matsumoto T, Sekiguchi S, Yamamoto N, Takahashi N, Shinohara M, and Yamada T (2018). S-net project: Construction of large-scale seismic and tsunami observation system on seafloor along the Japan Trench, 20, 12000. Geophysical Research Abstracts 20: 12000
|
|
Waldhauser F and Ellsworth WL (2000). A double-difference earthquake location algorithm: Method and application to the Northern Hayward Fault, California. Bull Seismol Soc Amer 90(6): 1 353–1 368. doi: 10.1785/0120000006
|
|
Walter JI, Ogwari P, Thiel A, Ferrer F, Woelfel I, Chang JC, Darold AP, and Holland AA (2019). The Oklahoma Geological Survey Statewide Seismic Network. Seismol Res Lett 91(2A): 611–621. doi: 10.1785/0220190211
|
|
Wang HF, Zeng X, Miller DE, Fratta D, Feigl KL, Thurber CH, and Mellors RJ (2018). Ground motion response to an ML4.3 earthquake using co-located distributed acoustic sensing and seismometer arrays. Geophys J Int 213(3): 2020 – 2036 . doi: 10.1093/gji/ggy102
|
|
Wang X, Williams EF, Karrenbach M, Herráez MG, Martins HF, and Zhan ZW (2020). Rose Parade Seismology: Signatures of Floats and Bands on Optical Fiber. Seismol Res Lett 91(4): 2 395–2 398. doi: 10.1785/0220200091
|
|
Williams EF, Fernández-Ruiz MR, Magalhaes R, Vanthillo R, Zhan ZW, González-Herráez M, and Martins H F (2019). Distributed sensing of microseisms and teleseisms with submarine dark fibers. Nature Communications 10(1): 5778. doi: 10.1038/s41467-019-13262-7
|
|
Winter K, Lombardi D, Diaz-Moreno A, and Bainbridge R (2021). Monitoring icequakes in East Antarctica with the Raspberry Shake. Seismol Res Lett, doi: 10.1785/0220200483
|
|
Wu YM, Liang WT, Mittal H, Chao WA, Lin CH, Huang BS, and Lin CM (2016). Performance of a low-cost earthquake early warning system (P-Alert) during the 2016 ML6.4 Meinong (Taiwan) Earthquake. Seismol Res Lett 87(5): 1050– 1059 . doi: 10.1785/0220160058
|
|
Wu YM, Chen DY, Lin TL, Hsieh CY, Chin TL, Chang WY, Li WS, Ker SH (2013). A high-density seismic network for earthquake early warning in Taiwan based on low cost sensors. Seismol Res Lett 84(6): 1 048–1 054. doi: 10.1785/0220130085
|
|
Yang L, Liu X, and Beroza GC (2021). Revisiting evidence for widespread seismicity in the upper mantle under Los Angeles. Science Advances 7(4): eabf2862. doi: 10.1126/sciadv.abf2862
|
|
Yu CQ, Zhan ZW, Lindsey NJ, Ajo-Franklin JB, and Robertson M (2019). The potential of DAS in teleseismic studies: Insights from the Goldstone experiment. Geophys Res Lett 46(3): 1 320–1 328. doi: 10.1029/2018gl081195
|
|
Yue H, and Lay T (2011). Inversion of high-rate (1 sps) GPS data for rupture process of the 11 March 2011 Tohoku earthquake (MW9.1). Geophys Res Lett 38(7): L00G09. doi: 10.1029/2011gl048700
|
|
Zhan ZW (2020). Distributed acoustic sensing turns fiber-optic cables into sensitive seismic antennas. Seismol Res Lett 91(1): 1–15. doi: 10.1785/0220190112
|
|
Zhan ZW, Cantono M, Kamalov V, Mecozzi A, Müller R, Yin S, and Castellanos JC (2021). Optical polarization-based seismic and water wave sensing on transoceanic cables. Science 371(6532): 931–936. doi: 10.1126/science.abe6648
|
震球期刊
EQS
Supported by: Beijing Renhe Information Technology Co., Ltd. 
| Feature | Broadband | Nodes | Micro-electromechanical systems (MEMS) | Distributed acoustic sensing (DAS) | Other optic sensing (M2018, Z2021*) |
| Typical spacing | 1–100 km | 100 m–10 km | 1–10 km | 1–10 m | N/A |
| # of sensors | 10–100 | 100–1000 | 100–1000 | 1000–10000 | N/A |
| Array Aperture | Flexible | Flexible | Flexible | < 100 km | Depend on fiber length |
| Sensitivity | High | Intermediate | Low | Intermediate | Intermediate |
| Power | Cable/solar | Battery | Cable/battery | Cable | Cable |
| Real-time telemetry | Yes | Mostly no | Yes/No | Yes | Yes |
| Submarine | Ocean Bottom Seismometer | No | No | Yes | Yes |
| Note: M2018 and Z2021 represent Marra et al. (2018) and Zhan ZW et al., (2021), respectively. | |||||
DownLoad:
CSV
