Search for: All records

We present the first global-scale database of 4.3 billion P- and S-wave picks extracted from 1.3 PB continuous seismic data via a cloud-native workflow. Using cloud computing services on Amazon Web Services, we launched ~145,000 containerized jobs on continuous records from 47,354 stations spanning 2002-2025, completing in under three days. Phase arrivals were identified with a deep learning model, PhaseNet, through an open-source Python ecosystem for deep learning, SeisBench. To visualize and gain a global understanding of these picks, we present preliminary results about pick time series revealing Omori-law aftershock decay, seasonal variations linked to noise levels, and dense regional coverage that will enhance earthquake catalogs and machine-learning datasets. We provide all picks in a publicly queryable database, providing a powerful resource for researchers studying seismicity around the world. This report provides insights into the database and the underlying workflow, demonstrating the feasibility of petabyte-scale seismic data mining on the cloud and of providing intelligent data products to the community in an automated manner. 

Abstract Distributed acoustic sensing (DAS) on submarine fiber-optic cables is providing new observational insights into solid Earth processes and ocean dynamics. However, the availability of offshore dark fibers for long-term deployment remains limited. Simultaneous telecommunication and DAS operating at different wavelengths in the same fiber, termed optical multiplexing, offers one solution. In May 2024, we collected a four-day DAS dataset utilizing an L-band DAS interrogator and multiplexing on the submarine cables of the Ocean Observatory Initiative’s Regional Cabled Array offshore central Oregon. Our findings show that multiplexed DAS has no impact on communications and is unaffected by network traffic. Moreover, the quality of DAS data collected via multiplexing matches that of data obtained from dark fiber. With a machine-learning event detection workflow, we detect 31 T waves and the S wave of one regional earthquake, demonstrating the feasibility of continuous earthquake monitoring using the multiplexed offshore DAS. We also examine ocean waves and ocean-generated seismic noise. We note high-frequency seismic noise modulated by low-frequency ocean swell and hypothesize about its origins. The complete dataset is freely available. 

Abstract With the rise of data volume and computing power, seismological research requires more advanced skills in data processing, numerical methods, and parallel computing. We present the experience of conducting training workshops in various forms of delivery to support the adoption of large-scale high-performance computing (HPC) and cloud computing, advancing seismological research. The seismological foci were on earthquake source parameter estimation in catalogs, forward and adjoint wavefield simulations in 2D and 3D at local, regional, and global scales, earthquake dynamics, ambient noise seismology, and machine learning. This contribution describes the series of workshops delivered as part of research projects, the learning outcomes for participants, and lessons learned by the instructors. Our curriculum was grounded on open and reproducible science, large-scale scientific computing and data mining, and computing infrastructure (access and usage) for HPC and the cloud. We also describe the types of teaching materials that have proven beneficial to the instruction and the sustainability of the program. We propose guidelines to deliver future workshops on these topics. 

The Antarctic ice sheet is buttressed by floating ice shelves that calve icebergs along large fractures called rifts. Despite the significant influence exerted by rifting on ice shelf geometry and buttressing, the scarcity of in situ observations of rift propagation contributes considerable uncertainty to understanding rift dynamics. Here, we report the first‐ever seismic recording of a multiple‐kilometer rift propagation event. Remote sensing and seismic recordings reveal that a rift in the Pine Island Glacier Ice Shelf extended 10.53 km at a speed of 35.1 m/s, the fastest known ice fracture at this scale. We simulate ocean‐coupled rift propagation and find that the dynamics of water flow within the rift limit the propagation rate, resulting in rupture two orders of magnitude slower than typically predicted for brittle fracture. Using seismic recordings of the elastic waves generated during rift propagation, we estimate that ocean water flows into the rift at a rate of at least 2,300 cubic meters during rift propagation and causes mixing in the subshelf cavity. Our observations support the hypotheses that large ice shelf rift propagation events are brittle, hydrodynamically limited, and exhibit sensitive coupling with the surrounding ocean. 

Abstract Elevated seismic noise for moderate‐size earthquakes recorded at teleseismic distances has limited our ability to see their complexity. We develop a machine‐learning‐based algorithm to separate noise and earthquake signals that overlap in frequency. The multi‐task encoder‐decoder model is built around a kernel pre‐trained on local (e.g., short distances) earthquake data (Yin et al., 2022,https://doi.org/10.1093/gji/ggac290) and is modified by continued learning with high‐quality teleseismic data. We denoise teleseismic P waves of deep Mw5.0+ earthquakes and use the clean P waves to estimate source characteristics with reduced uncertainties of these understudied earthquakes. We find a scaling of moment and duration to beM₀ ≃ τ⁴, and a resulting strong scaling of stress drop and radiated energy with magnitude ( and ). The median radiation efficiency is 5%, a low value compared to crustal earthquakes. Overall, we show that deep earthquakes have weak rupture directivity and few subevents, suggesting a simple model of a circular crack with radial rupture propagation is appropriate. When accounting for their respective scaling with earthquake size, we find no systematic depth variations of duration, stress drop, or radiated energy within the 100–700 km depth range. Our study supports the findings of Poli and Prieto (2016,https://doi.org/10.1002/2016jb013521) with a doubled amount of earthquakes investigated and with earthquakes of lower magnitudes. 

The commercial cloud offers on-demand computational resources that could be revolutionary for the seismological community, especially as seismic datasets continue to grow. However, there are few educational examples for cloud use that target individual seismological researchers. Here, we present a reproducible earthquake detection and association workflow that runs on Microsoft Azure. The Python-based workflow runs on continuous time-series data using both template matching and machine learning. We provide tutorials for constructing cloud resources (both storage and computing) through a desktop portal and deploying the code both locally and remotely on the cloud resources. We report on scaling of compute times and costs to show that CPU-only processing is generally inexpensive, and is faster and simpler than using GPUs. When the workflow is applied to one year of continuous data from a mid-ocean ridge, the resulting earthquake catalogs suggest that template matching and machine learning are complementary methods whose relative performance is dependent on site-specific tectonic characteristics. Overall, we find that the commercial cloud presents a steep learning curve but is cost-effective. This report is intended as an informative starting point for any researcher considering migrating their own processing to the commercial cloud. 

Abstract Large-scale processing and dissemination of distributed acoustic sensing (DAS) data are among the greatest computational challenges and opportunities of seismological research today. Current data formats and computing infrastructure are not well-adapted or user-friendly for large-scale processing. We propose an innovative, cloud-native solution for DAS seismology using the MinIO open-source object storage framework. We develop data schema for cloud-optimized data formats—Zarr and TileDB, which we deploy on a local object storage service compatible with the Amazon Web Services (AWS) storage system. We benchmark reading and writing performance for various data schema using canonical use cases in seismology. We test our framework on a local server and AWS. We find much-improved performance in compute time and memory throughout when using TileDB and Zarr compared to the conventional HDF5 data format. We demonstrate the platform with a computing heavy use case in seismology: ambient noise seismology of DAS data. We process one month of data, pairing all 2089 channels within 24 hr using AWS Batch autoscaling. 

SUMMARY Seismograms contain multiple sources of seismic waves, from distinct transient signals such as earthquakes to continuous ambient seismic vibrations such as microseism. Ambient vibrations contaminate the earthquake signals, while the earthquake signals pollute the ambient noise’s statistical properties necessary for ambient-noise seismology analysis. Separating ambient noise from earthquake signals would thus benefit multiple seismological analyses. This work develops a multitask encoder–decoder network named WaveDecompNet to separate transient signals from ambient signals directly in the time domain for 3-component seismograms. We choose the active-volcanic Big Island in Hawai’i as a natural laboratory given its richness in transients (tectonic and volcanic earthquakes) and diffuse ambient noise (strong microseism). The approach takes a noisy 3-component seismogram as input and independently predicts the 3-component earthquake and noise waveforms. The model is trained on earthquake and noise waveforms from the STandford EArthquake Dataset (STEAD) and on the local noise of seismic station IU.POHA. We estimate the network’s performance by using the explained variance metric on both earthquake and noise waveforms. We explore different neural network designs for WaveDecompNet and find that the model with long-short-term memory (LSTM) performs best over other structures. Overall, we find that WaveDecompNet provides satisfactory performance down to a signal-to-noise ratio (SNR) of 0.1. The potential of the method is (1) to improve broad-band SNR of transient (earthquake) waveforms and (2) to improve local ambient noise to monitor the Earth’s structure using ambient noise signals. To test this, we apply a short-time average to a long-time average filter and improve the number of detected events. We also measure single-station cross-correlation functions of the recovered ambient noise and establish their improved coherence through time and over different frequency bands. We conclude that WaveDecompNet is a promising tool for a broad range of seismological research. 

Abstract We cluster a global database of 3529 Mw>5.5 earthquakes in 1995–2018 based on a dynamic time warping distance between earthquake source time functions (STFs). The clustering exhibits different degrees of complexity of the STF shapes and suggests an association between STF complexity and earthquake source parameters. Most of the thrust events have simple STF shapes across all depths. In contrast, earthquakes with complex STF shapes tend to be located at shallow depths in complicated tectonic regions, exhibit long source duration compared with others of similar magnitude, and tend to have strike-slip mechanisms. With 2D dynamic modeling of dynamic ruptures on heterogeneous fault properties, we find a systematic variation of the simulated STF complexity with frictional properties. Comparison between the observed and synthetic clustering distributions provides useful constraints on frictional properties. In particular, the characteristic slip-weakening distance could be constrained to be short (<0.1  m) and depth dependent if stress drop is in general constant. 

Abstract Megathrust earthquakes exhibit a ubiquitous seismic radiation style: low‐frequency (LF) seismic energy is efficiently emitted from the shallowest portion of the fault, whereas high‐frequency (HF) seismic energy is efficiently emitted from the deepest part of the fault. Although this is observed in many case‐specific studies, we show that it is ubiquitous in global megathrust earthquakes between 1995 and 2021. Previous studies have interpreted this as an effect of systematic depth variation in either the plate interface frictional properties (Lay et al., 2012) or the P wavespeeds (Sallarès & Ranero, 2019). This work suggests an alternative hypothesis: the interaction between waves and ruptures due to the Earth's free surface is the leading mechanism that generates this behavior. Two‐dimensional dynamic rupture simulations of subduction zone earthquakes support this hypothesis. Our simulations show that the interaction between the seismic waves reflected at the Earth's free surface and the updip propagating rupture results in LF radiation at the source. In contrast, the downdip propagation of rupture is less affected by the free surface and is thus dominated by HF radiation typical of buried faults. To a second degree, the presence of a realistic Earth structure derived from P‐wave velocity (V_P) tomographic images and realistic V_P/V_Sratio estimated in boreholes further enhances the contrast in source radiation. We conclude that the Earth's free surface is necessary to explain the observed megathrust earthquake radiation style, and the realistic structure of subduction zone is necessary to better predict earthquake ground motion and tsunami potential.