skip to main content

Attention:

The NSF Public Access Repository (PAR) system and access will be unavailable from 11:00 PM ET on Friday, December 13 until 2:00 AM ET on Saturday, December 14 due to maintenance. We apologize for the inconvenience.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


  • NSF Public Access
  • Search Results
  • Toward Fully Autonomous Seismic Networks: Backprojecting Deep Learning-Based Phase Time Functions for Earthquake Monitoring on Continuous Recordings
Title: Toward Fully Autonomous Seismic Networks: Backprojecting Deep Learning-Based Phase Time Functions for Earthquake Monitoring on Continuous Recordings
Abstract Accurate and (near) real-time earthquake monitoring provides the spatial and temporal behaviors of earthquakes for understanding the nature of earthquakes, and also helps in regional seismic hazard assessments and mitigations. Because of the increase in both the quality and quantity of seismic data, an automated earthquake monitoring system is needed. Most of the traditional methods for detecting earthquake signals and picking phases are based on analyses of features in recordings of an individual earthquake and/or their differences from background noises. When seismicity is high, the seismograms are complicated, and, therefore, traditional analysis methods often fail. With the development of machine learning algorithms, earthquake signal detection and seismic phase picking can be more accurate using the features obtained from a large amount of earthquake recordings. We have developed an attention recurrent residual U-Net algorithm, and used data augmentation techniques to improve the accuracy of earthquake detection and seismic phase picking on complex seismograms that record multiple earthquakes. The use of probability functions of P and S arrivals and potential P and S arrival pairs of earthquakes can increase the computational efficiency and accuracy of backprojection for earthquake monitoring in large areas. We applied our workflow to monitor the earthquake activity in southern California during the 2019 Ridgecrest sequence. The distribution of earthquakes determined by our method is consistent with that in the Southern California Earthquake Data Center (SCEDC) catalog. In addition, the number of earthquakes in our catalog is more than three times that of the SCEDC catalog. Our method identifies additional earthquakes that are close in origin times and/or locations, and are not included in the SCEDC catalog. Our algorithm avoids misidentification of seismic phases for earthquake location. In general, our algorithm can provide reliable earthquake monitoring on a large area, even during a high seismicity period.  more » « less
Award ID(s):
1725729
PAR ID:
10396307
Author(s) / Creator(s):
; ; ;
Date Published:
Journal Name:
Seismological Research Letters
Volume:
93
Issue:
3
ISSN:
0895-0695
Page Range / eLocation ID:
1880 to 1894
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ( , Seismological Research Letters)
    null (Ed.)
    Abstract Seismograms are convolution results between seismic sources and the media that seismic waves propagate through, and, therefore, the primary observations for studying seismic source parameters and the Earth interior. The routine earthquake location and travel-time tomography rely on accurate seismic phase picks (e.g., P and S arrivals). As data increase, reliable automated seismic phase-picking methods are needed to analyze data and provide timely earthquake information. However, most traditional autopickers suffer from low signal-to-noise ratio and usually require additional efforts to tune hyperparameters for each case. In this study, we proposed a deep-learning approach that adapted soft attention gates (AGs) and recurrent-residual convolution units (RRCUs) into the backbone U-Net for seismic phase picking. The attention mechanism was implemented to suppress responses from waveforms irrelevant to seismic phases, and the cooperating RRCUs further enhanced temporal connections of seismograms at multiple scales. We used numerous earthquake recordings in Taiwan with diverse focal mechanisms, wide depth, and magnitude distributions, to train and test our model. Setting the picking errors within 0.1 s and predicted probability over 0.5, the AG with recurrent-residual convolution unit (ARRU) phase picker achieved the F1 score of 98.62% for P arrivals and 95.16% for S arrivals, and picking rates were 96.72% for P waves and 90.07% for S waves. The ARRU phase picker also shown a great generalization capability, when handling unseen data. When applied the model trained with Taiwan data to the southern California data, the ARRU phase picker shown no cognitive downgrade. Comparing with manual picks, the arrival times determined by the ARRU phase picker shown a higher consistency, which had been evaluated by a set of repeating earthquakes. The arrival picks with less human error could benefit studies, such as earthquake location and seismic tomography. 
    more » « less
  2. ; ; ; ; ( , Seismological Research Letters)
    Abstract

    Seismicity at restless volcanoes commonly features a variety of signal types reflecting both volcanotectonic and fluid-driven source processes. However, traditional catalogs of seismicity are often incomplete, especially concerning events with emergent onsets such as those driven by the dynamics of magmatic and hydrothermal fluids. The detection of all discrete events and continuous seismic tremors, regardless of the underlying source processes, would therefore improve the ability of monitoring agencies to forecast eruptions and mitigate their associated hazards. We present a workflow for generalized detection of seismic events based on the network covariance matrix (Seydoux et al., 2016). Our contributions enable the method to simultaneously detect continuous and short-duration (<∼10 s) events, provide information about the frequency content of the signals, and to refine the initial detection times by an order of magnitude (from window lengths of 75 to 7.5 s). We test the workflow on a 15-month record of seismicity with 23 stations at Mammoth Mountain, California (July 2012–October 2013) and detect 62% of long-period events and 94% of volcanotectonic events in the existing Northern California Earthquake Data Center catalog. In addition, ∼3000 events are not included in the catalog, and thousands of tremor signals are found. The method is suitable for near-real-time analysis of continuous waveforms and can provide a valuable supplement to existing algorithms to improve the completeness of catalogs used for monitoring volcanoes.

     
    more » « less
  3. ; ; ; ( , Journal of Geophysical Research: Solid Earth)
    Abstract

    We derive new, 3D, isotropic models of seismic compressional and shear wavespeeds, Vp and Vs, respectively, their ratio, Vp/Vs, and a catalog of relocated earthquakes for Southern California from more than 10 million P‐ and S‐wave arrivals associated with over 0.3 million earthquakes that occurred between 2000 and 2020. We augment high‐quality analyst‐reviewed phase arrival picks from the Southern California Earthquake Data Center with S‐wave arrival picks obtained with an automated algorithm, and we derive new wavespeed models via traveltime tomography formulated using Poisson‐Voronoi cells (Fang et al., 2020,https://doi.org/10.1785/0220190141). The results contribute to improved regional wavespeed models, particularly the Vp/Vs model, and absolute event locations. The obtained models correlate well with regional geological features and yield more accurate synthetic waveforms than other regional models do for waves with periods shorter than 5 s in much of the modeled region. The derived event catalog exhibits tighter spatial clustering than the standard regional catalog, thereby helping to characterize subsurface features of major faults. The regional 1D averaged Vp/Vs ratio shows high values at shallow depths, decreases to a minimum at about 10 km, then increases again at greater depths below 15 km. Deep seismicity correlates well with regions of Vp/Vs ratio lower than 1.75, which may indicate an increased brittle‐to‐ductile transition depth with an important influence on crustal mechanics. The new wavespeed models and seismic catalog can be useful for various studies including analyses of seismicity patterns and simulations of crustal deformation and ground motion.

     
    more » « less
  4. ; ; ( , Journal of Geophysical Research: Solid Earth)
    Abstract

    We develop an automated processing procedure to derive a new catalog of earthquake locations, magnitudes, and potencies and analyze 9 years of data between 2008 and 2016 in the San Jacinto fault‐zone region. Our procedure accounts for detailed 3‐D velocity structure using a probabilistic global‐search location inversion and obtains high‐precision relative event locations using differential travel times measured by cross‐correlating waveforms. The obtained catalog illuminates spatiotemporal seismicity patterns in the fault zone with observations for 108,800 earthquakes in the magnitude range −1.8 to 5.4. Inside a focus region consisting of an 80‐km by 50‐km rectangle oriented parallel to the main fault trace, we estimate a 99% detection rate of earthquakes with magnitude 0.6 and greater and detect and locate about 60% more events than those present in the Southern California Seismic Network catalog. The results provide the most complete catalog available for the focused study region during the analyzed period and include both deeper events and very shallow patches of seismicity not present in the regional catalog. The seismicity exhibits a variety of complex patterns that contain important information on deformation processes in the region. The fraction of event pairs with waveforms having cross‐correlation coefficients ≥0.95 is only about 3%, indicating diverse processes operating in the fault zone.

     
    more » « less
  5. ; ; ( , Seismological Research Letters)
    Abstract We present the high-resolution Parkfield matched filter relocated earthquake (PKD-MR) catalog for the 2004 Mw 6 Parkfield earthquake sequence in central California. We use high-quality seismic data recorded by the borehole High Resolution Seismic Network combined with matched filter detection and relocations from cross-correlation derived differential travel times. We determine the magnitudes of newly detected events by computing the amplitude ratio between the detections and templates using a principal component fit. The relocated catalog spans from 6 November 2003 to 28 March 2005 and contains 13,914 earthquakes, which is about three times the number of events listed in the Northern California Seismic Network catalog. Our results on the seismicity rate changes before the 2004 mainshock do not show clear precursory signals, although we find an increase in the seismic activity in the creeping section of the San Andreas fault (SAF) (about ∼30 km northwest of the mainshock epicenter) in the weeks prior to the mainshock. We also observe a decrease in the b-value parameter in the Gutenberg–Richter relationship in the creeping section in the weeks prior to the mainshock. Our results suggest stress is increasingly released seismically in the creeping section, accompanied by a decreasing aseismic creeping rate before the mainshock occurrence. However, b-value and seismicity rates remain stable in the Parkfield section where the 2004 mainshock ruptured. This updated catalog can be used to study the evolution of aftershocks and their relations to afterslip following the 2004 Parkfield mainshock, seismicity before the mainshock, and how external stresses interact with the Parkfield section of the SAF and the 2004 sequence. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email