Alysha D. Armstrong, Zachary Claerhout, Ben Baker, Keith D. Koper
{"title":"A Deep-Learning Phase Picker with Calibrated Bayesian-Derived Uncertainties for Earthquakes in the Yellowstone Volcanic Region","authors":"Alysha D. Armstrong, Zachary Claerhout, Ben Baker, Keith D. Koper","doi":"10.1785/0120230068","DOIUrl":null,"url":null,"abstract":"ABSTRACT Traditional seismic phase pickers perform poorly during periods of elevated seismicity due to inherent weakness when detecting overlapping earthquake waveforms. This weakness results in incomplete seismic catalogs, particularly deficient in earthquakes that are close in space and time. Supervised deep-learning (DL) pickers allow for improved detection performance and better handle the overlapping waveforms. Here, we present a DL phase-picking procedure specifically trained on Yellowstone seismicity and designed to fit within the University of Utah Seismograph Stations (UUSS) real-time system. We modify and combine existing DL models to label the seismic phases in continuous data and produce better phase arrival times. We use transfer learning to achieve consistency with UUSS analysts while maintaining robust models. To improve the performance during periods of enhanced seismicity, we develop a data augmentation strategy to synthesize waveforms with two nearly coincident P arrivals. We also incorporate a model uncertainty quantification method, Multiple Stochastic Weight Averaging-Gaussian (MultiSWAG), for arrival-time estimates and compare it to dropout—a more standard approach. We use an efficient, model-agnostic method of empirically calibrating the uncertainties to produce meaningful 90% credible intervals. The credible intervals are used downstream in association, location, and quality assessment. For an in-depth evaluation of our automated method, we apply it to continuous data recorded from 25 March to 3 April 2014, on 20 three-component stations and 14 vertical-component stations. This 10-day period contains an Mw 4.8 event, the largest earthquake in the Yellowstone region since 1980. A seismic analyst manually examined more than 1000 located events, including ∼855 previously unidentified, and concluded that only two were incorrect. Finally, we present an analyst-created, high-resolution arrival-time data set, including 651 new arrival times, for one hour of data from station WY.YNR for robust evaluation of missed detections before association. Our method identified 60% of the analyst P picks and 81% of the S picks.","PeriodicalId":9444,"journal":{"name":"Bulletin of the Seismological Society of America","volume":"35 1","pages":"0"},"PeriodicalIF":2.6000,"publicationDate":"2023-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Bulletin of the Seismological Society of America","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1785/0120230068","RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"GEOCHEMISTRY & GEOPHYSICS","Score":null,"Total":0}
引用次数: 0
Abstract
ABSTRACT Traditional seismic phase pickers perform poorly during periods of elevated seismicity due to inherent weakness when detecting overlapping earthquake waveforms. This weakness results in incomplete seismic catalogs, particularly deficient in earthquakes that are close in space and time. Supervised deep-learning (DL) pickers allow for improved detection performance and better handle the overlapping waveforms. Here, we present a DL phase-picking procedure specifically trained on Yellowstone seismicity and designed to fit within the University of Utah Seismograph Stations (UUSS) real-time system. We modify and combine existing DL models to label the seismic phases in continuous data and produce better phase arrival times. We use transfer learning to achieve consistency with UUSS analysts while maintaining robust models. To improve the performance during periods of enhanced seismicity, we develop a data augmentation strategy to synthesize waveforms with two nearly coincident P arrivals. We also incorporate a model uncertainty quantification method, Multiple Stochastic Weight Averaging-Gaussian (MultiSWAG), for arrival-time estimates and compare it to dropout—a more standard approach. We use an efficient, model-agnostic method of empirically calibrating the uncertainties to produce meaningful 90% credible intervals. The credible intervals are used downstream in association, location, and quality assessment. For an in-depth evaluation of our automated method, we apply it to continuous data recorded from 25 March to 3 April 2014, on 20 three-component stations and 14 vertical-component stations. This 10-day period contains an Mw 4.8 event, the largest earthquake in the Yellowstone region since 1980. A seismic analyst manually examined more than 1000 located events, including ∼855 previously unidentified, and concluded that only two were incorrect. Finally, we present an analyst-created, high-resolution arrival-time data set, including 651 new arrival times, for one hour of data from station WY.YNR for robust evaluation of missed detections before association. Our method identified 60% of the analyst P picks and 81% of the S picks.
期刊介绍:
The Bulletin of the Seismological Society of America, commonly referred to as BSSA, (ISSN 0037-1106) is the premier journal of advanced research in earthquake seismology and related disciplines. It first appeared in 1911 and became a bimonthly in 1963. Each issue is composed of scientific papers on the various aspects of seismology, including investigation of specific earthquakes, theoretical and observational studies of seismic waves, inverse methods for determining the structure of the Earth or the dynamics of the earthquake source, seismometry, earthquake hazard and risk estimation, seismotectonics, and earthquake engineering. Special issues focus on important earthquakes or rapidly changing topics in seismology. BSSA is published by the Seismological Society of America.