Seismic phase pickers based on deep neural networks have been extensively used recently,demonstrating their advantages on both performance and efficiency.However,these pickers are trained with and applied to different...Seismic phase pickers based on deep neural networks have been extensively used recently,demonstrating their advantages on both performance and efficiency.However,these pickers are trained with and applied to different data.A comprehensive benchmark based on a single dataset is therefore lacking.Here,using the recently released DiTing dataset,we analyzed performances of seven phase pickers with different network structures,the efficiencies are also evaluated using both CPU and GPU devices.Evaluations based on F1-scores reveal that the recurrent neural network(RNN)and EQTransformer exhibit the best performance,likely owing to their large receptive fields.Similar performances are observed among PhaseNet(UNet),UNet++,and the lightweight phase picking network(LPPN).However,the LPPN models are the most efficient.The RNN and EQTransformer have similar speeds,which are slower than those of the LPPN and PhaseNet.UNet++requires the most computational effort among the pickers.As all of the pickers perform well after being trained with a large-scale dataset,users may choose the one suitable for their applications.For beginners,we provide a tutorial on training and validating the pickers using the DiTing dataset.We also provide two sets of models trained using datasets with both 50 Hz and 100 Hz sampling rates for direct application by end-users.All of our models are open-source and publicly accessible.展开更多
Current deep neural networks(DNN)used for seismic phase picking are becoming more complex,which consumes much computing time without significant accuracy improvement.In this study,we introduce a cascaded classificatio...Current deep neural networks(DNN)used for seismic phase picking are becoming more complex,which consumes much computing time without significant accuracy improvement.In this study,we introduce a cascaded classification and regression framework for seismic phase picking,named as the classification and regression phase net(CRPN),which contains two convolutional neural network(CNN)models with different complexity to meet the requirements of accuracy and efficiency.The first stage of the CRPN are shallow CNNs used for rapid detection of seismic phase and picking P and S arrival times for earthquakes with magnitude larger than 2.0,respectively.The second stage of CRPN is used for high precision classification and regression.The regression is designed to reduce the time difference between the probability maximum and the real arrival time.After being trained using 500,000 P and S phases,the CRPN can process 400 hours’seismic data per second,whose sampling rate is 1 Hz and 25 Hz for the two stages,respectively,on a Nvidia K2200 GPU,and pick 93%P and 89%S phases with the error being reduced by 0.1s after regression correction.展开更多
Automatic phase picking is a critical procedure for seismic data processing, especially for a huge amount of seismic data recorded by a large-scale portable seismic array. In this study is presented a new method used ...Automatic phase picking is a critical procedure for seismic data processing, especially for a huge amount of seismic data recorded by a large-scale portable seismic array. In this study is presented a new method used for automatic accurate onset phase picking based on the proporty of dense seismic array observations. In our method, the Akaike＇s information criterion （AIC） for the single channel observation and the least-squares cross-correlation for the multi-channel observation are combined together. The tests by the seismic array observation data after triggering with the short-term average/long-term average （STA/LTA） technique show that the phase picking error is less than 0.3 s for local events by using the single channel AIC algorithm. In terms of multi-channel least-squares cross-correlation technique, the clear teleseismic P onset can be detected reliably. Even for the teleseismic records with high noise level, our algorithm is also able to effectually avoid manual misdetections.展开更多
Current popular deep learning seismic phase pickers like PhaseNet and EQTransformer suffer from performance drop in China.To mitigate this problem,we build a unified set of customized seismic phase pickers for differe...Current popular deep learning seismic phase pickers like PhaseNet and EQTransformer suffer from performance drop in China.To mitigate this problem,we build a unified set of customized seismic phase pickers for different levels of use in China.We first train a base picker with the recently released DiTing dataset using the same U-Net architecture as PhaseNet.This base picker significantly outperforms the original PhaseNet and is generally suitable for entire China.Then,using different subsets of the DiTing data,we fine-tune the base picker to better adapt to different regions.In total,we provide 5 pickers for major tectonic blocks in China,33 pickers for provincial-level administrative regions,and 2 special pickers for the Capital area and the China Seismic Experimental Site.These pickers show improved performance in respective regions which they are customized for.They can be either directly integrated into national or regional seismic network operation or used as base models for further refinement for specific datasets.We anticipate that this picker set will facilitate earthquake monitoring in China.展开更多
Reservoir earthquake characteristics such as small magnitude and large quantity may result in low monitoring efficiency when using traditional methods.However,methods based on deep learning can discriminate the seismi...Reservoir earthquake characteristics such as small magnitude and large quantity may result in low monitoring efficiency when using traditional methods.However,methods based on deep learning can discriminate the seismic phases of small earthquakes in a reservoir and ensure rapid processing of arrival time picking.The present study establishes a deep learning network model combining a convolutional neural network(CNN) and recurrent neural network(RNN).The neural network training uses the waveforms of 60 000 small earthquakes within a magnitude range of 0.8-1.2 recorded by 73 stations near the Dagangshan Reservoir in Sichuan Province as well as the data of the manually picked P-wave arrival time.The neural network automatically picks the P-wave arrival time,providing a strong constraint for small earthquake positioning.The model is shown to achieve an accuracy rate of 90.7 % in picking P waves of microseisms in the reservoir area,with a recall rate reaching 92.6% and an error rate lower than 2%.The results indicate that the relevant network structure has high accuracy for picking the P-wave arrival times of small earthquakes,thus providing new technical measures for subsequent microseismic monitoring in the reservoir area.展开更多
The phase identification and travel time picking are critical for seismic tomography,yet it will be challenging when the numbers of stations and earthquakes are huge.We here present a method to quickly obtain P and S ...The phase identification and travel time picking are critical for seismic tomography,yet it will be challenging when the numbers of stations and earthquakes are huge.We here present a method to quickly obtain P and S travel times of pre-determined earthquakes from mobile dense array with the aid from long term phase records from co-located permanent stations.The records for 1768 M≥2.0 events from 2011 to 2013 recorded by 350 ChinArray stations deployed in Yunnan Province are processed with an improved AR-AIC method utilizing cumulative envelope and rectilinearity.The reference arrivals are predicted based on phase records from 88 permanent stations with similar spatial coverage,which are further refined with AR-AIC.Totally,718573 P picks and 512035 S picks are obtained from mobile stations,which are 28 and 22 times of those from permanent stations,respectively.By comparing the automatic picks with manual picks from 88 permanent stations,for M≥3.0 events,81.5%of the P-pick errors are smaller than 0.5 second and 70.5%of S-pick errors are smaller than1 second.For events with a lower magnitude,76.5%P-pick errors fall into 0.5 second and 69.5%S-pick errors are smaller than 1 second.Moreover,the Pn and Sn phases are easily discriminated from directly P/S,indicating the necessity of combining traditional auto picking and integrating machine learning method.展开更多
High-quality datasets are critical for the development of advanced machine-learning algorithms in seismology.Here,we present an earthquake dataset based on the ChinArray Phase I records(X1).ChinArray Phase I was deplo...High-quality datasets are critical for the development of advanced machine-learning algorithms in seismology.Here,we present an earthquake dataset based on the ChinArray Phase I records(X1).ChinArray Phase I was deployed in the southern north-south seismic zone(20°N-32°N,95°E-110°E)in 2011-2013 using 355 portable broadband seismic stations.CREDIT-X1local,the first release of the ChinArray Reference Earthquake Dataset for Innovative Techniques(CREDIT),includes comprehensive information for the 105,455 local events that occurred in the southern north-south seismic zone during array observation,incorporating them into a single HDF5 file.Original 100-Hz sampled three-component waveforms are organized by event for stations within epicenter distances of 1,000 km,and records of≥200 s are included for each waveform.Two types of phase labels are provided.The first includes manually picked labels for 5,999 events with magnitudes≥2.0,providing 66,507 Pg,42,310 Sg,12,823 Pn,and 546 Sn phases.The second contains automatically labeled phases for 105,442 events with magnitudes of−1.6 to 7.6.These phases were picked using a recurrent neural network phase picker and screened using the corresponding travel time curves,resulting in 1,179,808 Pg,884,281 Sg,176,089 Pn,and 22,986 Sn phases.Additionally,first-motion polarities are included for 31,273 Pg phases.The event and station locations are provided,so that deep learning networks for both conventional phase picking and phase association can be trained and validated.The CREDIT-X1local dataset is the first million-scale dataset constructed from a dense seismic array,which is designed to support various multi-station deep-learning methods,high-precision focal mechanism inversion,and seismic tomography studies.Additionally,owing to the high seismicity in the southern north-south seismic zone in China,this dataset has great potential for future scientific discoveries.展开更多
基金jointly supported by the National Natural Science Foundation of China (No. 42074060)the Special Fund, Institute of Geophysics, China Earthquake Administration (CEA-IGP) (Nos. DQJB19B29, DQJB20B15, and DQJB22Z01)supported by XingHuo Project, CEA (No. XH211103)
文摘Seismic phase pickers based on deep neural networks have been extensively used recently,demonstrating their advantages on both performance and efficiency.However,these pickers are trained with and applied to different data.A comprehensive benchmark based on a single dataset is therefore lacking.Here,using the recently released DiTing dataset,we analyzed performances of seven phase pickers with different network structures,the efficiencies are also evaluated using both CPU and GPU devices.Evaluations based on F1-scores reveal that the recurrent neural network(RNN)and EQTransformer exhibit the best performance,likely owing to their large receptive fields.Similar performances are observed among PhaseNet(UNet),UNet++,and the lightweight phase picking network(LPPN).However,the LPPN models are the most efficient.The RNN and EQTransformer have similar speeds,which are slower than those of the LPPN and PhaseNet.UNet++requires the most computational effort among the pickers.As all of the pickers perform well after being trained with a large-scale dataset,users may choose the one suitable for their applications.For beginners,we provide a tutorial on training and validating the pickers using the DiTing dataset.We also provide two sets of models trained using datasets with both 50 Hz and 100 Hz sampling rates for direct application by end-users.All of our models are open-source and publicly accessible.
基金This work is financially supported by the National Key R&D Program of China(No.2018YFC1504200)National Natural Science Foundation of China(No.41661164035)+1 种基金the LU JIAXI International Team Program from the KC Wong Education Foundation and CAS(No.GJTD-2018-12)We thank two anonymous reviewers for their valuable comments and suggestions,which substantially improve the article.
文摘Current deep neural networks(DNN)used for seismic phase picking are becoming more complex,which consumes much computing time without significant accuracy improvement.In this study,we introduce a cascaded classification and regression framework for seismic phase picking,named as the classification and regression phase net(CRPN),which contains two convolutional neural network(CNN)models with different complexity to meet the requirements of accuracy and efficiency.The first stage of the CRPN are shallow CNNs used for rapid detection of seismic phase and picking P and S arrival times for earthquakes with magnitude larger than 2.0,respectively.The second stage of CRPN is used for high precision classification and regression.The regression is designed to reduce the time difference between the probability maximum and the real arrival time.After being trained using 500,000 P and S phases,the CRPN can process 400 hours’seismic data per second,whose sampling rate is 1 Hz and 25 Hz for the two stages,respectively,on a Nvidia K2200 GPU,and pick 93%P and 89%S phases with the error being reduced by 0.1s after regression correction.
基金National Natural Science Foundation of China (Grant No. 40234043).
文摘Automatic phase picking is a critical procedure for seismic data processing, especially for a huge amount of seismic data recorded by a large-scale portable seismic array. In this study is presented a new method used for automatic accurate onset phase picking based on the proporty of dense seismic array observations. In our method, the Akaike＇s information criterion （AIC） for the single channel observation and the least-squares cross-correlation for the multi-channel observation are combined together. The tests by the seismic array observation data after triggering with the short-term average/long-term average （STA/LTA） technique show that the phase picking error is less than 0.3 s for local events by using the single channel AIC algorithm. In terms of multi-channel least-squares cross-correlation technique, the clear teleseismic P onset can be detected reliably. Even for the teleseismic records with high noise level, our algorithm is also able to effectually avoid manual misdetections.
基金the National Key R&D Program of China(No.2021YFC3000700)the Special Fund of the Institute of Geophysics,China Earthquake Administration(Nos.DQJB22X08 and DQJB21Z05).
文摘Current popular deep learning seismic phase pickers like PhaseNet and EQTransformer suffer from performance drop in China.To mitigate this problem,we build a unified set of customized seismic phase pickers for different levels of use in China.We first train a base picker with the recently released DiTing dataset using the same U-Net architecture as PhaseNet.This base picker significantly outperforms the original PhaseNet and is generally suitable for entire China.Then,using different subsets of the DiTing data,we fine-tune the base picker to better adapt to different regions.In total,we provide 5 pickers for major tectonic blocks in China,33 pickers for provincial-level administrative regions,and 2 special pickers for the Capital area and the China Seismic Experimental Site.These pickers show improved performance in respective regions which they are customized for.They can be either directly integrated into national or regional seismic network operation or used as base models for further refinement for specific datasets.We anticipate that this picker set will facilitate earthquake monitoring in China.
基金supported by the National Key R&D Program of China(2018YFC1503200)the National Natural Science Foundation of China(41790463,41804063,42074060)the Scientific Research InstitutesBasic Research and Development Operations Special Fund of the Institute of Geophysics,China Earthquake Administration(DQJB19B29,DQJB20B27)。
文摘Reservoir earthquake characteristics such as small magnitude and large quantity may result in low monitoring efficiency when using traditional methods.However,methods based on deep learning can discriminate the seismic phases of small earthquakes in a reservoir and ensure rapid processing of arrival time picking.The present study establishes a deep learning network model combining a convolutional neural network(CNN) and recurrent neural network(RNN).The neural network training uses the waveforms of 60 000 small earthquakes within a magnitude range of 0.8-1.2 recorded by 73 stations near the Dagangshan Reservoir in Sichuan Province as well as the data of the manually picked P-wave arrival time.The neural network automatically picks the P-wave arrival time,providing a strong constraint for small earthquake positioning.The model is shown to achieve an accuracy rate of 90.7 % in picking P waves of microseisms in the reservoir area,with a recall rate reaching 92.6% and an error rate lower than 2%.The results indicate that the relevant network structure has high accuracy for picking the P-wave arrival times of small earthquakes,thus providing new technical measures for subsequent microseismic monitoring in the reservoir area.
基金This project is supported by the National Key Research and Development Program of China(2018YFC1503200)the Fundamental Research Funds for the Institute of Geophysics of China Earthquake Administration(DQJB19B29)+1 种基金the National Natural Science Foundation of China(41790463)the Science and Technology Projects of Zhejiang Earthquake Agency(2019zjj05)。
文摘The phase identification and travel time picking are critical for seismic tomography,yet it will be challenging when the numbers of stations and earthquakes are huge.We here present a method to quickly obtain P and S travel times of pre-determined earthquakes from mobile dense array with the aid from long term phase records from co-located permanent stations.The records for 1768 M≥2.0 events from 2011 to 2013 recorded by 350 ChinArray stations deployed in Yunnan Province are processed with an improved AR-AIC method utilizing cumulative envelope and rectilinearity.The reference arrivals are predicted based on phase records from 88 permanent stations with similar spatial coverage,which are further refined with AR-AIC.Totally,718573 P picks and 512035 S picks are obtained from mobile stations,which are 28 and 22 times of those from permanent stations,respectively.By comparing the automatic picks with manual picks from 88 permanent stations,for M≥3.0 events,81.5%of the P-pick errors are smaller than 0.5 second and 70.5%of S-pick errors are smaller than1 second.For events with a lower magnitude,76.5%P-pick errors fall into 0.5 second and 69.5%S-pick errors are smaller than 1 second.Moreover,the Pn and Sn phases are easily discriminated from directly P/S,indicating the necessity of combining traditional auto picking and integrating machine learning method.
基金funded by the National Key R&D Program of China (No. 2021YFC3000702)the Special Fund of the Institute of Geophysics, China Earthquake Administration (No. DQJB20B15)+2 种基金the National Natural Science Foundation of China youth Grant (No. 41804059)the Joint Funds of the National Natural Science Foundation of China (No. U223920029)the Science for Earthquake Resilience of China Earthquake Administration (No. XH211103)
文摘High-quality datasets are critical for the development of advanced machine-learning algorithms in seismology.Here,we present an earthquake dataset based on the ChinArray Phase I records(X1).ChinArray Phase I was deployed in the southern north-south seismic zone(20°N-32°N,95°E-110°E)in 2011-2013 using 355 portable broadband seismic stations.CREDIT-X1local,the first release of the ChinArray Reference Earthquake Dataset for Innovative Techniques(CREDIT),includes comprehensive information for the 105,455 local events that occurred in the southern north-south seismic zone during array observation,incorporating them into a single HDF5 file.Original 100-Hz sampled three-component waveforms are organized by event for stations within epicenter distances of 1,000 km,and records of≥200 s are included for each waveform.Two types of phase labels are provided.The first includes manually picked labels for 5,999 events with magnitudes≥2.0,providing 66,507 Pg,42,310 Sg,12,823 Pn,and 546 Sn phases.The second contains automatically labeled phases for 105,442 events with magnitudes of−1.6 to 7.6.These phases were picked using a recurrent neural network phase picker and screened using the corresponding travel time curves,resulting in 1,179,808 Pg,884,281 Sg,176,089 Pn,and 22,986 Sn phases.Additionally,first-motion polarities are included for 31,273 Pg phases.The event and station locations are provided,so that deep learning networks for both conventional phase picking and phase association can be trained and validated.The CREDIT-X1local dataset is the first million-scale dataset constructed from a dense seismic array,which is designed to support various multi-station deep-learning methods,high-precision focal mechanism inversion,and seismic tomography studies.Additionally,owing to the high seismicity in the southern north-south seismic zone in China,this dataset has great potential for future scientific discoveries.
