Ke Jia
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Ke Jia has a total of 46 co-authors. Here are the top co-authors by number of shared publications:
- Shiyong Zhou
School of Earth and Space Sciences Peking University Beijing China
7 shared publications - Jiancang Zhuang
The Institute of Statistical Mathematics, Research Organization of Information and Systems, Tokyo, Japan
4 shared publications - Jun Hu
State Key Laboratory of Geohazard Prevention and Geoenvironment Protection, Chengdu University of Technology , Chengdu 610059
4 shared publications Changsheng Jiang
Institute of Geophysics China Earthquake Administration Beijing China
3 shared publications- Changsheng Jiang
Institute of Geophysics, China Earthquake Administration, Beijing 100081, China
3 shared publications - Kai Deng
College of Geophysics, Chengdu University of Technology , Chengdu 610059
3 shared publications
Latest Publications
- Open Access
- 9 May 2026
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 246 (1)
One-step prediction random forest for induced seismic hazard forecasting: application to the Luxian area, Southern Sichuan Basin, China
SUMMARY: Hydraulic fracturing in unconventional gas development has intensified concerns over induced seismicity, generating significant seismic hazards with potential risk implications for surrounding environments and communities. Accurate prediction and transparent interpretation of such hazards remain open challenges in seismology and engineering practice. This study addresses these challenges by developing a One-Step Prediction random forest framework to model the spatiotemporal relationships among hydraulic fracturing well deployment, geological factors, historical seismicity and the likelihood of future seismic occurrences. A seismic energy labelling scheme based on one-step prediction enables the framework to estimate potential seismic energy release and identify dominant controlling factors through feature attribution. Building on these results, the conventional traffic-light risk management system is conceptually extended to a traffic-light hazard management scheme, which incorporates theoretical insights from risk analysis to improve interpretability and operational relevance. Predictive performance across low-, medium- and high-hazard scenarios is assessed using confusion matrix analysis, while Shapley Additive exPlanations-based interpretability confirms that the framework preserves physical consistency by linking geological and operational drivers with seismic energy release. The findings advance methodological innovation in induced seismicity research by combining hazard prediction with a hazard-management framework inspired by risk theory, providing both theoretical insights and practical tools for shale gas development.
- 9 May 2026
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 246 (1)
Open AccessOne-step prediction random forest for induced seismic hazard forecasting: application to the Luxian area, Southern Sichuan Basin, China
SUMMARY: Hydraulic fracturing in unconventional gas development has intensified concerns over induced seismicity, generating significant seismic hazards with potential risk implications for surrounding environments and communities. Accurate prediction and transparent interpretation of such hazards remain open challenges in seismology and engineering practice. This study addresses these challenges by developing a One-Step Prediction random forest framework to model the spatiotemporal relationships among hydraulic fracturing well deployment, geological factors, historical seismicity and the likelihood of future seismic occurrences. A seismic energy labelling scheme based on one-step prediction enables the framework to estimate potential seismic energy release and identify dominant controlling factors through feature attribution. Building on these results, the conventional traffic-light risk management system is conceptually extended to a traffic-light hazard management scheme, which incorporates theoretical insights from risk analysis to improve interpretability and operational relevance. Predictive performance across low-, medium- and high-hazard scenarios is assessed using confusion matrix analysis, while Shapley Additive exPlanations-based interpretability confirms that the framework preserves physical consistency by linking geological and operational drivers with seismic energy release. The findings advance methodological innovation in induced seismicity research by combining hazard prediction with a hazard-management framework inspired by risk theory, providing both theoretical insights and practical tools for shale gas development.
- Open Access
- 3 April 2026
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 245 (3)
CapsNet-enhanced seismic event classification: benchmarking and system deployment on the DiTing 2.0 data set
SUMMARY: Automated seismic event classification is a critical component of modern earthquake monitoring anda network operations, yet many previous studies have been limited by small data sets or binary tasks. This work presents the first systematic benchmark of mainstream deep learning models, including convolutional neural networks, Transformer-based architectures and Capsule Networks (CapsNet), on the large-scale DiTing 2.0 AI seismic data set, which contains 19 384 labelled three-component waveforms spanning three classes (natural earthquakes, quarry blasts and mine collapses). Thus, we provide reproducible baselines for the seismic events classifications. Several CapsNet variants optimized for DiTing 2.0 are evaluated, and our results show that the CapsNet+Res model using MFCC representations and data augmentation achieves 91.08 per cent accuracy (weighted F1 = 91.10 per cent) on the held-out test set. The multistation voting further improves event-level accuracy to 97.52 per cent, while a companion noise–event classifier attains 98.47 per cent accuracy. Functionality testing on demonstration continuous records confirms reliable end-to-end operation within a transferable and user-friendly system, underscoring its feasibility for seismic network applications. Overall, this study bridges methodological development and operational application by providing robust baselines, insights into the interplay of input features and architectures, and a practical platform for automated classification; large-scale continuous-data evaluation, cross-regional transfer validation and adaptive learning strategies remain important directions for future research.
- 3 April 2026
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 245 (3)
Open AccessCapsNet-enhanced seismic event classification: benchmarking and system deployment on the DiTing 2.0 data set
SUMMARY: Automated seismic event classification is a critical component of modern earthquake monitoring anda network operations, yet many previous studies have been limited by small data sets or binary tasks. This work presents the first systematic benchmark of mainstream deep learning models, including convolutional neural networks, Transformer-based architectures and Capsule Networks (CapsNet), on the large-scale DiTing 2.0 AI seismic data set, which contains 19 384 labelled three-component waveforms spanning three classes (natural earthquakes, quarry blasts and mine collapses). Thus, we provide reproducible baselines for the seismic events classifications. Several CapsNet variants optimized for DiTing 2.0 are evaluated, and our results show that the CapsNet+Res model using MFCC representations and data augmentation achieves 91.08 per cent accuracy (weighted F1 = 91.10 per cent) on the held-out test set. The multistation voting further improves event-level accuracy to 97.52 per cent, while a companion noise–event classifier attains 98.47 per cent accuracy. Functionality testing on demonstration continuous records confirms reliable end-to-end operation within a transferable and user-friendly system, underscoring its feasibility for seismic network applications. Overall, this study bridges methodological development and operational application by providing robust baselines, insights into the interplay of input features and architectures, and a practical platform for automated classification; large-scale continuous-data evaluation, cross-regional transfer validation and adaptive learning strategies remain important directions for future research.
- Open Access
- 23 March 2026
- Published by Elsevier in SSRN Electronic Journal
Progressive Activation of a Complex Fault System by Hydraulic Fracturing: Insights from High-Resolution Microseismic Monitoring
Hydraulic fracturing operations in unconventional shale reservoirs can induce seismicity through complex physical processes. This study investigates fault activation during a 19-stage hydraulic fracturing operation in the Weiyuan shale gas field, Sichuan Basin, China, using a high-resolution microseismic catalog derived from an advanced detection and location workflow (PALM). Our integrated analysis, incorporating precise hypocenter relocation, focal mechanism solution, stress field inversion, and waveform cross-correlation, reveals the activation of a multi-fault system comprising both thrust and strike-slip structures. We observe a transition in the dominant triggering mechanism: early seismicity was primarily driven by pore-pressure diffusion on proximal faults, while later stages were characterized by remote fault activation via poroelastic stress transfer. A key finding is the systematic rotation of the principal stress axes near the injection zone, providing direct evidence of injection-induced stress perturbations. These findings demonstrate that pre-existing, critically stressed faults can be activated through a combination of mechanisms, with poroelastic effects capable of triggering seismicity beyond the immediate pore pressure front. Our work provides a framework for interpreting induced seismicity in complex geological settings and highlights stress field rotation as a diagnostic indicator, offering critical insights for seismic hazard assessment and mitigation.
- 23 March 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessProgressive Activation of a Complex Fault System by Hydraulic Fracturing: Insights from High-Resolution Microseismic Monitoring
Hydraulic fracturing operations in unconventional shale reservoirs can induce seismicity through complex physical processes. This study investigates fault activation during a 19-stage hydraulic fracturing operation in the Weiyuan shale gas field, Sichuan Basin, China, using a high-resolution microseismic catalog derived from an advanced detection and location workflow (PALM). Our integrated analysis, incorporating precise hypocenter relocation, focal mechanism solution, stress field inversion, and waveform cross-correlation, reveals the activation of a multi-fault system comprising both thrust and strike-slip structures. We observe a transition in the dominant triggering mechanism: early seismicity was primarily driven by pore-pressure diffusion on proximal faults, while later stages were characterized by remote fault activation via poroelastic stress transfer. A key finding is the systematic rotation of the principal stress axes near the injection zone, providing direct evidence of injection-induced stress perturbations. These findings demonstrate that pre-existing, critically stressed faults can be activated through a combination of mechanisms, with poroelastic effects capable of triggering seismicity beyond the immediate pore pressure front. Our work provides a framework for interpreting induced seismicity in complex geological settings and highlights stress field rotation as a diagnostic indicator, offering critical insights for seismic hazard assessment and mitigation.
- 11 January 2025
- Published by Society of Exploration Geophysicists in Geophysics
- Vol. 90 (6) , MR441-MR447
Quantifying the role of fluid injection in induced seismicity at Changning salt mine
Understanding the causal mechanisms linking industrial fluid injection to induced seismicity is crucial for seismic hazard assessment. This study investigates the relationship between salt-mining operations and anomalous seismicity in Changning, Sichuan, China. We use two independent approaches: (1) calculating static Coulomb stress changes induced by fluid injection/extraction and (2) estimating stress perturbations via rate-and-state friction law. Key findings reveal that within 2 km of the wells, injection-induced stress perturbations promoted seismicity. Spatiotemporal correlation with positive Coulomb stress changes confirms injection modulated near-field fault stability. Beyond 2 km, minimal injection-induced stresses preclude direct triggering. Instead, external large events and cascading stress transfer from near-field earthquakes may dominate the mid- to far-field seismicity. Critically, cumulative seismic moment release exceeded the theoretical bound, demonstrating that tectonic processes supplied considerable seismic energy. The dual-path stress analysis quantifies injection’s fractional role and effectively evaluates the impact of fluid injection in complex geologic settings.
- 11 January 2025
- Published by Society of Exploration Geophysicists in Geophysics
- Vol. 90 (6) , MR441-MR447
Quantifying the role of fluid injection in induced seismicity at Changning salt mine
Understanding the causal mechanisms linking industrial fluid injection to induced seismicity is crucial for seismic hazard assessment. This study investigates the relationship between salt-mining operations and anomalous seismicity in Changning, Sichuan, China. We use two independent approaches: (1) calculating static Coulomb stress changes induced by fluid injection/extraction and (2) estimating stress perturbations via rate-and-state friction law. Key findings reveal that within 2 km of the wells, injection-induced stress perturbations promoted seismicity. Spatiotemporal correlation with positive Coulomb stress changes confirms injection modulated near-field fault stability. Beyond 2 km, minimal injection-induced stresses preclude direct triggering. Instead, external large events and cascading stress transfer from near-field earthquakes may dominate the mid- to far-field seismicity. Critically, cumulative seismic moment release exceeded the theoretical bound, demonstrating that tectonic processes supplied considerable seismic energy. The dual-path stress analysis quantifies injection’s fractional role and effectively evaluates the impact of fluid injection in complex geologic settings.
- Open Access
- 4 September 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (17) , 7857
Machine Learning Applications in Seismology
The comprehension of earthquakes and natural hazards, including volcanic eruptions and landslides, as well as explosions, through observational data is a pivotal activity within the field of seismology
- 4 September 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (17) , 7857
Open AccessMachine Learning Applications in Seismology
The comprehension of earthquakes and natural hazards, including volcanic eruptions and landslides, as well as explosions, through observational data is a pivotal activity within the field of seismology
- Open Access
- 24 May 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (11) , 6424
Small Earthquakes Can Help Predict Large Earthquakes: A Machine Learning Perspective
Earthquake prediction is a long-standing problem in seismology that has garnered attention from the scientific community and the public. Despite ongoing efforts to understand the physical mechanisms of earthquake occurrence, there is no convincing physical or statistical model for predicting large earthquakes. Machine learning methods, such as random forest and long short-term memory (LSTM) neural networks, excel at identifying patterns in large-scale databases and offer a potential means to improve earthquake prediction performance. Differing from physical and statistical approaches to earthquake prediction, we explore whether small earthquakes can be used to predict large earthquakes within the framework of machine learning. Specifically, we attempt to answer two questions for a given region: (1) Is there a likelihood of a large earthquake (e.g., M ≥ 6.0) occurring within the next year? (2) What is the maximum magnitude of an earthquake expected to occur within the next year? Our results show that the random forest method performs best in classifying large earthquake occurrences, while the LSTM method provides a rough estimation of earthquake magnitude. We conclude that small earthquakes contain information relevant to predicting future large earthquakes and that machine learning provides a promising avenue for improving the prediction of earthquake occurrences.
- 24 May 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (11) , 6424
Open AccessSmall Earthquakes Can Help Predict Large Earthquakes: A Machine Learning Perspective
Earthquake prediction is a long-standing problem in seismology that has garnered attention from the scientific community and the public. Despite ongoing efforts to understand the physical mechanisms of earthquake occurrence, there is no convincing physical or statistical model for predicting large earthquakes. Machine learning methods, such as random forest and long short-term memory (LSTM) neural networks, excel at identifying patterns in large-scale databases and offer a potential means to improve earthquake prediction performance. Differing from physical and statistical approaches to earthquake prediction, we explore whether small earthquakes can be used to predict large earthquakes within the framework of machine learning. Specifically, we attempt to answer two questions for a given region: (1) Is there a likelihood of a large earthquake (e.g., M ≥ 6.0) occurring within the next year? (2) What is the maximum magnitude of an earthquake expected to occur within the next year? Our results show that the random forest method performs best in classifying large earthquake occurrences, while the LSTM method provides a rough estimation of earthquake magnitude. We conclude that small earthquakes contain information relevant to predicting future large earthquakes and that machine learning provides a promising avenue for improving the prediction of earthquake occurrences.
- Open Access
- 27 March 2022
- Published by Wiley in ESS Open Archive
Physics-Based Earthquake Simulations of the Anninghe-Zemuhe-Daliangshan-Xiaojiang Fault system in the Southwestern China
Seismic hazard evaluation is important for urban construction, earthquake disaster prevention and mitigation. However, because of the long recurrence times of large earthquakes (hundreds to thousands of years), seismic hazard assessment based on relatively short-term (tens of years) observational seismic catalog is difficult. Synthetic seismic catalogs provide a useful way to obtain long-term seismic hazard. Based on Coulomb failure criterion, we apply a quasi-static physics-based earthquake simulator (Virtual Quake) in a tectonically complicated region with several major faults, namely the Anninghe, Zemuhe, Daliangshan and Xiaojiang faults, in the Southwestern China. Slip rates of those major faults are constrained by the GPS data, and frictions are constrained by the laboratory experiments. Considering the stress interactions among the fault system, a synthetic catalog over 20, 000 years is simulated. The simulated catalog shows consistence with the observed seismicity in spatial distribution of large earthquakes, b value and mean seismic rate. The synthetic catalog also shows that the mean intervals of M≥7.0 earthquakes for the Anninghe, Zemuhe, Daliangshan and Xiaojiang faults are 299a, 867a, 361a and 90a, respectively. Our results provide a helpful index to evaluate seismic hazard in such a complicated region.
- 27 March 2022
- Published by Wiley in ESS Open Archive
Open AccessPhysics-Based Earthquake Simulations of the Anninghe-Zemuhe-Daliangshan-Xiaojiang Fault system in the Southwestern China
Seismic hazard evaluation is important for urban construction, earthquake disaster prevention and mitigation. However, because of the long recurrence times of large earthquakes (hundreds to thousands of years), seismic hazard assessment based on relatively short-term (tens of years) observational seismic catalog is difficult. Synthetic seismic catalogs provide a useful way to obtain long-term seismic hazard. Based on Coulomb failure criterion, we apply a quasi-static physics-based earthquake simulator (Virtual Quake) in a tectonically complicated region with several major faults, namely the Anninghe, Zemuhe, Daliangshan and Xiaojiang faults, in the Southwestern China. Slip rates of those major faults are constrained by the GPS data, and frictions are constrained by the laboratory experiments. Considering the stress interactions among the fault system, a synthetic catalog over 20, 000 years is simulated. The simulated catalog shows consistence with the observed seismicity in spatial distribution of large earthquakes, b value and mean seismic rate. The synthetic catalog also shows that the mean intervals of M≥7.0 earthquakes for the Anninghe, Zemuhe, Daliangshan and Xiaojiang faults are 299a, 867a, 361a and 90a, respectively. Our results provide a helpful index to evaluate seismic hazard in such a complicated region.
- Open Access
- 2 August 2021
- Published by American Geophysical Union (AGU) in Geophysical Research Letters
- Vol. 48 (15)
Stress Transfer Along the Western Boundary of the Bayan Har Block on the Tibet Plateau From the 2008 to 2020 Yutian Earthquake Sequence in China
Eight Ml ≥ 7.0 earthquakes have occurred around the Bayan Har block, NW Tibet, China, since 2000, resulting in a large number of casualties. Near the western boundary of the Bayan Har block, four Ml ≥ 6.0 Yutian earthquakes have occurred from 2008 to 2020. Stress interactions among them are comprehensively investigated by applying the ETAS (Epidemic‐Type Aftershock Sequence) model and calculating(Coulomb failure stress change) from a 3D linear viscoelastic mode. The combined (coseismic plus postseismic)induced by proceeding Yutian earthquakes on hypocenters of the 2012, 2014, and 2020 Yutian earthquakes are −1.5 × 10−4, 3.6 × 10−3, and 1.5 × 10−1 MPa, respectively. The background probabilities of the 2008, 2012, 2014, and 2020 Yutian earthquakes are 0.87, 0.97, 1.5 × 10−3, and 8.7 × 10−5, respectively. Combining those two independent approaches, we conclude that the 2008 and 2012 Yutian earthquakes are more like background earthquakes and that the 2014 and 2020 Yutian earthquakes were triggered by the proceeding Yutian earthquakes.
- 2 August 2021
- Published by American Geophysical Union (AGU) in Geophysical Research Letters
- Vol. 48 (15)
Open AccessStress Transfer Along the Western Boundary of the Bayan Har Block on the Tibet Plateau From the 2008 to 2020 Yutian Earthquake Sequence in China
Eight Ml ≥ 7.0 earthquakes have occurred around the Bayan Har block, NW Tibet, China, since 2000, resulting in a large number of casualties. Near the western boundary of the Bayan Har block, four Ml ≥ 6.0 Yutian earthquakes have occurred from 2008 to 2020. Stress interactions among them are comprehensively investigated by applying the ETAS (Epidemic‐Type Aftershock Sequence) model and calculating(Coulomb failure stress change) from a 3D linear viscoelastic mode. The combined (coseismic plus postseismic)induced by proceeding Yutian earthquakes on hypocenters of the 2012, 2014, and 2020 Yutian earthquakes are −1.5 × 10−4, 3.6 × 10−3, and 1.5 × 10−1 MPa, respectively. The background probabilities of the 2008, 2012, 2014, and 2020 Yutian earthquakes are 0.87, 0.97, 1.5 × 10−3, and 8.7 × 10−5, respectively. Combining those two independent approaches, we conclude that the 2008 and 2012 Yutian earthquakes are more like background earthquakes and that the 2014 and 2020 Yutian earthquakes were triggered by the proceeding Yutian earthquakes.
- Open Access
- 5 May 2021
- Published by Wiley in ESS Open Archive
Stress transfer along the western boundary of the Bayan Har Block on the Tibet Plateau from the 2008 to 2020 Yutian earthquake sequence in China
Eight Ml≥7.0 earthquakes have occurred around the Bayan Har block, NW Tibet, China, since 2000, resulting in a large number of casualties and countless economic loss. Near the western boundary of the Bayan Har block, four Ml≥6.0 Yutian earthquakes have occurred from 2008 to 2020. Stress interactions among them are comprehensively investigated by applying the ETAS (Epidemic-Type Aftershock Sequence) model and calculating (Coulomb failure stress change). The viscoelastic induced by proceeding Yutian earthquakes on hypocenters of the 2012, 2014 and 2020 Yutian earthquakes are -1.5004×10 -4 , 3.5820×10 -3 and 1.4770×10 -1 MPa, respectively. The background probabilities of the 2008, 2012, 2014 and 2020 Yutian earthquakes are 0.87, 0.97, 1.5×10 -3 and 8.7×10 -5 , respectively. Combining those two independent approaches, we conclude that the 2008 and 2012 Yutian earthquakes are more like background earthquakes and that the 2014 and 2020 Yutian earthquakes were triggered by the proceeding Yutian earthquakes.
- 5 May 2021
- Published by Wiley in ESS Open Archive
Open AccessStress transfer along the western boundary of the Bayan Har Block on the Tibet Plateau from the 2008 to 2020 Yutian earthquake sequence in China
Eight Ml≥7.0 earthquakes have occurred around the Bayan Har block, NW Tibet, China, since 2000, resulting in a large number of casualties and countless economic loss. Near the western boundary of the Bayan Har block, four Ml≥6.0 Yutian earthquakes have occurred from 2008 to 2020. Stress interactions among them are comprehensively investigated by applying the ETAS (Epidemic-Type Aftershock Sequence) model and calculating (Coulomb failure stress change). The viscoelastic induced by proceeding Yutian earthquakes on hypocenters of the 2012, 2014 and 2020 Yutian earthquakes are -1.5004×10 -4 , 3.5820×10 -3 and 1.4770×10 -1 MPa, respectively. The background probabilities of the 2008, 2012, 2014 and 2020 Yutian earthquakes are 0.87, 0.97, 1.5×10 -3 and 8.7×10 -5 , respectively. Combining those two independent approaches, we conclude that the 2008 and 2012 Yutian earthquakes are more like background earthquakes and that the 2014 and 2020 Yutian earthquakes were triggered by the proceeding Yutian earthquakes.
- Open Access
- 27 January 2021
- Published by Wiley in ESS Open Archive
Stress transfer in the western boundary of Bayan Har Block in Tibet Plateau from the 2008 to 2020 Yutian Earthquake sequence in China
Eight M≥7.0 earthquakes have occurred around Bayan Har block since 2000, resulting huge casualties and countless economic loss. Specifically, four M≥6.0 earthquakes occurred near the western boundary of Bayan Har block from 2008 to 2020. Stress interactions among them are comprehensively investigated by applying ETAS (epidemic-type aftershock sequence) model and calculating dCFS (Coulomb failure stress change). The viscoelastic dCFS induced by proceeding Yutian earthquakes on hypocenters of 2012, 2014 and 2020 Yutian earthquakes are -1.5004×10 -4 , 3.5820×10 -3 and 1.4770×10 -1 MPa, respectively. The background probabilities of 2008, 2012, 2014 and 2020 Yutian earthquakes are 0.87, 0.97, 1.5×10 -3 and 8.7×10 -5 , respectively. Combining those two kinds of evidence, we conclude that 2008 and 2012 Yutian earthquakes are more like background earthquakes and 2014 and 2020 Yutian earthquakes are triggered by proceeding Yutian earthquakes.
- 27 January 2021
- Published by Wiley in ESS Open Archive
Open AccessStress transfer in the western boundary of Bayan Har Block in Tibet Plateau from the 2008 to 2020 Yutian Earthquake sequence in China
Eight M≥7.0 earthquakes have occurred around Bayan Har block since 2000, resulting huge casualties and countless economic loss. Specifically, four M≥6.0 earthquakes occurred near the western boundary of Bayan Har block from 2008 to 2020. Stress interactions among them are comprehensively investigated by applying ETAS (epidemic-type aftershock sequence) model and calculating dCFS (Coulomb failure stress change). The viscoelastic dCFS induced by proceeding Yutian earthquakes on hypocenters of 2012, 2014 and 2020 Yutian earthquakes are -1.5004×10 -4 , 3.5820×10 -3 and 1.4770×10 -1 MPa, respectively. The background probabilities of 2008, 2012, 2014 and 2020 Yutian earthquakes are 0.87, 0.97, 1.5×10 -3 and 8.7×10 -5 , respectively. Combining those two kinds of evidence, we conclude that 2008 and 2012 Yutian earthquakes are more like background earthquakes and 2014 and 2020 Yutian earthquakes are triggered by proceeding Yutian earthquakes.