AGAbhijit Ghosh

Abhijit Ghosh

*This is an algorithmically generated author record based on publicly available information.

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


Abhijit Ghosh has a total of 83 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Rupture Propagation Direction at the Junction of the Garlock and San Andreas Fault System: A Machine Learning Classification Approach Driven by Earthquake Simulations

    by Shankho Niyogi, Abhijit Ghosh, Evan Marschall, Roby Douilly, David Oglesby

    Anticipating potential earthquake rupture paths in the San Andreas fault system poses considerable challenges due to its complex geometry, stress conditions, and frictional properties. A commonly used approach relies on independently varying parameters through iterations of earthquake simulations to better understand the effects of such parameters on rupture. However, such approaches are computationally expensive. In this study we implement a machine learning algorithm trained on outputs from the Rate‐and‐State earthquake simulator (RSQSim) to understand the importance of various parameters on the rupture path at the branch intersection of San Andreas and Garlock fault. We show how machine learning models employing gradient boosting (xgboost) and Random Forest trained on a data set of ∼9,800 simulated earthquakes can capture the rupture classification as well as extract feature importance of fault parameters. We demonstrate that for certain scenarios the machine learning models have considerable testing accuracy in classifying the rupture path, and augmentation of features in the input training data leads to improvements in the accuracy both in terms of precision and recall. Furthermore, our machine learning models suggest that the pre‐earthquake conditions of the fault on which the rupture nucleated are the dominant parameters which affect rupture path at the branch intersection. This approach stands as a promising tool for rupture propagation and fault parameter studies, particularly for understanding which parameters are crucial for determining branching behavior. Consequently, it can help interpret the behavior of both modeled and real world faults.

  • Top Cited Papers
    Open Access

    AI‐PAL: Self‐Supervised AI Phase Picking via Rule‐Based Algorithm for Generalized Earthquake Detection

    by Yijian Zhou, Hongyang Ding, Abhijit Ghosh, Zengxi Ge

    Delineating fault structures through microseismicity is crucial for earthquake hazard assessment, yet constructing high‐resolution catalogs over extended periods remains challenging. This study introduces AI‐PAL, a novel deep learning‐driven workflow that employs a Self‐Attention RNN (SAR) model trained with detections from PAL, an established rule‐based algorithm (Zhou, Yue, et al., 2021, https://doi.org/10.1785/0220210111), for generalized earthquake detection. PAL utilizes short‐term‐average over long‐term‐average algorithm for event detection, ensuring consistent performance across different datasets. AI‐PAL leverages these rule‐based picks as training labels, enabling self‐supervised learning of the SAR model across arbitrary regions, thereby enhancing PAL's detection capabilities. We applied SAR‐PAL to two distinct regions that are featured by recent large earthquakes: (a) the preseismic period of the Ridgecrest‐Coso region (2008–2019), and (b) the pre‐to‐postseismic period of the East Anatolian Fault Zone (EAFZ, 2020–2023/04). Our results demonstrate that SAR‐PAL offers slightly higher detection completeness than the quake template matching matched filter catalog, while boosts over 100 times faster processing and a superior temporal stability, avoiding detection gaps during background periods. Compared to PhaseNet and GaMMA, two widely recognized phase picker and associator, SAR‐PAL proved more scalable, achieving ∼2.5 times more event detections in the EAFZ case, along with a ∼7 times higher phase association rate. We further experimented training PhaseNet and SAR with PAL detections and routine catalogs, and found that no other combinations matched the detection performance of SAR‐PAL. The enhanced catalogs built by SAR‐PAL reveals geometrical complexities of the Ridgecrest faults and the Erkenek‐Pütürge segment of EAFZ, offering insights into their contrasting roles during the large earthquake.


  • Open Access

    Abundant Quasi‐Repeating Earthquakes Occurring Within Repeater Sequences on the Erkenek‐Pütürge Fault (SE Turkey)

    by Yijian Zhou, Abhijit Ghosh

    Repeaters are known as earthquakes that rupture the same asperity in similar manners. However, the complexity of repeater asperities is poorly investigated. In this study, we developed a new strategy to systematically detect quasi‐repeaters that rupture the same asperities in different ways. It begins by identifying repeaters using waveform cross‐correlation (CC) ≥0.9 and CC‐measured Δ(S‐P) times ≤0.01 s. We then adopt a much lower CC threshold to detect quasi‐repeaters, classifying repeater sequences as simple (lacking quasi‐repeaters) or complex (including quasi‐repeaters). Applying this workflow to the Erkenek‐Pütürge fault in Turkey, we detected 43 repeater sequences during 2020–2023/04. Notably, 30 sequences (∼70%) are complex. The repeater magnitudes show weak correlation with their recurrence interval, with only complex sequences following the 2023 M w 7.8 mainshock an exception. Moreover, ∼74% sequences have the first two events following the mainshocks exhibiting decreasing magnitude. These results suggest that slip rate and asperity complexity may be key factors in such modulation.


  • Open Access

    A New Fault Slip Mode Unveiled in the Regional Dynamic Triggering of the 2023 Turkey Kahramanmaraş Earthquake Sequence

    by Heng Luo, Yijian Zhou, Zeyan Zhao, Mehmet Köküm, Teng Wang, Han Yue, Zexin Wang, Nan Hu, Abhijit Ghosh, Xiaodong Song, et al.

    Faults can slip at vastly different rates, generating both high‐stress‐drop regular earthquakes and low‐stress‐drop slow slip events (SSEs). Here, we document a transitional mode of high‐stress‐drop but “silent” slip with two Mw 5 events that were triggered by the 2023 Turkey Kahramanmaraş earthquake sequence. Specifically, we identify eight fault slip events from radar interferometry, none of which are reported in seismic catalogs. Except for four typical aseismic creep events, two events represent earthquakes whose waveforms were obscured within the coda waves of the mainshocks. Notably, we find two silent events that did not radiate discernible high‐frequency seismic energy nor produce local aftershocks but did exhibit high stress drops of 4–6 MPa, similar to regular earthquakes. These two “silent” events likely represent a new fault slip mode positioned between regular earthquakes and SSEs. The identification of these diverse shallow slip events, which are missed in seismic catalogs, calls for reevaluation of faulting mechanisms, the nature of static/dynamic triggering effects, and seismic hazard assessments.


  • Open Access

    Seismic monitoring of 2020 Baghjan oil-well blowout incident in Assam, India

    by Santanu Baruah, Shankho Niyogi, Abhijit Ghosh, Davide Piccinini, Gilberto Saccorotti, Alan L. Kafka, Danica Roth, Mahendra Kumar Yadava, Manoj K. Phukan, G. Narahari Sastry, et al.

    Characterization of a productive oil/gas well blowout through seismological methods is relatively uncommon. In this paper, we conduct an in-depth seismic evaluation of one of the world’s most significant onshore oil well blowout incidents, which occurred in 2020 at the Baghjan oil field in Assam, northeast India. We show that the blowout and related on-site activities generated distinct signals that can be distinguished by their spectral characteristics, temporal variation in geometric spreading, and sharp attenuation of daytime noise in comparison to the nighttime. A micro-earthquake potentially triggered by the blowout was also detected. Furthermore, we show how seismic data can be used to reasonably estimate blowout gas exit velocity and flame height. Our results demonstrate that a detailed characterization and spatiotemporal variation of blowout activity can be successfully captured through seismic monitoring, opening new opportunities for hazard mitigation and cost-effective disaster management for such catastrophic events.


  • Open Access

    Construction of Long-Term Seismic Catalog with Deep Learning: A Workflow for Localized Self-Attention RNN (LoSAR)

    by Yijian Zhou, Hongyang Ding, Abhijit Ghosh, Zengxi Ge

    Characterizing fault behaviors prior to large earthquakes through long-term seismicity is crucial for seismic hazard assessment, yet constructing high-resolution catalogs over extended periods poses significant challenges. This study introduces LoSAR, a novel deep learning-driven workflow that enhances phase picking by Localizing a Self-Attention Recurrent neural network with local data, addressing the generalization problem common in data-driven approaches. We apply LoSAR to two distinct regions that are both featured by recent large earthquakes: (1) preseismic period of the Ridgecrest-Coso region (2008-2019), and (2) pre-postseismic period of the East Anatolian Fault Zone (EAFZ, 2020-2023/04). Through detailed comparisons, we demonstrate that LoSAR offers slightly higher detection completeness than the QTM matched filter catalog, while boosts an over 100 times faster processing and a superior temporal stability, avoiding low-magnitude gaps during background periods. Against PhaseNet and GaMMA, two established phase picker and associator, LoSAR proves more scalable and generalizable, achieving roughly 2.5 times more event detections in the EAFZ case, along with a ~7 times higher phase association rate. By leveraging the two enhanced catalogs and b-value analysis, we gain insights into the preseismic fault behaviors: (1) The Ridgecrest faults are characterized by sparse and distributed seismicity across a band of ~20 km, revealing multiple orthogonal preexisting faults; coupled with a low b-value that signifies this area as a persistent asperity; (2) The Erkenek-Pütürge segment of EAFZ exhibits complex fault geometry that forms a persistent rupture barrier, which consists of a hidden conjugate fault system that presents as a ~10-km wide fault zone.


  • Open Access

    Neural Shading Fields for Efficient Facial Inverse Rendering

    by G. Rainer, L. Bridgeman, A. Ghosh

    Given a set of unstructured photographs of a subject under unknown lighting, 3D geometry reconstruction is relatively easy, but reflectance estimation remains a challenge. This is because it requires disentangling lighting from reflectance in the ambiguous observations. Solutions exist leveraging statistical, data‐driven priors to output plausible reflectance maps even in the under‐constrained single‐view, unknown lighting setting. We propose a very low‐cost inverse optimization method that does not rely on data‐driven priors, to obtain high‐quality diffuse and specular, albedo and normal maps in the setting of multi‐view unknown lighting. We introduce compact neural networks that learn the shading of a given scene by efficiently finding correlations in the appearance across the face. We jointly optimize the implicit global illumination of the scene in the networks with explicit diffuse and specular reflectance maps that can subsequently be used for physically‐based rendering. We analyze the veracity of results on ground truth data, and demonstrate that our reflectance maps maintain more detail and greater personal identity than state‐of‐the‐art deep learning and differentiable rendering methods.


  • Tremor signals during fluid injection are generated by fault slip

    by Shankho Niyogi, Abhijit Ghosh, Abhash Kumar, Richard W. Hammack

    Seismic tremor signals, also known as long-period, long-duration signals, have been reported in several locations where fluid injection for enhanced oil and gas exploration is taking place. However, the origin of these signals remains poorly constrained. We studied seismic tremor signals in Wellington Field, Kansas, using a seismic array during a carbon dioxide injection program. We show that these signals are generated below the surface during the time of carbon dioxide injection. They have a distinct spectral signature, similar to those observed in glacial and volcanic environments. The tremor sources are located near the injection site and aligned with preexisting faults. Modeling results imply that such tremors are generated by frictional slip on fault. These observations may reveal an important deformation mode, which is useful for studying associated stress, seismicity, and triggering, as well as for tracking fault activities during injection operations of all fluids, including supercritical carbon dioxide.


  • Open Access

    Construction of Long-term Seismic Catalog with Deep Learning and Characterization of Preseismic Fault Behavior in the Ridgecrest-Coso Region (2008-2019)

    by Yijian Zhou, Abhijit Ghosh, Lihua Fang, Han Yue, Shiyong Zhou

    Long-term seismicity is an effective tool to infer fault properties at depth, but the catalog construction is challenging because of the large data volume. We propose a new deep learning-based workflow that follows a “Train-Detect-Pick” procedure, which solves the generalization problem in AI pickers. We apply the new workflow on the preseismic phase (2008-2019) of Ridgecrest-Coso region. Results show that the new workflow realizes efficient and stable detection, and well substitutes matched filter. Our new catalog helps characterize the preseismic fault behavior: (1) the Ridgecrest area has a distributed deformation, and the 2019-ruptured segment has a persistent asperity; (2) the central Garlock fault is unfavorable for rupture propagation, because of its discontinuous geometry and low coupling ratio; (3) the Coso geothermal field generates intense and shallow seismicity, which has a high b-value that does not correlate with seismicity rate and industrial production, thus suggest a low stress level.