Scholarlink

Subduction zones in Earth’s tectonics are potent geological locations for generating devastating earthquakes, including both inter-plate and intra-plate events. The northeastern region of the Indian subcontinent is particularly active in terms of subduction zones, having experienced several major earthquakes in the past like 1930 Dhubri (Mw-7.1), 1950 Assam (Mw-8.4), 1988 Manipur (Mw-7.2), 2009 Assam (Mw-8.4), 2011 Sikkim (Mw-6.9), and 2021 Assam (Mw-6). This study presents a comprehensive ground motion model specifically tailored to the northeastern Indian subduction zone. Addressing the gap of the less comprehensive recorded dataset for the region, the NGA-Subduction dataset is combined with the available recorded data. This dataset is then used to develop the ground motion model for the region by utilizing the artificial neural network and extreme learning machine technique. The formulation chosen for the model comprises the earthquake magnitude, faulting mechanism, source-to-site distance, shear wave velocity, depth, faulting mechanism, and flag for differentiating between both datasets. A detailed residual analysis of the model is also part of this work. Further, the application of the developed model extends to a focused hazard assessment of northeast India. The difference in the hazard estimates for northeastern India using clustered and declustered catalogs is estimated. The novel outcomes of this work offer insights into the seismic behavior of the northeastern Indian subduction zone, providing a robust foundation for enhanced earthquake preparedness and risk mitigation strategies.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine Learning-Based Ground Motion Model for Subduction Zone Earthquakes and Its Application Toward Hazard Estimation of Northeast India

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Romani Choudhary,
  • J. Dhanya

摘要

Subduction zones in Earth’s tectonics are potent geological locations for generating devastating earthquakes, including both inter-plate and intra-plate events. The northeastern region of the Indian subcontinent is particularly active in terms of subduction zones, having experienced several major earthquakes in the past like 1930 Dhubri (Mw-7.1), 1950 Assam (Mw-8.4), 1988 Manipur (Mw-7.2), 2009 Assam (Mw-8.4), 2011 Sikkim (Mw-6.9), and 2021 Assam (Mw-6). This study presents a comprehensive ground motion model specifically tailored to the northeastern Indian subduction zone. Addressing the gap of the less comprehensive recorded dataset for the region, the NGA-Subduction dataset is combined with the available recorded data. This dataset is then used to develop the ground motion model for the region by utilizing the artificial neural network and extreme learning machine technique. The formulation chosen for the model comprises the earthquake magnitude, faulting mechanism, source-to-site distance, shear wave velocity, depth, faulting mechanism, and flag for differentiating between both datasets. A detailed residual analysis of the model is also part of this work. Further, the application of the developed model extends to a focused hazard assessment of northeast India. The difference in the hazard estimates for northeastern India using clustered and declustered catalogs is estimated. The novel outcomes of this work offer insights into the seismic behavior of the northeastern Indian subduction zone, providing a robust foundation for enhanced earthquake preparedness and risk mitigation strategies.