数据驱动的重力站季节性微震信号模拟研究

Simulation on Seasonal Microseismic Signals at Gravity Station Using Data-Driven Approaches

ES评分 0

DOI 10.14075/j.jgg.2024.12.586
刊名
Journal of Geodesy and Geodynamics
年,卷(期) 2025, 45(11)
作者
作者单位

1 中国地震局地震大地测量重点实验室, 武汉市洪山测路40号, 430071;
2 山东省地震局泰安地震监测中心站, 山东省泰安市罗汉崖2号, 271000

摘要
重力站季节性微震信号受随机性、间歇性、周期性因素的综合影响, 难以进行物理模拟和推估。本文基于乌什连续重力观测站2008-2024年1 Hz重力固体潮数据, 在采用视垂直位移分析方法提取DF频段微震信号的基础上, 提出改进的变分模态分解(improved variational mode decomposition, IVMD)算法与长短期记忆(long short-term memory, LSTM)神经网络算法相耦合的方法, 建立重力季节性微震信号的数据模型。结果表明, 与LSTM模型相比, IVMD-LSTM模型更能反映季节性信号的年变差异, 且利用数据驱动建立的模型数据与GLDAS和ERA5模型雪深数据的相关系数提升至0.5。本文研究结果可为基于重力微震季节性信号的模型建立和地震分析预报应用提供一种新的处理方法。
Abstract
The seasonal microseismic signals at gravity stations are difficult to physically simulate and estimate due to the combined effects of randomness, intermittency, and periodicity. Based on the 1 Hz gravity solid tide data of Wushi continuous gravity observation station from 2008 to 2024, we extract the microseismic signals in DF frequency band using the apparent vertical displacement analysis method. An improved variational mode decomposition(IVMD) coupled with the long short-term memory(LSTM) neural network algorithm is proposed to establish a data model for seasonal gravity microseismic signals. The results show that the IVMD-LSTM model can better reflect the annual variation differences of seasonal signals compared to the LSTM model. Moreover, the correlation coefficient between model data using data-driven methods and the snow depth data from GLDAS and ERA5 models increased to 0.5. This study can provide a new processing method for model establishment using seasonal gravity microseismic signals and their application in earthquake analysis and forecasting.
关键词
gPhone重力仪; 视垂直位移季节性变化; 长短期记忆神经网络; 变分模态分解
KeyWord
gPhone gravimeter; seasonal variation of apparent vertical displacement; long short-term memory(LSTM) neural network; variational mode decomposition(VMD)
基金项目
页码 1127-1133
  • 参考文献
  • 相关文献
  • 引用本文

王诤冉, 韦进, 李惠玲, 杨光亮, 胡敏章. 数据驱动的重力站季节性微震信号模拟研究 [J]. 大地测量与地球动力学. 2025; 45; (11). 1127 - 1133.

  • 文献评论

相关学者

相关机构