欢迎访问《地震》,

地震 ›› 2018, Vol. 38 ›› Issue (4): 1-13.

• •    下一篇

基于GPS观测的2015年尼泊尔MS8.1地震多尺度震后形变特征

苏小宁1,2, 孟国杰2   

  1. 1.同济大学测绘与地理信息学院, 上海 200092;
    2.中国地震局地震预测研究所地震预测重点实验室, 北京 100036
  • 收稿日期:2017-10-23 发布日期:2019-08-15
  • 通讯作者: 孟国杰,研究员。E-mail:guojiemeng@126.com
  • 作者简介:苏小宁(1986-),男,甘肃白银人,助理研究员,从事GPS数据处理与地壳变形
  • 基金资助:
    国家自然科学基金(41604007); 地震预测研究所基本科研业务费专项(2016IES010203; 2017IES0401;
    2018IEF010102)

Multiscale Postseismic Deformation based on cGPS Time Series Following the 2015 Gorkha Nepal MS8.1 Earthquake

SU Xiao-ning1,2, MENG Guo-jie2   

  1. 1.College of Surveying and Geo-informatics, Tongji University, Shanghai 200092, China;
    2.Institute of Earthquake Forecasting Key Laboratory of Earthquake Forecast, CEA, Beijing 100036, China
  • Received:2017-10-23 Published:2019-08-15

摘要: 2015年4月25日青藏高原南缘的喜马拉雅地震带发生尼泊尔MS8.1大地震, 中国藏南区域和尼泊尔境内的GPS连续站均观测到了明显的同震和震后形变。 收集中国大陆构造环境监测网络藏南区域GPS基准站和尼泊尔境内的GPS连续观测站数据, 采用统一策略精密解算后获得了测站坐标时序。 利用震前两年以上的数据拟合测站长期构造运动和周期性非构造运动, 并从震后的时间序列扣除之后获得了GPS测站的震后形变。 采用震后余滑的对数型模型、 粘弹性松弛的指数型模型以及二者组合的综合模型等3种震后形变模型拟合了GPS测站观测到的震后形变。 采用基于贝叶斯估计的马尔科夫链蒙特卡洛(Markov Chain Monte Carlo, MCMC)算法反演了模型参数。 结果显示, 对数型模型加指数型模型的综合模型能够更好地拟合GPS观测到的震后形变, 表明GPS观测到的震后形变至少包含两种不同的变形机制, 具有多尺度特征。 指数型模型的特征时间为876.623年, 对应的等效粘滞系数为8.30×1020 Pa·s。 对比不同测站两年内所积累的震后形变与同震比值, 相比震中附近的测站, 中国藏南区域GPS测站观测到的震后形变所占比例更高, 可能指示喜马拉雅地震带主要断裂的持续性加载, 其地震危险性值得进一步关注。

关键词: 尼泊尔地震, GPS, 贝叶斯估计, 对数型模型, 指数型模型, 多尺度特征

Abstract: On April 25th 2015, MS8.1 Nepal earthquake occurred in the Himala earthquake belt that locates in the south margin of the Tibetan Plateau. We analyzed the continuous GPS data from Crustal Movement Observation Network of China and Nepal continuous GPS network. We firstly adopted the multi-scales wavelet method to investigate the pre-seismic effect of tectonic activities on the new set GPS stations, and corrected the coordinate time series of GPS sites. Then,retrieved the postseismic deformations in these time series, and studied the effects of the afterslip based on logarithmic model and visco-ealstic relaxation based on exponential model induced by this event. The parameters of three poseismic deformation models including afterslip, viscoelastic relaxion and the combination of both were inverted by MCMC algorithm of Bayesion frame, showing that observed postseismic deformation involved at least two different mechanisms. We found that the postseismic deformation was featured with variable temporal scales, and controlled by effects of both the afterslip and visco-ealstic relaxation. The charasteristic time of exponential model is 876.623 a, with a viscosity of 8.30×1020 Pa·s. The inverted cumulative deformation shows that this effect is ongoing along the main fault, while the seismic risk in the upper part of unruptured area still deserved further attention.

Key words: 2015 Nepal earthquake, GPS, Bayesian framewor, Logarithmic model, Exponential model, Multi-scale characteristic

中图分类号: