基于高频GNSS地震学的震源参数与破裂过程实时反演研究

We propose a method to calculate the real-time displacement using only one single station equipped by the high-rate GNSS, which consists of double positioning systems, one is the US GPS and the other the Chinese Compass systems. The accuracy and reliability of the method will be verified through theoretical experiments. In order to deal with the insufficiency of GNSS sites at locations near strong earthquakes, we intend to carry out real-time data fusion between accelerograms from strong motion instruments and displacement data from GNSS to acquisite surface velocity waveforms including both low and high frequency information. We also put forward a real-time earthquake moment magnitude estimation method based on high-rate GNSS time-series displacement data, to cope with saturation problem that commonly exists in earthquake magnitude estimations using seismograms acquired by seismometers. Magnitude estimation methods are then evaluated to determine its scope of application and threshold of each method can be determined, through comparison between traditional seismology method and the newly-proposed method. We study self-adapting strategy in earthquake rupture process inversion that be obtained in real time based on high-rate GNSS sites and strong motion stations. We acknowledge this project as high-rate real-time GNSS seismology, based on which we can try to obtain the dynamic stress/strain evolution and the early judgment basis for earthquake rupture propagation direction. To justify and improve the method, we take the 2008 Wenchuan earthquake as the example, conducting earthquake rupture process inversion in real time and retrospective manner. In the mean while, we propose to study the role of the Pengguan fault and the Xiaoyudong fault during the Wenchuan rupture process, through elaborate analysis the relationship among station residual variations, station observations, fault model and slip distribution variations, etc. The proposed project is critical not only to better understand the Wenchuan faulting mechanism and rupture mechanics, but also provide technical basis for the national earthquake early warning system, which is under the transforming process from “two systems” fusion (using both velocity and acceleration data ) to “three systems” fusion (using velocity, acceleration and displacement data).

开展基于高频GNSS（GPS/BDS）的实时单站定位方法研究，并验证其精度；开展加速度与GNSS位移实时融合处理研究，获取包含高低频的地表速度数据；针对地震学方法在估计震级时出现饱和现象，提出基于GNSS多种方法相结合的实时震级估算方法，通过与地震学方法比较，给出震级估算方法的适用范围和判断阈值；基于高频GNSS地震学，研究震源破裂过程自适性和指纹匹配实时反演算法，获取断层破裂动态过程，给出地表动态应力应变图像、地震动图像，分析研究断层破裂方向判断依据。在此基础上，以汶川地震为例，开展断层破裂过程回溯性实时反演研究，详细分析汶川地震发生过程中，台站残差变化与台站数据、断层模型、滑动分布变化等因素的关系，判断彭灌断层、小鱼洞断层在映秀-北川断层破裂过程中所起的作用。项目研究成果不仅对认识汶川地震的震源破裂过程有重要意义，也为我国地震预警系统从“二网融合”到“三网融合”发展提供坚实的技术保障。

高频GNSS地震学能够实时解算测站速度/位移、估计强震震级并反演震源破裂过程，推动了强震预警从“速度+加速度”到“速度+加速度+位移”的发展。本项目采用高频GNSS（GPS/BDS）数据，研究了实时单站位移解算方法、GNSS-强震数据融合和震级估算方法以及震源破裂过程反演算法，并以我国及周边数次强震为例开展回溯性研究，探讨了相关算法的精度和适用性。. 研发了“多卫星系统精密单点定位算法”和“GPS/BDS单站测速算法”，实现了对单测站GNSS数据的实时高精度解算，测速精度达mm/s、位移精度达cm级。发展了一套稳健的强震动基线校正方法，能够实现对加速度数据基线漂移的自动、快速校正；提出了一种基于自适应卡尔曼滤波的GNSS与强震仪数据融合算法，融合后的强震动积分位移与GNSS位移互差在cm级，且融合后的数据既解决了GNSS观测采样率偏低、短期精度相对不足的问题，也保存了强震仪的高频信息、消除了强震数据积分过程中基线漂移问题；基于实时GNSS速度和位移记录，开发了PGD/PGV震级快速估计、震中位置快速确定以及矩心矩张量反演等算法；开发了“基于Okada模型的强震破裂过程自适应算法”，实现了快速断层动态破裂反演中断层破裂尺度的动态调整；研发了“基于三角元模型的指纹匹配算法”，建立了“鲜水河-安宁河-则木河”断层破裂模型指纹库。以上述相关算法为支撑，利用高频GNSS和强震动数据开展了2008年汶川地震震源参数回溯性实时反演，结果表明实时地表位移数据可以快速准确获取强震震级和破裂优势方向，高频GNSS能够对现有地震预警系统提供很好的补充。此外，相关算法在日本3-11地震、尼泊尔地震、玛多地震及2200个合成地震中均得到了充分的验证。在算法研发的基础上，开发了“基于GNSS的地震预警系统平台”，该原型系统对大规模低成本地震预警系统搭建的技术体系进行了验证，为后续的工程建设提供了技术参考。