Study Of Noise Signal Clustering And Seismicity At The Weifang Segment Of Tanlu Fault Zone

As the most prominent fault system in eastern China,the Tanlu fault zone has experienced multi-phase tectonic evolution.The Shandong section(about 34°N～37°N)of Tanlu fault zone hosted several historical destructive earthquakes,such as the M7+Anqiu Earthquake and M8+ Tancheng Earthquake that occurred at the northern and southern parts of the section in 70 BC and 1668,respectively.Locations of the presentday seismicity correlate closely with the surface rupture zone of the M8+ Tancheng Earthquake(hereafter referred to as the Tancheng segment),whereas very few earthquakes occurred on the fault segment of the M7+ Anqiu Earthquake(Weifang segment),resulting in a seismic gap.Many literatures have proposed that the Weifang seismic gap is facing a high risk for a potential destructive earthquake.However,it is unknown whether such a gap is simply due to lack of detectability of the existing seismic network.The Weifang gap still lacks of investigation for noise signal and microearthquake activity,as well as fine imaging of fault zone structure.Since the Tanlu fault zone runs through several developed and populated regions of eastern China,a destructive earthquake would cause signficant damages to the Chinese society and economy.Therefore,it is necessary to carry out investigation on the nature of the Weifang seismic gap.Ambient noise is the pervasive component of continuous seismic records.Noise signals are generally believed to be generated from non-tectonic activities,such as ocean waves and human activities.Understanding the signal characteristics and origins of noises is of great importance for modern seismic data analysis,such as microearthquake detection,empirical Green function calculation,and model training of machine learning.We make use of an unsupervised machine learning technique for noise clustering using the data from a dense array of short-period seismographs(Sarray)deployed within and around the Weifang seismic gap.The noise field of the Sarray comprises 6 types of signals related to various non-tectonic sources.The noise of Label 0 presents a lower main frequency with a higer amplitude in general;Label-1 type noise shows a lower amplitude,but with a broader frequency band;Label-2 noise has a dominant frequency of ～11 Hz;Noise of Label 3 presents two amplitude peaks in the frequency range of 2～20 Hz;Noise of Label 4 possesses a broader band of frequency and a higher amplitude;Label-5 noise presents a clear peak at ～50 Hz.After comprehensively comparing recording environments of different types of noises,we speculated possible physical sources of the6 types of noise signals,which range from road and rail traffic,wind,and nearby power lines.Quantitatively analyzing the spatiotemporal evolution of anthropogenic noises,we found a close correlation between the distribution of anthropogenic noises and urbanization.A well determined 3-D map of seismicity is useful in outlining fault zone structure and quantifying seismic risk of seismic gap.After processing 1-year seismic records of a broadband array of 31 stations(Barray)using advanced methods including the REST,Match&Locate and Grow Clust,we obtained a catalog of local earthquakes within the Weifang seismic gap between August of 2017 and July of 2018.The well-located seismicity spreads across several branches of the Tanlu fault zone,suggesting that stress release is more diffused at this segment.The b value of the detected seismicity in the Weifang segment is 1.02±0.02,slightly lower than that of the Tancheng segment,which might indicate that two segments have different levels of stress.The total annual release of seismic energy along the Tancheng segment is ～44 times larger than that released along the Weifang segment.This is consistent with GPS observation that the Weifang segment has a higher slip deficit.The low amount of energy released by small earthquakes along the Weifang segment might suggest that it has a high risk for a potential destructive earthquake in the future.The well-located 3-D seismicity also outlines several branches of the Tanlu fault zone,such as the northwest dipping Tangwu-Gegou fault and AnqiuJuxian fault in the western part of the fault zone.In the eastern part,we found at least three branches with steeper dipping angles that dip to the southeast in the south and switch to the northwest in the north.We also conducted a time delay analysis recorded by the dense Sarray to quantify lateral variations of P-and S-wave velocities of the crust across the array.The analysis reveals low velocity zones and large velocity contrasts across branches of the Tanlu fault zone.The analysis also shows that the well-located earthquakes occurred mainly in the transition zones between high-and low-velocity structures,which are likely the locations of fault branches developed from stress concentration.