The Structure Of The Crust Beneath North China Craton And Xingmeng Orogenic Belt

The North China Craton(NCC)is characterized by by frequent magmatic and volcanic activities,forming various faults and folds.There are many contradictions in the crust mantle structure of the North China Craton.There are still many controversies and doubts in the study of NCC.What is the mechanism of lithospheric crust destruction? Most earthquakes in North China occur in the middle and upper crust,so what are the characteristics of the velocity structure of the middle and upper crust?These problems need further study.The tectonic evolution of the Xing'an Mongolian orogenic belt(XMOB)has experienced many stages,such as subduction and closure of the paleo Asian Ocean,crustal accretion,continental collision,post orogenic collapse and extension,and then superimposed the Mongolia-Okhotsk tectonic domain and the paleo Pacific tectonic domain.The discussion of the tectonic evolution process of the area and the closure position of the paleo Asian Ocean is conducive to a profound understanding of the tectonic evolution history of the XMOB.The crust records the information of continental tectonic evolution.Through the exploration of the crustal structure,we can increase the understanding of the earth's interior,infer the informations of plate tectonics evolution,and analyze the seismogenic environment.Receiver function and ambient noise methods are both effective means to explore underground structures.With the development of machine learning related methods,machine learning is more widely used in the field of seismology,showing great application potential.This paper mainly includes three aspects:(1)We collected waveforms from 188 broadband seismic stations deployed in the NCC.Based on these data,we obtained the sediment thickness,crustal thickness,and S-wave velocity structure of the crust through the joint inversion of ambient noise surface wave and receiver function.Based on the geological and geophysical studying of the North China Craton,the deep dynamic model of the study area is discussed.(2)We collected waveforms from 44 broadband stations located in XMOB to calculate receiver functions.The crustal thickness and Poisson's ratio were obtained by using H-? stacking.Combined with previous studies,this paper speculates on the hot research issues in the Xing'an Mongolian orogenic belt,that is,the closed position of the Paleo-Asian Ocean and the triggering mechanism of its extension.(3)We present a new method for derivation of crustal thickness from receiver function data using Extreme Gradient Boosting(XGBoost)algorithm.We compare it with traditional methods and discuss the advantages and disadvantages of machine learning application.The results demonstrate:In the NCC,the sediment thickness corresponds to the geological structure.The crustal thickness changes from 29 km to 46 km,thinning sharply when moving northwest to southeast.In the middle and upper crust,the North China Basin shows slight high-speed anomalies,which may be related to the relatively high rock strength caused by multiple subsidence since the Cenozoic.In the lower crust,the low velocity anomalies are found around the Tangshan-Tianjin,Zhangbei and Taihang Mountain orogenic belt.A significantly high-velocity anomaly is observed beneath the NCB,and we speculate that it may be caused by the underplating and crystallization of the basic mafic magma from the mantle,accompanied by large-scale extension after the lower crust detachment.Located in several regions with strong earthquakes,there is a certain range of(?10 km)high-speed bodies under the sedimentary layer,where the high-speed bodies are wrapped by the low S-wave velocity.In the XMOB,the crustal thickness of the study area ranges from 38.7 to 42.7 km,with an average thickness of 41.2 km.There is a great difference in crustal thickness on both sides of Solonker suture zone.The characteristics of crustal thickness support the geodynamic model that the Paleo-Asian Ocean subducted and closed at the Solonker suture zone.The Poisson's ratios in the study area are low,ranging from 0.215 to 0.277,with an average value of 0.243,suggesting that the rock composition of the area is dominated by felsic-acid rocks.There exists a negative correlation between the Poisson's ratio and the crustal thickness.Combined with the lower values of Poisson's ratio,we speculate that crustal compression thickening mainly occurs in the felsic upper crust.We show how such a XGBoost model can be trained on a set of receiver function samples and how the uncertainty can be calculated.The technique is applied for NCC to derive crustal thickness,and we compare it with the result of conventional H-?stacking,CCP stacking and inversion.The crustal thicknesses obtained by XGBoost model are highly consistent with the conventional methods.Our model can be used to estimate the receiver function that is influenced by the sedimentary layer and does not depend on the distribution of stations.The effectiveness of XGBoost model make us to quickly and accurately estimate the crustal thickness from large datasets of receiver function.