Crustal Lg-wave Attenuation Model And Its Application In Sichuan-yunnan Region Of China

The attenuation of seismic wave can be described by quality factor Q.The high-frequency regional Lg-waves were used in this paper to obtain the Q tomography of Sichuan-Yunnan region and its neighborhood.The Lg-wave amplitude attenuation is very sensitive to physical properties in the crust including the temperature,partial melting and pore fluid.Thus,the₂)can provide new constraint for the lateral difference of crustal tectonic interpretation.The Sichuan-Yunnan region and its surrounding areas are mainly composed of Songpan-Ganzi Belt,Sichuan Basin,Middle-Yunnan block,Indo-China block and Yangtze craton,in which the escape patterns of the Tibetan Plateau materials have been under debate.We can obtain the lateral crustal structure of attenuation of the Sichuan-Yunnan region and its neighbor area and further discuss the dynamic mechanism of the escape patterns of Tibetan plateau matter through the Lg-wave attenuation.Therefore,we have solved two fundamental issues that pertain to Q tomography and carried out a discuss about the tectonic implication of Q model in Sichuan-Yunnan region by constructing a high-resolution broadband Lg-wave Q model in Sichuan-Yunnan region and its surrounding areas.The related research findings are Listed as follows.The data input to Lg Q tomographic inversion are the Lg amplitudes that are corrected by1D geometrical spreading and taken logarithm operation,which is called“reduced amplitudes”,the distribution of the errors in the reduced amplitudes are of importance whether the inversion can be solved as a least-squared problem.Comprehensive statistics were performed for investigating error characteristics in Lg-wave Q tomography based on vertical component waveforms recorded by 262 seismic stations in and around the Sichuan-Yunnan region from 133 earthquakes with magnitudes between 3.5 and 6.5 from July 2007 to May 2013.Under the first-order approximation,we use a random variableto represent the modeling error,so that we turn the difficult mission of the statistical analysis of the error in reduced amplitude into a simple work of analyzing the statistics of.The various statistical test results show that the modeling errors obeyed a normal distribution under the first-order approximation,but with long-tailed outliers.By using the data screening technique,we generated a new dataset with nearly perfect normally distributed for Q tomography which can be solved as a least-squared problem.Unlike those various iterative methods,we used the Q tomographic method based on singular value decomposition(SVD).The iterative methods start with an initial Q model and interactively improve the model for better data fit,but they cannot quantitatively estimate the resolution and error in Q model.The Q tomographic method developed via PROPACK package can overcome the shortcomings that SVD cannot apply for the data-intensive modeling.Thus,we can perform singular value decomposition on a data kernel matrix,which enables us to solve for the model resolution,covariance matrices and along with the high-resolution broadband Lg-wave Q model at 1.0 Hz,in which areas with dense ray path coverages can be resolved within a range of?100 km with a relative error of less than 3%.From the Q_Lgmodel we obtained,it can be discerned that Songpan-Ganzi Belt,Sanjiang fault area and western Sichuan have a significant high Q feature,correlating with previously observed low velocity,high conductivity and high Poisson's ratio,may indicate possible high temperature and/or partial melting within the crust.Higher-Q phenomenon appear at eastern Sichuan basin and Yangtze craton may indicate the materials of Tibetan plateau was blocked by the Sichuan basin,forcing it to migrate southward.Meanwhile,a higher-Q area appears at Emeishan Large Igneous Province of Middle-Yunnan block with low-Q located in both sides,it may suggest that the crustal materials diverted into eastern and western sides due to resistance from the Emeishan Large Igneous Province.Also,through our results and other geological and geophysical observations,a dynamic model is suggested here by combining rigid block extrusion model,plastic deformation model and the crustal flow model to explain crustal material escape pattern.