Real‐Time Earthquake Early Warning With Deep Learning: Application to the 2016 M 6.0 Central Apennines, Italy Earthquake

Comparison of Single-Trace and Multiple-Trace Polarity Determination for Surface Microseismic Data Using Deep Learning

Abstract For surface microseismic monitoring, determination of the P-wave first-motion polarity is important because (1) it has been widely used to determine focal mechanisms and (2) the location accuracy of the diffraction-stack-based method is improved greatly using polarization correction. The convolutional neural network (CNN) is a form of deep learning algorithm that can be applied to predict the polarity of a seismogram automatically. However, the existing network designed for polarity detection utilizes only individual trace information. In this study, we design a multitrace-based CNN (MT-CNN) architecture using several neighbor traces combined as training samples, which could utilize the polarity information of neighbor sensors in the surface microseismic array. We use 17,227 field seismograms with labeled polarities to train two different neural networks that predict the polarities by a single trace or by multiple traces. The performance of the test set and field example of two CNN architectures shows that the MT-CNN significantly produces fewer polarity prediction errors and leads to more accurate focal mechanism solutions for microseismic events.

 全干涉成像的微地震定位方法研究

An effective workflow for updating 1D velocity model and event location in microseismic monitoring

Cross double-difference inversion method for microseismic location and velocity model update

Cross double-difference inversion for simultaneous velocity model update and microseismic event location

Cross double-difference inversion for microseismic event location using data from a single monitoring well

The double-difference (DD) location method has long been applied for locating a cluster of earthquakes with data recorded at surface seismic stations. This method has also been used for locating microseismic events with multiple monitoring wells during hydraulic fracturing. We first extended the approach for locating a cluster of microseismic events with data recorded from a single well. To do this, we have reduced the 3D location problem to 2D by projecting all of the events onto a vertical ([Formula: see text]) plane, with a vertical [Formula: see text]-axis and a horizontal [Formula: see text]-axis representing the distance to the monitoring well, considering the symmetric character of the 1D velocity model and the vertical monitoring well. We then performed a 2D location inversion and projected the results back to 3D using the event azimuths, which were determined from a separate analysis of the initial P-wave polarizations. However, although the DD method could determine relative locations of events reasonably well, it yielded poor absolute locations. We have developed a cross DD (CDD) approach using the cross traveltime difference between the P-wave arrival of one event and the S-wave arrival of another event for inversion instead of the arrival-time differences of the same phases as in the DD method. The CDD method contains more information on absolute locations than the DD method, resulting in a much more stable absolute location determination. The synthetic and field data tests indicated that the CDD method could improve the accuracy of relative and absolute event locations in microseismic clusters.

Cross double-difference inversion method for microseismic location

Double difference method for locating microseismic events from a single well