Machine Learning Models for Classifying Variations in Emergent and Impulsive Seismic Noise

Date: 4/24/2019

Time: 06:00 PM

Room: Grand Ballroom

The proper classification of emergent and impulsive noise signals is critical for reducing false detections of microearthquakes and understanding ongoing ground motions. Continuous seismic waveforms contain numerous natural and anthropogenic signals whereas tectonic seismic events occupy only a small percentage of each day. A dense array of 1,100 vertical geophones recorded ground motions at the San Jacinto fault zone for 30 days in 2014 that provides detailed test data to detect microearthquakes and observe surface/atmospheric processes that manifest as impulsive and emergent seismic signals. Recent studies utilizing the spatially dense seismic array have demonstrated that ongoing low-amplitude seismic motion is dominated by various weak sources originating at the surface from anthropogenic and atmosphere interaction. Labeling classes of waveforms from wind generated ground motions, air-traffic, automobiles, and other non-tectonic signals can provide insightful information for designing a training data set. We apply a new methodology that uses subtle changes in correlations to label continuous waveforms as random noise, non-random noise, or a mixture of signals, and focus our efforts on identifying different classes of non-tectonic signals in the non-random noise to build a machine learning training data set. The data allow us to produce millions of 1 second labeled waveforms and present results showing the variability in the noise signals. The noise signals are used to train a convolutional neural network (ConvNet) to classify continuous waveforms. The ConvNet contains 4 convolution layers and 2 fully connected layer using rectified linear unit activation functions on each layer and a softmax activation function for the output layer. Results of coherent non-tectonic signals across the array provide insight on shallow crustal deformation and surface generated ground motions.

Presenting Author: Christopher Johnson

Authors