Data Science
Modern trends in geophysical and seismic data analysis involve plowing through very large amounts of data to extract maximum information on Earth structures and processes. Our group’s work has always been geared toward improving existing tools and developing new ones to maximize information extraction from large data sets. Additional examples of recent work include: 1) Using unconventional sources of seismic energy (i.e., anthropogenic noise) to obtain accurate and reliable measurements of spatio-temporal seismic velocity models in underground mines; 2) Harnessing modern statistical learning algorithms to characterize the full seismic wave field, classify earthquakes and subduction zone behavior, and perform improved automatic earthquake detections and locations; and 3) Implementing Bayesian algorithms to obtain reliable uncertainty estimate in geophysical inversions.
Pascal Audet
Professor of Geophysics & University Research Chair
Publications
Blockly earthquake transformer: A deep learning platform for custom phase picking
Deep-learning (DL) algorithms are increasingly used for routine seismic data processing tasks, including seismic event detection and …
Capturing seismic velocity changes in receiver functions with optimal transport
Temporal changes in seismic velocities are an important tool for tracking structural changes within the crust during transient …
Probabilistic inversion of circular phase spectra: Application to two-station phase-velocity dispersion estimation in western Canada
Periodic directional and temporal measurements are common in seismology, and necessitate specific statistical analyses that are …
QuakeLabeler: a fast seismic data set creation and annotation toolbox for AI applications
The production and preparation of data sets are essential steps in machine learning (ML) applications. With the increasing volume and …
Shear‐wave velocity structure of sediments on Cascadia's continental margin from probabilistic inversion of seafloor compliance data
Several seismic techniques, both passive and active, exist for estimating the shear-wave velocity Vs structure of shallow sedimentary …
Probabilistic inversion of seafloor compliance for oceanic crustal shear velocity structure using mixture density neural networks
Measurements of various physical properties of oceanic sediment and crustal structures provide insight into a number of geological and …
Azimuthal anisotropy in Bayesian surface-wave tomography: application to northern Cascadia and Haida Gwaii
Surface wave tomography is a valuable tool for constraining azimuthal anisotropy at regional scales. However, sparse and uneven …
Automatic detection and location of seismic events from time-delay projection mapping and neural network classification
The past several decades have seen an exponential increase in the volume of available seismic data, and with it has come the need to …
Uncovering the physical controls of deep subduction zone slow slip using supervised classification of subducting plate features
Deep slow slip events (SSEs) at subduction zones have significantly contributed to refining our understanding of the megathrust …
Uncovering the physical controls of deep subduction zone slow slip using supervised classification of subducting plate features
Deep slow slip events (SSEs) at subduction zones have significantly contributed to refining our understanding of the megathrust …
Curie depth estimation from magnetic anomaly data: a re-assessment using multitaper spectral analysis and Bayesian inference
The maximum depth of magnetization in the Earth’s crust is generally thought to coincide with the Curie temperature of magnetite (580 …
Recovery of P Waves from Ambient‐Noise Interferometry of Borehole Seismic Data around the San Andreas Fault in Central California
Studies on the recovery and processing of surface waves from ambient‐noise interferometry are now standard practice in seismology. It …
Seismic Interferometry Using Persistent Noise Sources for Temporal Subsurface Monitoring
In passive source seismology, seismic interferometry typically refers to the cross correlation of ambient noise to construct an …
Interferometric methods for spatio temporal seismic monitoring in underground mines
In active underground mining environments, monitoring of the rockmass has important implications for both safety and productivity. …
Supervised machine learning on a network scale: application to seismic event classification and detection
A new method using a machine learning technique is applied to event classification and detection at seismic networks. This method is …
Directional wavelet analysis on the sphere: Application to gravity and topography of the terrestrial planets
The spectral relations (admittance and correlation) between gravity and topography are often used to obtain information on the density …