Earthquakes follow a well-known power-law size relation, with smaller events occurring much more often than larger events. Earthquake catalogs are thus dominated by small earthquakes yet are still missing a much larger number of even smaller events because of signal fidelity issues. To overcome these limitations, we applied a template-matching detection technique to the entire waveform archive of the regional seismic network in Southern California. This effort resulted in a catalog with 1.81 million earthquakes, a 10-fold increase, which provides important insights into the geometry of fault zones at depth, foreshock behavior and nucleation processes, and earthquake-triggering mechanisms. The rich detail resolved in this type of catalog will facilitate the next generation of analyses of earthquakes and faults.

The ongoing Sheldon, NV sequence occurred within the remote northwestern corner of the state, where station coverage is sparse and hence the initial catalog locations are poorly constrained. The sequence is of particular scienti c interest due to its persistent, swarm-like seismicity and large moment release (26 M4+ and 261 M3+ events through July 2016, with the largest being M4.8).

This article provides an overview of current applications of machine learning (ML) in seismology. ML techniques are becoming increasingly widespread in seismology, with applications ranging from identifying unseen signals and patterns to extracting features that might improve our physical understanding. The survey of the applications in seismology presented here serves as a catalyst for further use of ML. Five research areas in seismology are surveyed in which ML classification, regression, clustering algorithms show promise: earthquake detection and phase picking, earthquake early warning (EEW), ground-motion prediction, seismic tomography, and earthquake geodesy. We conclude by discussing the need for a hybrid approach combining data-driven ML with traditional physical modeling.

Earthquake source spectra contain fundamental information about the dynamics of earthquake rupture. However, the inherent tradeoffs in separating source and path effects, when combined with limitations in recorded signal bandwidth, make it challenging to obtain reliable source spectral estimates for large earthquake data sets. We present here a stable and statistically robust spectral decomposition method that iteratively partitions the observed waveform spectra into source, receiver, and path terms. Unlike previous methods of its kind, our new approach provides formal uncertainty estimates and does not assume self‐similar scaling in earthquake source properties. Its computational efficiency allows us to examine large data sets (tens of thousands of earthquakes) that would be impractical to analyze using standard empirical Green's function‐based approaches. We apply the spectral decomposition technique to P wave spectra from five areas of active contemporary seismicity in Southern California: the Yuha Desert, the San Jacinto Fault, and the Big Bear, Landers, and Hector Mine regions of the Mojave Desert. We show that the source spectra are generally consistent with an increase in median Brune‐type stress drop with seismic moment but that this observed deviation from self‐similar scaling is both model dependent and varies in strength from region to region. We also present evidence for significant variations in median stress drop and stress drop variability on regional and local length scales. These results both contribute to our current understanding of earthquake source physics and have practical implications for the next generation of ground motion prediction assessments.