My research interest focuses on the overarching goal of understanding the physics of earthquake processes in active fault zones, including but not limited to the study of tectonic tremor (i.e., one type of slow earthquake), the dynamic triggering of tremor and earthquakes, the interaction between slow and typical earthquakes, the source parameters of repeating seismic events, and temporal changes in shallow crust properties.
Observations of Triggered Tremor
Tectonic tremor is one type of seismic event that occurs at the transition of a fault. While tremor can occur spontaneously as ambient" tremor, it can also be activated by distant earthquakes as "triggered" tremor. Investigating into the dynamic triggering of tremor can help seismologists evaluate background ambient tremor activities and more thoroughly understand the faulting process. Publications: For details, see the "Triggered tremor summary map" below. Green marks indicate the study regions in which I have been involved.
Summary of Global Triggered Tremor Sources
Interaction Between Tectonic Tremor and Nearby Earthquakes
Tectonic tremor is a stress-sensitive seismic event that belongs to the slow earthquake family. From observations
of tremor, which provide useful clues for the evaluation of stress variation in nearby earthquakes, we have learned that tremor activity increases in the few weeks before the occurrence of a nearby earthquake (Chao et al., JGR, 2017 ). Other publication: Han et al., GRL, 2014.
A useful approach to detecting seismic events with similar waveforms is the waveform cross-correlation technique. We apply this technique to (1) auto-detect ambient tremor (Chao et al., JGR, 2017), (2) identify repeating earthquake families appearing in the earthquake catalog (Chao et al., SPE, 2016), (3) detect new events with the template waveforms using the matched-filter approach (Tang et al., GRL, 2010; Aguiar et al., GRL, 2017), and (4) measure shear-wave splitting (Chao and Peng, GJI, 2009).
Earthquakes Source Analysis
We have developed a series of algorithms that identify accurate source parameters (e.g., location, moment magnitude, stress drop, fault size, and slip) for micro-seismicity during the hydraulic fracturing operations (Chao et al., SPE, 2016).
Machine Learning for the Study of Geoscience Problems
To tackle geoscience algorithms, we are currently collaborating with computer scientists and developing several machine-learning approaches, including those for (1) applying a decision-tree method to quantifying the important factors of the triggering threshold of triggered seismic events (Tang et al., JGR, in revision), and (2) adapting convolutional neural network for the auto-identification of triggered earthquake (Tang et al., SRL, in revision) and triggered tremor (Chao et al., in prep.).
We have developed a MATLAB package, CrazyTremor, for searching for triggered seismic events (Chao and Yu, SRL, 2018). The software can facilitate the analytical process of identifying triggered seismic events from a sizeable seismic dataset.
Click here to download and learn more about CrazyTremor.
Temporal Variation of Shallow Crust Properties
We have measured the temporal variation of shear-wave splitting in the shallow surface and found an apparent decrease in the velocity of S-waves within the top 200m crust layer following an M6.4 earthquake.
Tremor Sources Characteristics