UC Riverside

Tectonic stress is released in the form of aseismic and seismic signatures. Slow and fast/regular earthquakes represent different extremes the seismic spectrum and although we know a lot about the latter the mechanisms of slow earthquakes are still not very well understood. Ever since the discovery of slow earthquakes (Rogers & Dragert, 2003), seismic and geodetic observations and laboratory studies have helped elucidate the nature of these events whose nature varies a lot even amongst themselves. Slow earthquakes include low frequency earthquakes (LFEs), tremors, very low frequency earthquakes (VLFEs), slow slip events (SSE) and episodic tremor and slip (ETS Although their nature differs from location to location, they follow an empirical linear moment duration scaling law. Each of these events has a distinct signal duration and frequency range that makes its detection challenging. These events have often been observed to affect regular seismicity. Studies have shown that slow earthquakes have often precede large regular earthquakes of different sizes and forms. A couple of slow earthquakes led up to the mainshock of the Mw 9.0 Tohoku earthquake 2011 (Kato et al., 2012). A foreshock sequence associated with multiple slow-slip events spatiotemporally preceded the Mw 8.1 Iquique earthquake in 2014 (Kato & Nakagawa, 2014; Ruiz et al., 2014). Such evidence is not limited to only subduction zones. The Mw 7.9 Izmit earthquakes were also preceded by 44 minutes of slow earthquake activity characterized by long-period signals that increased until the mainshock (Bouchon et al., 2011). The most enigmatic example would be the transition of a VLFE into an Mw 3.7 earthquake in a strike-slip setting in Alaska (Tape et al., 2018). The fact that these transition slip behaviors can occur in the same part of the fault interface as regular earthquakes raise questions about the mechanism that transforms one type of slip behavior to another. The 2011 Tohoku earthquake ruptured a part of the fault previously associated with VLFE activity (Ide, Shelly, et al., 2007), raising questions about reassessing our understanding of basic fault mechanics. The research shared herein analyses the physical processes behind spatiotemporal behavior of earthquakes, especially VLFEs, in various tectonic settings. The first project examines the offshore region of Cascadia subduction zone. The study begins in chapter 2 where we discover widespread occurrence of discrete VLFEs offshore Cascadia using ocean-bottom seismometers. Barring occasional regular fast earthquakes, VLFEs are the only seismic stress indicators in offshore CSZ. In the first section of Chapter 2, using centroid moment tensor inversion and matched filtering, we detected sequences of 12 distinct families of VLFEs. The VLFEs north of 43N have a focal mechanism consistent with subduction zone deformation in the area. However, the VLFEs, along 43N-46N, show strike-slip faulting, attributed to sediment consolidation, subduction bending, and transpressional regimes created by complex plate tectonics. It challenges a canonical view of seismogenic zone in Cascadia characterized by a frictionally homogeneous fault segment producing only regular fast earthquakes. The second project involves the discovery of VLFEs and their temporal relationship to nearby regular earthquakes on the Island of Taiwan. This study detects discrete very low frequency earthquakes (VLFEs) using a grid search moment tensor inversion algorithm (Ghosh et al., 2015; Hutchison & Ghosh, 2019). By applying a matched filtered technique, we have created a robust VLFE catalog for three years. The two VLFEs closer to the tremor-producing region show a temporal relationship, but the western VLFE is the most active among the three. Our VLFE catalog of high temporal resolution allows us to identify a significant increase in VLFE activities preceding earthquake swarms. An empirical comparison of the VLFE catalog with regional and local cataloged fast earthquakes reveal two such instances. We show that fluid migration from deeper to shallower crust explains this modulation of regular fast earthquakes by VLFEs. In the following chapter we use Beam Back Projection (BBP) to detect and located tremors in Cholame. Our results show five times more detections than the conventional ECC method. The use of beam back-projection (BBP) has helped us to identify shallow tremors that have never been observed before in this area. We have strong evidence showing persistent shallow tremor activity in the seismogenic zone with our high-resolution seismic array. We create a robust catalog of events for one year and show that the presence of high fluid pressure may be attributed to the crustal heterogeneity that we observe. In the final chapter of this study, we use the experience from the two projects discussed above and create a robust automatic VLFE detection code package. This code can detect, locate, and provide a matched filtered catalog of all events occurring within a user specified region. Earlier all these parts have been used separately and the method was more prone to systematic and human error. We use 2 datasets to verify our observations and finally make the code package publicly available for use. In summary, this research portrays how important slow earthquakes are in the grander scheme of plate tectonics which can improve our understanding of plate interactions.