1. Incorporating deterministic fault maturity properties as a priori elements for Bayesian type earthquake early warning algorithms and final rupture length prediction
The jury is out as to once an earthquake starts, how quickly we can determine its final size. Different studies, using distinct approaches, looking the moment of large earthquakes have come to different conclusions. Meglar and Hayes, [2019] found that by looking at the moment acceleration (i.e. how rapidly the moment rate increases), the relative size of an earthquake can be determined 10 seconds after the arrival of a P-wave. Meier et al., [2017] uses temporally normalized moment rate functions to show that we are unlikely to be able to differentiate the size of an earthquake until it is halfway over.
Using a generic empirical relationship between slip and rupture length, dictated by fault maturity [Perrin et al., 2016], to generate slip-rupture length templates, we backproject peak ground displacements (PGDs) along the temporally evolving rupture outputs from the earthquake early warning algorithm FinDer (Finite-Fault Rupture Detector) [Böse et al., 2012; 2015; 2017] to forward predict the final rupture length, magnitude, and associated probability distributions of large earthquakes. We typically find that we are able to predict the final rupture length and magnitude before half of the earthquake has occurred. [Hutchison et al., submitted, GRL]
Using a generic empirical relationship between slip and rupture length, dictated by fault maturity [Perrin et al., 2016], to generate slip-rupture length templates, we backproject peak ground displacements (PGDs) along the temporally evolving rupture outputs from the earthquake early warning algorithm FinDer (Finite-Fault Rupture Detector) [Böse et al., 2012; 2015; 2017] to forward predict the final rupture length, magnitude, and associated probability distributions of large earthquakes. We typically find that we are able to predict the final rupture length and magnitude before half of the earthquake has occurred. [Hutchison et al., submitted, GRL]
2. Slow Earthquakes
One of my primary interests lies in using seismology as a tool to better understand the underlying dynamics in fault zones -- particularly in the transition zone, the role of slow earthquakes in the earthquake cycle, source properties of slow earthquakes, and the mechanics of the transition zone. As a result of improved instrumentation, it has only been within the last 20 years that the geophysical community has discovered transitional behavior between the two end members of sliding behavior along faults: creeping and locked. (Does creep even exist on a very small scale?)
This transitional behavior is characterized by an umbrella of both seismic and aseismic events called slow earthquakes that are thought to follow a unique scaling relationship where their moment release is proportional to their duration, whereas the moment release of a regular earthquake is proportional to its duration cubed. Though, new research is beginning to call this relationship into question, demonstrating that slow earthquakes may behave more similarly to regular earthquakes than previously thought. The transition zone is thought to be regulated by rate weakening asperities surrounded by a rate strengthening background, though a variety of models within this conceptual framework still exist and require further constraints.
Beyond their source properties and their role in the earthquake cycle, I am also interested in the ways slow earthquakes can be used as a toolset in tomography, to generate GMPEs, and to determine properties such as attenuation factor.
This transitional behavior is characterized by an umbrella of both seismic and aseismic events called slow earthquakes that are thought to follow a unique scaling relationship where their moment release is proportional to their duration, whereas the moment release of a regular earthquake is proportional to its duration cubed. Though, new research is beginning to call this relationship into question, demonstrating that slow earthquakes may behave more similarly to regular earthquakes than previously thought. The transition zone is thought to be regulated by rate weakening asperities surrounded by a rate strengthening background, though a variety of models within this conceptual framework still exist and require further constraints.
Beyond their source properties and their role in the earthquake cycle, I am also interested in the ways slow earthquakes can be used as a toolset in tomography, to generate GMPEs, and to determine properties such as attenuation factor.
2.1 Very Low Frequency Earthquakes
Using a grid-search centroid moment tensor inversion method, I look for VLFEs in Cascadia. We successfully use these detections as template events to detect additional events. This indicates that VLFEs are repeating events. We determined that VLFE activity increases during episodic tremor and slip (ETS) events and decreases during the inter-ETS period. VLFEs are also consistent with slow-slip behavior, as detected by GPS, even when tremor is not. [Hutchison & Ghosh, 2016; Hutchison & Ghosh, 2018, JGR]
2.2 Transitional seismic activity in the San Jacinto Fault (Anza Gap)
Using visual inspection, multi-beam backprojection and envelope cross correlation, we detect and locate several discrete instances of tremor in June, 2011 in the Anza Gap. We interpret these events to be a manifestation of transient activity, particularly given the recent discovery of a slow-slip event in the Anza Gap following the El Mayor-Cucupah earthquake [Hutchison & Ghosh, 2017].
2.3 Regional and teleseismic triggered tremor, microseismic events, and foreshocks preceding the 2016 Mw 5.2 Borrego earthquake suggestive of dynamically triggered creep as a delayed triggering mechanism
On June 10, 2016, we observe a small tremor event in the Anza Gap triggered by the passage of Rayleigh waves from a Mw 6.1 earthquake in Puerto Morazán, Nicaragua, the second instance of triggered tremor observed in the San Jacinto Fault (SJF). The increase in dynamic stress, ~24 kPa, is consistent with values calculated for this region following the first instance of triggered tremor after the 2002 Denali Earthquake [Wang et al., 2013], indicating a greater fault strength, higher ambient stress, or combination of the two, than the Cholame section of the San Andreas Fault (SAF), where triggered tremor typically requires a dynamic stress increase of 10-20 kPa. Tremor is followed by microseismic events in the same region and radiate southeast, towards the trifurcation region of the SJF. The majority of these events are not included in the catalog. Energy from a second event, a Mw 6.2 from the Solomon Islands, arrives in the region ~2 hours later. Throughout the second teleseism, microseismic events continue. After teleseismic energy ceases, using a waveform based stacking method [Grigoli et al., 2018], a series of 10 foreshocks are detected, in addition to 7 cataloged events, which spatiotemporally precede a Mw 5.2 Borrego mainshock. Seemingly, these cascading events suggest the delayed dynamic triggering of the Borrego earthquake. In the context of previous studies suggesting regional deep creep, we propose that a dynamically triggered creep event, manifested through continued microseismic activity during and after the passage of teleseismic energy, ultimately triggered a mainshock. {Hutchison et al., 2019 - in revision]
3. Use of Archival Sources to Enhance Understanding of Natural Hazards
At times the geologic record is incapable of providing details that can and has often been communicated amongst people. Using a collection of archived accounts from a series of cascading disasters in the capitol city of Guatemala in 1717, we established a relationship between a volcanic eruption that was likely the source of continued seismic activity and later, a phreatic explosion. [Hutchison et al., 2016]