Date of Award

2019-01-01

Degree Name

Doctor of Philosophy

Department

Geological Sciences

Advisor(s)

Aaron A. Velasco

Abstract

Technological advances in combination with the onslaught of data availability allow for large seismic data streams to automatically and systematically be recorded, processed, and stored. Here, we develop an automated approach to identify small, local earthquakes within these large continuous seismic data records. Our aim is to automate the process of detecting small seismic events triggered by a distant large earthquake, recorded at a single station. Specifically, we apply time-domain short-term average (STA) to long-term average (LTA) ratio algorithms to three-component data to create a catalog of detections. We remove some of the false detections by requiring the detection be recorded on a minimum of two channels. To calibrate the algorithm, we compare our automatic detection catalog to a set of analyst-derived P-wave arrival times for a subset of small earthquakes occurring in the December 2008 Yellowstone swarm. Of the four STA/LTA algorithms we test (1 s/10 s; 4 s/40 s; 8 s/80 s; 16 s/160 s), the 1 s/10 s and 4 s/40 s detectors proved most effective at identifying the majority of events in the swarm. We apply these detectors to ±45 hrs and ±5 hrs of USArray data from the 2011 Japan M 9.0 and the 2010 Chile M 8.8 earthquakes, respectively. Using time-of-day versus number of detection relationships, we identify 38 of the 728 available stations that exhibit strong anthropogenic noise following the 2011 Japan earthquake. Our detection algorithm identified three regional earthquakes concurrent with the passage of the S and surface waves of the Chile mainshock at USArray station R11A that locate in the Coso region of California, as well as events in Texas following the Japan earthquake. This chapter is based on the paper (Velasco et. al., 2016).

Using 13 years of data (2004-2016) from the EarthScope USArray Transportable Array and the Southern California Seismic Network, we search for remotely triggered seismicity in the Coso Geothermal Field (CGF), California. We first apply a short term to long term average ratio detector to high-pass (5 Hz) filtered waveforms spanning ±5 hours encompassing 211 M ≥ 7 global earthquakes. We visually inspect these waveforms to identify uncatalogued local earthquakes. We use our augmented local earthquake catalog to investigate remote earthquake triggering in the CGF region. Flagging mainshocks with a statistically significant increase in seismicity following the P-wave arrival, we find that of the 211 remote mainshocks, 32 (15%) triggered seismicity in the CGF. An additional 9 mainshocks had local earthquakes coincident with the surface waves, but no statistically significant rate increase was found. Of the 41 (19%) triggering mainshocks, 28 and 13 exhibit instantaneous and delayed triggering, respectively. We find no correlation between triggering and mainshock depth, peak dynamic stress, nor mainshock focal mechanism type. However, the CGF is optimally oriented for remote triggering from mainshocks in the West Pacific, since dynamic stresses align favorably with the local stress field (maximum horizontal stress).

Induced seismicity, earthquakes caused by anthropogenic activity, has more than doubled in the last several years resulting from practices related to oil and gas production. Furthermore, large earthquakes have been shown to promote the triggering of other events within two fault lengths (static triggering), due to static stresses caused by physical movement along the fault, and also remotely from the passage of seismic waves (dynamic triggering). Thus, in order to understand the mechanisms for earthquake failure, we investigate regions where natural, induced, and dynamically triggered events occur, and specifically target Oklahoma. Utilizing 9 years of data (2008-2016) from EarthScope’s USArray Transportable Array, Oklahoma, and GS local networks, we search for dynamically triggered seismicity in Oklahoma (OK). We first apply a short term to long term average ratio detector to high-pass (5 Hz) filtered waveforms spanning ±5 hours encompassing 144 M ≥ 7 global earthquakes. We visually inspect these waveforms to identify uncatalogued local earthquakes. We use our augmented local earthquake catalog to investigate remote earthquake triggering in OK. Flagging mainshocks with a statistically significant increase in seismicity following the P-wave arrival, we find that of the 144 remote mainshocks, 24 (17%) statistically triggered seismicity in OK. An additional 31 mainshocks had local earthquakes coincident with the surface waves, but no statistically significant rate increase was found. Of the 55 triggering mainshocks, 51 and 4 exhibit instantaneous and delayed triggering, respectively. We find transient stresses can contribute to natural and induced stress states advancing the earthquake cycle and providing insight to the constantly changing stress state of induce systems. Utilizing the distribution of triggered earthquake populations in the OK region we identify regions particularly susceptible to earthquake hazards associated with sustained fluid injection.

Over 2.3 million people inhabit the El Paso-Juarez metropolitan area, a region that faces potential local seismic hazards related to the East Franklin Mountains fault (EFMF). The region is associated with low-level (M<4) seismicity, most recently a M=2.5 earthquake in March of 2012 located within the city limits of El Paso. Both El Paso and Juarez are located within the Rio Grande rift and are built upon Quaternary faults that pose a seismic hazard to the local population. To further assess seismic hazards in the region we conducted a paleoseismologic trenching study at a site in the main urban section of the EFMF fault system to characterize its mid-late Quaternary slip history. A paleoseismic trench was excavated across the East Franklin Mountains Fault at the McKelligan Canyon stepover on the only undisturbed section of the fault within the urbanized section of the fault. The trench site lies just east of Beaumont Medical center (BMC) on a 6.5-m high piedmont scarp about 1 km east of the range front. The trench exposed a spectacular exposure of the fault surface with a series of antithetic splays off the main fault as well as two minor subsidiary faults in the hanging wall. The excavation penetrated through a series of fault-related colluvial wedges in the hanging wall, into underlying alluvial fan strata that we correlate to alluvial fan strata exhumed in the footwall. Our field analysis of the trench leads to an interpretation of 5 colluvial wedges suggestive of ~1m of slip/event, with repeated patterns of surface rupturing suggestive of near-surface, subvertical mode-1 tensile fractures within the hanging wall developing into coseismic keystone graben at the fault scarp.

Language

en

Provenance

Received from ProQuest

File Size

158 pages

File Format

application/pdf

Rights Holder

Richard Alexander Alfaro-Diaz

Share

COinS