Detecting and Locating Small Earthquakes On Remote Oceanic Transform Faults

Open Access
Author:
Vieceli, Rhiannon Elizabeth
Graduate Program:
Geosciences
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
November 30, 2015
Committee Members:
  • Charles James Ammon, Thesis Advisor
  • Andrew Arnold Nyblade, Thesis Advisor
  • Sridhar Anandakrishnan, Thesis Advisor
Keywords:
  • Cross Correlation
  • Seismic event detection
  • Oceanic transform faults
  • Earthquake seismology
Abstract:
Although oceanic transform faults (OTFs) constitute a small fraction of the total plate boundary area, better constraints on OTF tectonic parameters (e.g. fault length, slip rate, thermal structure) compared to other tectonic boundaries make OTFs a useful focus of the investigation of earthquake processes. OTF earthquakes are also accompanied by a large fraction of aseismic deformation that makes them an interesting target for exploring the interaction of creep with slow and quick earthquakes. Because most typical OTFs are quite remote, even indirectly observing their deformation processes is a serious challenge. Standard teleseismic analysis methods have yielded valuable constraints on the first-order characteristics of moderate-to-large magnitude OTF earthquakes, but fundamental questions regarding rupture length and area as well as rupture-front propagation speed remain unknown in these systems; identifying the smaller-magnitude activity that often provides insight to some of these quantities is difficult. Short-period seismic arrays occasionally provide adequate information needed to locate small ($m_b \leq 4.5$) earthquakes along Mid-Atlantic transforms. In this work, we explore the possibility of detecting and locating small earthquakes along remote OTFs using intermediate-period waveform-based comparisons (e.g. cross-correlations) of template signals with the continuous seismic wavefield observed at seismic stations surrounding several OTFs. % The best results were obtained when using data with a bandwidth of $20-25 s \leq T \leq 50 s$ and when implementing the Match-and-Locate (M\&L) method of \emph{Zhang and Wen} [2015] that accounts for modest location differences between the template and target events. As expected, station coverage, template-station geometry, and the number of channels with high-quality data affect the results. When the M\&L method is used, valuable information on relative earthquake locations and magnitude estimates is obtained. While our results indicate that template events may be able to detect earthquakes occurring along the same transform at distances up to $\sim$250 km, the greatest results will be produced when searching for earthquakes closer to the template; a template event of $M_W$ 5.5 with 21 channels of high-quality data is expected to identify a co-located $M_W$~3.4 event roughly 73\% of the time and a template event of $M_W$ 5.0 with 66 channels of high-quality data is expected to identify a co-located $M_W$ 3.5 roughly 87\% of the time. % We applied the M\&L method to the Romanche Transform Fault of the Central Mid Atlantic, and detected 15 small earthquakes occurring near the eastern edge of the fault that may have gone undetected in between 2012-01-01 and 2014-07-01. We also applied the approach to characterize the foreshock and aftershock sequences of the 2015-02-13 $M_W$ 7.1 earthquake of the Charlie-Gibbs Transform Fault in the North Atlantic. We did not detect any uncatalogued foreshocks, but we identified six uncatalogued aftershocks. Still, the results are consistent with experience - oceanic transform systems produce substantially fewer aftershocks that comparable continental systems.