PageRank for Earthquakes

Aguiar, A. C.; Beroza, Gregory C.

doi:10.1785/0220130162

Cited by 26 publications

(27 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We first perform an initial cross correlation on all windows following Brown et al [2008] and later use an adaptation of Google's PageRank algorithm [Page et al, 1999] to find directly and indirectly linked windows that comprise robust LFEs. We focus on windows with high PageRank values and stack all other windows linked, both directly and indirectly, to the highly ranked window to create templates following Aguiar and Beroza [2014]. Once templates are established, we use them to search continuous data by template matching.…”

Section: Pagerank Applied To Tremor In Taiwanmentioning

confidence: 99%

“…This is a narrower frequency band than the typical 1-8 Hz used in other areas; however, narrow bandwidth tremor is also observed in Guerrero, Mexico (1-2 Hz) [Frank and Shapiro, 2014;Frank et al, 2013;Payero et al, 2008] and in the Hikurangi subduction zone, in northern New Zealand [Kim et al, 2011]. We perform a cross correlation of each window with all others, using sliding windows every 0.1 s during the 30 min of ambient tremor data ( Figure 1a) to find pairs that are above a threshold of 3σ, where σ = 1.253 × mean absolute deviation, based on a Gaussian assumption [Aguiar and Beroza, 2014]. We later apply the PageRank algorithm to create an LFE template for each of the 10 stations available during this time period.…”

Section: Ambient Tremormentioning

confidence: 99%

“…More recent studies show that tremor in Taiwan is not only triggered but also occurs spontaneously [Idehara et al, 2014;Sun et al, 2015]. In this study we focus on analyzing LFEs within tremor in Taiwan by using PageRank as an earthquake detection method [Aguiar and Beroza, 2014] to find LFEs during both triggered and ambient tremor without a priori knowledge of LFE waveforms. Large location uncertainties have left it unclear whether the downward extension of the Chaochou-Lishan Fault or subduction of the Eurasian plate is the tremor source region.…”

Section: Introductionmentioning

confidence: 99%

“…Large location uncertainties have left it unclear whether the downward extension of the Chaochou-Lishan Fault or subduction of the Eurasian plate is the tremor source region. In this study we focus on analyzing LFEs within tremor in Taiwan by using PageRank as an earthquake detection method [Aguiar and Beroza, 2014] to find LFEs during both triggered and ambient tremor without a priori knowledge of LFE waveforms. We compare LFEs found from both data sets and use local earthquakes to find the P wave signal of the LFEs to aid in the location of the events.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Tectonic tremor and LFEs on a reverse fault in Taiwan

Aguiar

Chao

Beroza

2017

Geophysical Research Letters

View full text Add to dashboard Cite

We compare low‐frequency earthquakes (LFEs) from triggered and ambient tremor under the southern Central Range, Taiwan. We apply the PageRank algorithm used by Aguiar and Beroza (2014) that exploits the repetitive nature of the LFEs to find repeating LFEs in both ambient and triggered tremor. We use these repeaters to create LFE templates and find that the templates created from both tremor types are very similar. To test their similarity, we use both interchangeably and find that most of both the ambient and triggered tremor match the LFE templates created from either data set, suggesting that LFEs for both events have a common origin. We locate the LFEs by using local earthquake P wave and S wave information and find that LFEs from triggered and ambient tremor locate to between 20 and 35 km on what we interpret as the deep extension of the Chaochou‐Lishan Fault.

show abstract

Section: Pagerank Applied To Tremor In Taiwanmentioning

confidence: 99%

Section: Ambient Tremormentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Tectonic tremor and LFEs on a reverse fault in Taiwan

Aguiar

Chao

Beroza

2017

Geophysical Research Letters

View full text Add to dashboard Cite

show abstract

“…The observation that LFEs occur repeatedly at localized asperities with the same source mechanism has led to multichannel network detection approaches to search for waveform repetitions and automatically identify LFEs for use as new templates (Brown et al, 2008;Aguiar and Beroza, 2014). A second class of approach, more akin to the manual/ visual approach taken by analysts and termed "cross-station approaches" by Rubin and Armbruster (2013) relies on the similarity of waveforms from an individual LFE recorded across multiple stations.…”

Section: Introductionmentioning

confidence: 99%

Detection and Location of Low‐Frequency Earthquakes Using Cross‐Station Correlation

Savard

Bostock

2015

Bulletin of the Seismological Society of America

View full text Add to dashboard Cite

Source geometry and simple propagation of S waves generated by lowfrequency earthquakes (LFEs) in northern Cascadia results in strong waveform coherence on horizontal channels for stations at smaller epicentral distances. As recognized by previous workers, this cross-station similarity can be exploited for detection of LFEs in tremor. We develop a cross-station correlation approach based on waveform coherence and travel-time consistency that exploits a full complement of network stations, and we demonstrate its application to separate arrays on Vancouver Island and Washington state. Our approach yields thousands of impulsive LFE detections per slow-slip episode and hundreds of events between episodes. LFE epicentral distributions reveal the presence of well-defined regions of high asperity density with those regions farther downdip that are active during interepisodic tremor and slow-slip periods. The LFE epicenters also display pronounced spatiotemporal clustering that compares favorably with independent tremor and LFE template catalogs and that defines rapid tremor reversals with complex propagation patterns. The LFE source depths can be estimated for some detections in which P waves are identified on the vertical channel through correlation with the horizontal-component S waveforms. LFE hypocenters cluster between two recently developed plate interface models below southern Vancouver Island. Our approach enables the identification of high signal-to-noise-ratio impulsive LFE detections from targeted regions on the plate boundary for generation of LFE templates to be used in studies of structure and seismogenesis in Cascadia.

show abstract