Three-component ambient noise beamforming in the Parkfield area

Katrin Löer*, Nima Riahi, Erik H. Saenger

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

33 Citations (Scopus)
8 Downloads (Pure)

Abstract

We apply a three-component beamforming algorithm to an ambient noise data set recorded at a seismic array to extract information about both isotropic and anisotropic surface wave velocities. In particular, we test the sensitivity of the method with respect to the array geometry as well as to seasonal variations in the distribution of noise sources. In the earth's crust, anisotropy is typically caused by oriented faults or fractures and can be altered when earthquakes or human activities cause these structures to change. Monitoring anisotropy changes thus provides time-dependent information on subsurface processes, provided they can be distinguished from other effects. We analyse ambient noise data at frequencies between 0. 08 and 0. 52 Hz recorded at a three-component array in the Parkfield area, California (US), between 2001 November and 2002 April. During this time, no major earthquakes were identified in the area and structural changes are thus not expected. We compute dispersion curves of Love and Rayleigh waves and estimate anisotropy parameters for Love waves. For Rayleigh waves, the azimuthal source coverage is too limited to perform anisotropy analysis. For Love waves, ambient noise sources are more widely distributed and we observe significant and stable surface wave anisotropy for frequencies between 0. 2 and 0. 4 Hz. Synthetic data experiments indicate that the array geometry introduces apparent anisotropy, especially when waves from multiple sources arrive simultaneously at the array. Both the magnitude and the pattern of apparent anisotropy, however, differ significantly from the anisotropy observed in Love wave data. Temporal variations of anisotropy parameters observed at frequencies below 0. 2 Hz and above 0. 4 Hz correlate with changes in the source distribution. Frequencies between 0. 2 and 0. 4 Hz, however, are less affected by these variations and provide relatively stable results over the period of study.

Original languageEnglish
Pages (from-to)1478-1491
Number of pages14
JournalGeophysical Journal International
Volume213
Issue number3
Early online date14 Feb 2018
DOIs
Publication statusPublished - 1 Jun 2018

Bibliographical note

ACKNOWLEDGEMENTS
Many thanks to Andrew Curtis and Ian Main for their critical and perceptive comments and to Sjoerd de Ridder for fruitful discussions. We are particularly grateful to Elmer Ruigrok for his detailed and constructive review, which has substantially improved our manuscript. We acknowledge the facilities of the IRIS Data Management System for providing access to the seismic data. This work is funded within the DFG project ‘SynPaTh’ and the EU project ‘GEMex’.

Keywords

  • North America
  • Seismic anisotropy
  • Seismic noise
  • Surface waves and free oscillations

Fingerprint

Dive into the research topics of 'Three-component ambient noise beamforming in the Parkfield area'. Together they form a unique fingerprint.

Cite this