Earth-prints repository, logo   DSpace

About DSpace Software
|earth-prints home page | roma library | bologna library | catania library | milano library | napoli library | palermo library
Please use this identifier to cite or link to this item: http://hdl.handle.net/2122/9494

Authors: Phillips, W. S.*
Mayeda, K. M.*
Malagnini, L.*
Title: How to Invert Multi-Band, Regional Phase Amplitudes for 2-D Attenuation and Source Parameters: Tests Using the USArray
Title of journal: Pure and Applied Geophysics
Series/Report no.: 3-5/171 (2014)
Publisher: Springer Verlag
Issue Date: 2014
DOI: 10.1007/s00024-013-0646-1
Keywords: 2-D Attenuation
Abstract: We inverted for laterally varying attenuation, absolute site terms, moments and apparent stress using over 460,000 Lg amplitudes recorded by the USArray for frequencies between 0.5 and 16 Hz. Corner frequencies of Wells, Nevada, aftershocks, obtained by independent analysis of coda spectral ratios, controlled the tradeoff between attenuation and stress, while independently determined moments from St. Louis University and the University of California constrained absolute levels. The quality factor, Q, was low for coastal regions and interior volcanic and tectonic areas, and high for stable regions such as the Great Plains, and Colorado and Columbia Plateaus. Q increased with frequency, and the rate of increase correlated inversely with 1-Hz Q, with highest rates in low-Q tectonic regions, and lowest rates in high-Q stable areas. Moments matched independently determined moments with a scatter of 0.2 NM. Apparent stress ranged from below 0.01 to above 1 MPa, with means of 0.1 MPa for smaller events, and 0.3 MPa for larger events. Stress was observed to be spatially coherent in some areas; for example, stress was lower along the San Andreas fault through central and northern California, and higher in the Walker Lane, and for isolated sequences such as Wells. Variance reduction relative to 1-D models ranged from 50 to 90 % depending on band and inversion method. Parameterizing frequency dependent Q as a power law produced little misfit relative to a collection of independent, multi-band Q models, and performed better than the omega-square source parameterization in that sense. Amplitude residuals showed modest, but regionally coherent patterns that varied from event to event, even between those with similar source mechanisms, indicating a combination of focal mechanism, and near source propagation effects played a role. An exception was the Wells mainshock, which produced dramatic amplitude patterns due to its directivity, and was thus excluded from the inversions. The 2-D Q plus absolute site models can be used for high accuracy, broad area source spectra, magnitude and yield estimation, and, in combination with models for all regional phases, can be used to improve discrimination, in particular for intermediate bands that allow coverage to be extended beyond that available for high frequency P-to-S discriminants.
Appears in Collections:04.06.07. Tomography and anisotropy
Papers Published / Papers in press

Files in This Item:

File Description SizeFormatVisibility
Phillips_Mayeda_Malagnini_PAGEOPH2013.pdfMain article5.66 MBAdobe PDFonly authorized users View/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.


Share this record
Del.icio.us

Citeulike

Connotea

Facebook

Stumble it!

reddit


 

Valid XHTML 1.0! ICT Support, development & maintenance are provided by CINECA. Powered on DSpace Software. CINECA