On the improvement of SKS splitting measurements by the Simultaneous Inversion of Multiple Waveforms (SIMW). Issue 3 (23rd December 2016)

• Record Type: Journal Article
• Authors: Roy, C.; Winter, A.; Ritter, J.R.R.; Schweitzer, J.
• Is Part Of: Geophysical journal international. Volume 208:Issue 3(2017:Mar.)
• Page Start: 1508
• Page End: 1523
• Publication Date: 2016-12-23
• Subjects: Mantle processes; Body waves -- Seismic anisotropy -- Dynamics of lithosphere and mantle; Geophysics -- Periodicals; 550
• DOI: 10.1093/gji/ggw470
• Languages: English
• ISSNs: 0956-540X

Abstract

Summary: The birefringence of core-refracted shear waves (e.g. SKS or SKKS) is often used to study seismic anisotropy in the Earth. However, depth resolution and multilayer anisotropy is generally poor for many regions on Earth. This is primarily due to SKS or SKKS phases that are not observable for different backazimuths either because of missing seismicity at the required distance range or because of a too low signal-to-noise ratio (SNR). We propose a new method called Simultaneous Inversion of Multiple Waveforms (SIMW), which allows the joint inversion of multiple core-refracted shear waves from different earthquakes within the same source region, observed by either the same seismic station or by a seismic network. The waveforms are concatenated into a combined signal, which is then inverted with the Silver & Chan method to determine the two splitting parameters: time delay δt, and fast polarization direction Φ. We apply our method to recordings at the large aperture Norwegian NORSAR Array and the German Gräfenberg array (GRF). Our results demonstrate that SIMW allows a stable determination of splitting results for low-amplitude or noisy SKS signals. Splitting parameter uncertainties can be reduced and reliable results are obtained for both arrays. Moreover, new backazimuth directions can be explored, enabling a more accurate derivation of two-layer anisotropy models. Our new methodology is particularly helpful for temporary station deployments with limited recording times in order to utilize as many as possible signals including such with low-amplitude and small SNR.