Global high-resolution phase velocity distributions of overtone and fundamental-mode surface waves determined by mode branch stripping

Abstract

Summary We apply the mode branch stripping (MBS) technique of van Heijst & Woodhouse (1997) to approximately 110 000 three‐component seismograms. We select high‐quality data with the reliability estimate of van Heijst & Woodhouse (1997). We assess the influence of different selection criteria and remove outliers using smooth degree‐12 phase velocity maps. We present final results in terms of Rayleigh‐ and Love‐wave fundamental‐mode and overtone global high‐resolution (l=40) phase velocity maps. We determine the optimum damping of the high‐resolution maps with a method based on cross‐validation. Our fundamental‐mode phase velocity maps are generally in good agreement with previous studies, especially with Ekström et al.(1997). They do, however, contain more short‐wavelength structure than previous studies as we apply relatively little damping. The Rayleigh wave overtone phase velocity measurements made with MBS are of high quality in broad frequency ranges. The measurements are generally well explained by phase velocity maps, and variance reductions for some modes, after rejecting outliers, are as high as 85 per cent. We compare our global phase velocity distributions to the previous results of Stutzmann & Montagner (1994) and to phase velocity maps predicted by 3‐D tomographic mantle models. Agreement of the model predictions with our Rayleigh wave phase velocity maps, in terms of both amplitude and observed structures, is good. For Love waves, the quality of the measurements made with MBS is not as high as for Rayleigh waves. The variance reductions achieved are lower and the agreement between model predictions and our phase velocity maps is less, especially in the frequency ranges where interference between different mode branches is strong. Finally, as an additional check on the quality of our overtone measurements, we present a comparison of the fundamental‐mode Rayleigh phase velocity distribution at 40 s and our fourth Rayleigh wave overtone phase velocity distribution at 62 s. These two modes are similarly sensitive to velocity anomalies in the top of the upper mantle. We show that the two phase velocity maps are in very close agreement.

Using probabilistic seismic tomography to test mantle velocity–density relationships

“…We have combined the NM data sets listed in Section 1 with two recent catalogues of SW fundamentals [31] and overtones [32]. In addition, we employ a small but important set of recently measured splitting functions for mantle v p multiplets [33] recently measured with generalized spectral ¢tting, a new NM catalogue compiled using an autoregressive form of spectral ¢tting [34], and a large new collection of NM fundamental and overtone splitting functions retrieved using regionalized stripping [35].…”

Minor-arc and major-arc global surface wave diffraction tomography

“…Some regions of the Earth, especially in the southern hemisphere, cannot be e ectively covered by minor-arc paths due to the sparseness of seismic stations. The use of major-arc data for boththe fundamental mode and overtone data van Heijst et al, 1999 would signi cantly improve the spatial and azimuthal coverage particularly for studies of azimuthal anisotropy. Spetzler et al 2002 discuss di raction tomography for major-arc measurements, but minor and major-arc observations have been previously used in tomographic studies only under the assumption of ray-theory e.g., Trampert & Woodhouse, 2003. In this paper we follow Spetzler et al 2002 to extend di raction tomography b y rede ning the zone of sensitivity and accommodating both minor-arc and major-arc measurements using the Born Rytov approximation.…”

A global horizontal shear velocity model of the upper mantle from multimode Love wave measurements

“…Love and Rayleigh wave dispersion maps for the fundamental mode exist (e.g. Trampert & Woodhouse 1995Larson & Ekström 2001;Visser et al 2007;Ekström 2011;Pasyanos et al 2014) and Rayleigh wave maps exist for some higher modes (van Heijst & Woodhouse 1997, 1999Shapiro & Ritzwoller 2002;Beucler et al 2003;Visser et al 2007Visser et al , 2008. Less work has been done on inverting Love wave measurements for β h (z) models (Li & Romanowciz 1996;Mégnin & Romanowicz 2000;Zhou et al 2005) and most existing β h (z) upper-mantle models are constrained using only fundamental mode measurements.…”