S-wave velocity below Europe from delay-time and waveform inversions

The upper mantle of Europe has been the subject of many tomographic studies of variations in P-wave velocity (e.g., Romanowicz 1980; Hovland et al., 1981; Spakman, 1988, 1991; and Spakman et aI., in preparation). In particular the studies of Spakman (1988, 1991) and Spakman et aI. (in preparation) provided detailed images of the upper mantle. However, these images leave us with questions as to how reliable they are, or, more specifically, in which regions they are reliable. Moreover, they do not provide sufficient information to determine the causes of the velocity heterogeneities. Our knowledge about the variations in composition and temperature in the upper mantle is limited. The availability of accurate measurements of both the P- and S-wave velocity would enable us to answer the question whether the velocity variations are caused by temperature variations, differences in composition, or both (Yan et al., 1989; Furlong et aI., 1992). As mentioned above, high-resolution P-wave models are available from seismic tomography, although the amplitudes of the velocity variations are not well determined. Several authors have studied the S wave velocity in the upper 200 km of the mantle with phase velocity measurements of surface waves (e.g., Berry and Knopoff, 1967; Panza et aI., 1980; Souriau, 1981; Mariller and Mueller, 1985). Snieder (1988) inverted waveforms of fundamental mode Rayleigh waves for a three-dimensional (3-D) model of the S velocity in the upper 200km of the mantle. From modelling broadband waveforms of S body waves Paulssen (1987) determined a set of I-D velocity models to 670 km depth for different paths crossing central and western Europe. However, these studies provided only limited information about the variations of the S wave velocity in the upper mantle on the length scale of anomalies in the 3-D P-wave velocity models. Therefore, the objective of the research described in this thesis is to find a model for variations in S wave velocity in which structure is resolved on length scales comparable to those in the 3-D P-wave velocity models. We discuss three different methods: body wave tomography with S-wave traveltime residuals published by the International Seismological Centre (ISC) (chapters 2 and 3); the inversion of phase velocities of surface waves (chapter 4); and the method of partitioned inversion (Nolet, 1990) of waveforms of both body- and surface waves (chapters 5,6, and 7).Paulssen (1987) determined a set of I-D velocity models to 670 km depth for different paths crossing central and western Europe. However, these studies provided only limited information about the variations of the S wave velocity in the upper mantle on the length scale of anomalies in the 3-D P-wave velocity models. Therefore, the objective of the research described in this thesis is to find a model for variations in S wave velocity in which structure is resolved on length scales comparable to those in the 3-D P-wave velocity models. We discuss three different methods: body wave tomography with S-wave traveltime residuals published by the International Seismological Centre (ISC) (chapters 2 and 3); the inversion of phase velocities of surface waves (chapter 4); and the method of partitioned inversion (Nolet, 1990) of waveforms of both body- and surface waves (chapters 5,6, and 7)