Southern Africa velocity models

Fig. 3. (a) Sensitivity test of the higher-mode waveforms to the depth to the base of the upper-mantle lid for the SLR seismogram of the 18 July 1986 earthquake (Fig. 1, event 2). The continuous line is the observed waveform, the dotted line is the synthetic for the southern Africa velocity model of Qiu et al. (1996), and the dashed line is the synthetic for the same velocity model but with the lid base increased to the depth indicated at the left of each seismogram, (b) Same as (a) but for the SLR seismogram of the 10 March 1989 earthquake (Fig. 1, event 5). (c) Same as (a) but for the SUR seismogram of the 24 July 1991 earthquake to the minimum S-wave velocity of the low-velocity zone (LVZ). 

Zhao, M., Langston, C. A., Nyblade, A. A. Owens, T. J. 1999. Upper mantle velocity structure beneath southern Africa from modeling regional seismic data. Journal of Geophysical Research, 104, 4783-4794. 

McWilliams 1977). Qiu et al. (1996) used earthquakes in southern Africa recorded at stations in Zimbabwe and South Africa (Fig. 1) to obtain an average velocity model for southern Africa. The main features of their model are a high shear-wave velocity lid in the upper mantle shown by both seismic and petrological data, below which there is a decrease in the shear-wave velocity shown by the seismic data. Priestley (1999) reexamined the seismograms studied by Qiu et al. (1996) and included additional data to determine... 

Fig. I. Source-receiver paths for regional earthquake seismograms used in the studies of Qiu et al. (1996) and Priestley (1999) superimposed on the major crustal subdivisions of southern Africa. A, seismograph stations jlk, earthquake locations dotted lines denote paths for the events used in the regional waveform modelling. Events 1-8 were used by Qiu et al. (1996) event 9 was used by Priestley (1999). Fundamental mode Rayleigh wave phase velocity dispersion from Priestley (1999) was measured for the SUR-BOSA and BOSA-LBTB paths. The bold lines denote the extent of the South African array.

Fig. 2. Comparison of the density, shear-wave velocity and compressional-wave velocity profiles beneath southern Africa from Priestley (1999) (bold continuous lines), and the density and velocity profiles for PREM (fine continuous lines). The shaded area denote estimates of the uncertainties in the density and velocity model of Priestley (1999) derived from the waveform fitting tests, the earthquake location errors as described by Qiu et al. (1996), and c. 2% anisotropy as proposed by Vinnik et al. (1995).

Fig. 8. Three-component waveform fits at three distance ranges for synthetic seismograms (dotted lines) computed from composite velocity model consisting of the seismic lithosphere of the southern African model of Priestley (1999) above 160 km depth and the southern Africa upper-mantle model derived from the global tomographic model S12WM13 below 160 km depth compared with the observed seismograms (continuous lines) of the 14 August 1994 earthquake (Fig. 1, event 8) recorded at three distance ranges.

Masters et a/. s (1996) model and that in Figure 2 can also be compared by computing the difference in the vertical S-wave travel time between southern Africa and an old ocean basin for each model. Using Nishimura Forsyth s (1989) model for an ocean basin whose age is greater than llOMa, the difference between the travel times between 42 and 400 km is —1.00 s for model S16B30 and —2.01 s for the velocity model shown in Figure 2. 

Recent work in southern Africa has focused on developing a teleseismic travel time tomography image for the upper mantle (James et al. 2001 James Fouch 2002). The objective of teleseismic travel time tomography is to determine a 3D seismic image for a volume of Earth beneath the seismie array based on the arrival times from a large number of source-receiver combinations. The 3D model is normally expressed in terms of perturbations to a reference model in which the velocity is a function only of radius, where the 3D perturbations account for that part of the travel time not explained by the reference model. 

---