HR: 09:45h
AN: S51A-07    [PDF]
TI: Receiver-function imaging of the crustal and mantle structure beneath South Africa: a comparison of stacking and migration
AU: * Levander, A
EM: alan@esci.rice.edu
AF: Department of Earth Science, Rice University, 6100 Main St., Houston, TX 77005 United States
AU: Niu, F
EM: niu@rice.edu
AF: Department of Earth Science, Rice University, 6100 Main St., Houston, TX 77005 United States
AB: Receiver functions analysis has been widely used in imaging the crustal and mantle structure in recent years. Signal enhancement techniques have evolved from single-station to common-conversion-points (CCP) stacking of receiver-function traces. Kirchhoff and other migration techniques are also introduced in receiver-function imaging. It is generally believed that CCP stacking is suitable for horizontally layered structure, while, migration techniques are more favored but require denser receiver array for imaging complicated media. Here we present an investigation on the advantages and disadvantages of the two techniques using a high-quality dataset in a region with a relatively simple structure. The Kaapvaal Seismic Array we used in this study is approximately a linear array extending from southwest to northeast. We chose 6 shallow teleseismic events roughly located at the direction of the array. Both CCP stacking and Kirchhoff migration are performed with the same dataset. Due to the sparse distribution of ($\sim$50 km), image above $\sim$200 km is not well reconstructed from migrations. The two techniques, however, produced very consistent images below $\sim$200 km, especially in the transition zone and lower mantle depths. The CCP stacking shows a very well-defined crust-mantle transition. Crustal thickness is found to be well correlated to geologic terranes, consistent to the previous studies. The crust beneath undisturbed Archean Kaapvaal craton is relatively thin ($\sim$35 km), while the surrounding Proterozoic belts have thick crusts up to $\sim$50 km. The upper mantle in this region seems to be extremely transparent in the images, except some artificial structures between 100 and 200 km resulted from crustal reverberations. The 410-km discontinuity is extremely well mapped, with an apparent topographic relief of $\sim$15 km across the array (1400 km). Images also show a very clear 660-km discontinuity across the array, except for the $\sim$250 km section of the southwest end of the array, where the discontinuity is diffused by as much as 50 km. Compared to the 410-km, the 660-km discontinuity has less topography ($\sim$5 km) except for the southwest end of the array. The transition zone thickness beneath the Kaapvaal craton is about 10-15 km thicker than those beneath the Namaqua-Natal and Limpopo belts. Although the observation could result from unmodeled velocity variations in the transition zone, a 10-15 km variation in transition zone thickness would need at least 3 to 4% variations in velocity, which are not observed by tomographic images. Lateral variations in the P to S conversion amplitude are also observed across the array, with the largest amplitude being found in the craton region. All the observations suggest that the distinction between the Archean craton and the Proterozoic belts in South Africa may continue to as deep as the transition zone.
DE: 7203 Body wave propagation
DE: 7205 Continental crust (1242)
DE: 7218 Lithosphere and upper mantle
DE: 7260 Theory and modeling
SC: Seismology [S]
MN: 2003 Fall Meeting