Insights into rifting from shear wave splitting and receiver functions: an example from Ethiopia

Seismic anisotropy beneath broad-band stations in the vicinity of the East African rift are compared with those on stable cratonic parts of Africa and Arabia. Such measurements offer potential constraints on rift processes, absolute plate motions (APM) and tectonic structure. New SKS shear wave splitting parameters are analysed beneath the broad-band stations of FURI and AAE (Ethiopia), BGCA (Central African Republic) and RAYN (Saudi Arabia). The number of events considered at the four stations varies from 13 to 32 and provides good azimuthal coverage. Stations on or near the rift show the polarization of the fast shear wave (phi) aligned parallel to the rift axis. The magnitude of the splitting delay (deltat) increases northward along the East African rift. Previously published measurements in Kenya show the smallest splitting value (1.0 s), whilst the Djibouti station, ATD, shows the largest splitting (1.6 s). The Ethiopian results (deltat = 1.38 + 0.03 s, phi = 36degrees + 1) show constancy in deltat and phi with respect to backazimuth, thus, suggesting a single anisotropic layer beneath the stations. There is no observed correlation of phi with APM direction. Less splitting (deltat) is observed beneath cratonic parts of Africa. BGCA in central Africa shows splitting parallel to the inferred direction of transpression, not the APM direction. Receiver-function analysis at FURI and AAE supports evidence from refraction experiments of thick crust (ca 40 km) in the region of continental rifting, however, the analysis shows a deeper interface at a depth of 90 km, also. This interface may mark the base of the lithosphere in this region. One interpretation of the splitting results is that the anisotropy at the Ethiopian stations is the result of aligned melt in this upper 90 km of lithosphere. A < 1 per cent volume fraction of melt aligned in thin (aspect-ratio <0.03) vertical ellipsoidal pockets generates sufficient splitting to explain the data. Higher splitting magnitudes in the north correlates with higher melt production observed in the Ethiopian part of the rift. Alternatively, anisotropy may be the result of the alignment of olivine in the asthenosphere parallel to the ridge axis, as material flows laterally to fill the gap caused by lithospheric extension. This would either suggest a northward-thickening anisotropic layer beneath the rift or enhanced olivine alignment as a result of changes in strain. Whatever the mechanism, it appears that the anisotropy is sensitive to, and provides insight into, the transition from continental to oceanic rifting in the northern Ethiopian rift.