A seismic velocity model for the upper mantle in northern Kenya derived from teleseismic earthquake data

Abstract

Teleseismic arrivals recorded by a network distributed over the northern part of the Kenya Rift in 1981 have been interpreted in terms of the structure of an anomalously low P-wave velocity upper mantle zone beneath the Kenya dome. The travel time residuals through a Jeffreys-Bullen Earth model have been compared with those derived previously by Savage and Long (1985). Their equivalence suggests that the large delays of up to about 1.9 seconds determined at stations well to the north of the apex of the Kenya dome are caused by a similar upper mantle low velocity zone as that presumed to exist beneath the centre and south of the dome. Slowness and azimuth anomalies of teleseismic arrivals recorded at a small aperture array at Ngurunit (NGU) in the Ndoto Mountains of the Turkana District on the north-eastern flank of the Kenya Rift show a relatively simple variation with azimuth and epicentral distance of event. It is suggested that this variation may be related to the shape of the upper surface of the anomalous mantle zone beneath the array. 3-dimensional ray-tracing through a low velocity mantle zone whose upper surface is described by a simple quadratic surface, enables a reasonable fit between observed and synthetic slowness and azimuth anomalies. The attitude of the quadratic surface implies a surprising thickening of the anomalous mantle zone to the north-east of the array. This together with previous geophysical results suggests a possible local thickening of the anomalous zone exists along most of the eastern flank of the Kenya Rift. Such a model has previously been proposed to explain the Quaternary volcanic activity from the Huri Hills in the north, through the Nyambeni range to the Chyulu Hills in the south, all lying about 150 – 200 km to the east of the main rift.