Magnetctelluric measurements across the Kenyan Rift Valley
Prior to a discussion of the experimental results, a review of regional induction studies is presented, which emphasises the strengtbs and limitations of the magnetotelluric technique. The fundamental theory and equation of the magnetotelluric data are discussed. Instrumentation used for recording magnetotelluric variations in Kenya is described. Magnetotelluric measure~ents were made at fourteen stations in Kenya, between July 1973 and May 1974. The ma.in group formed a 300 km long east-west line crossing the Rift Valley at the geographic equator. Telluric field and magnetic field variations were recorded in the period range 10 to 1000 seconds. Data from all sites have been analysed to obtain principal impedances and single station induction vectors. Possible sources· of systematic error involved in the data reduction process have been investigated. In particular, the biasing effect of strong source field polarisation has been taken into account. Transfer func t.Lons expressed in maximum- and minimum- response form indicate a current concentration at shallow depths flowing along the axis of the Rift ValJey and are in accord with the results of Banks and Ottey ~1-974). Maximum apparent resistivities measured at stations inside the Rift Valley and on its escarpments ('Rift stations') are more than one order of magnitude smaller than th~aximum apparent resistivities measured at stations outside the Rift Valley. In addition, the azimuths of maximum resistivity axes are oriented approximately parallel to the main rift faults at 'Rift stations' and approximately perpendicular to the main rift faults at stations away from the Rift Valley. The apparent resistivity data suggest that some tens of kilometers of conductive material lie under the rift valley and that the western and eastern edges of this conductor lie outside the region defined by the major rift faults. Only data which is believed to be relatively unaffected by sources of systematic error have been selected for interpretation in terms of conductivity models. This eliminates the highly anisotropic apparent resistivity data obtained at all stations away from the Rift Valley, and also induction arrows calculated at stations to the east of the rift valley, where the inductive response is complex. Greatest weight has been placed on the high quality isotropic data collected at 01 Joro Orok, a station on the eastern escarpment of the Rift Valley. Simultaneous magnetic variation measurements made at stations 200 km apart suggest that the horizontal di~ensions of the inducing field do not significantly affect the main conclusions of the interpretation. The magnetotelluric data provide the first independent evidence, for the existence 01 high conductivities at depths corresponding to the upper mantle below the Kenyan Rift ·Valley. The conclusions of previous Geomagnetic Depth Sounding in the region were heavily dependent upon other geophysical and geological data. An explanation of the upper mantle conductor in terms of partial melting is in accord with geological evidence of recent volcanic activity in the area. The depth to the top of and the thickness of the upper mantle conductor can not be resolved by the MT data since the mantle conductor is obscured by the presence of a good conductor at depths « 10 km ) corresponding to the upper crust. The high conductivity of upper crustal material can only reasonably be explained in terms of high temperatures and water saturation of the crust under the rift valley. The conductivity model and its geophysical implications are similar to those suggested by Hermance (1973b) and Berktold (1974) in their interpretations of magnetotelluric data collected in Iceland and Ethiopia, respectively. The possible existence of large quantities of water in the crust below the Kenyan Rift Valley and in the Sub-Icelandic and Sub-Ethiopian crusts suggests a relationship- between high water concentration', high temperatures and tectonic activity. The presence of water in rocks at high temperatures is of considerable interest in the understanding of the petrology of rocks in the region of the Rift Valley.