Mark R. Muller, Alan G. Jones, Stuart Fishwick, Chris Hatton, Herman Grütter, Rob L. Evans, Xavi Garcia, Marion P. Miensopust, Mark P. Hamilton and the SAMTEX Team Dublin Institute for Advanced Studies, 5 Merrion Square, Dublin 2, Ireland University of Leicester, Department of Geology, University Road, Leicester, LE1 7RH, U.K. MSA Geoservices, Johannesburg, South Africa BHP Billiton, 6 Hollard Street, Johannesburg 2001, South Africa Woods Hole Oceanographic Institution, Department of Geology and Geophysics, Clark South 172, 360 Woods Hole Road, Woods Hole, Massachusetts, 02543-1542, U.S.A. EMGS, Stiklestadveien 1, N-7041 Trondheim, Norway

[1] A regional-scale magnetotelluric (MT) experiment across the southern African Kaapvaal craton and surrounding terranes, called the Southern African Magnetotelluric Experiment (SAMTEX), has revealed complex structure in the lithospheric mantle. Large variations in maximum resistivity at depths to 200–250 km relate directly to age and tectonic provenance of surface structures. Within the central portions of the Kaapvaal craton are regions of resistive lithosphere about 230 km thick, in agreement with estimates from xenolith thermobarometry and seismic surface wave tomography, but thinner than inferred from seismic body wave tomography. The MT data are unable to discriminate between a completely dry or slightly “damp” (a few hundred parts per million of water) structure within the transitional region at the base of the lithosphere. However, the structure of the uppermost ∼150 km of lithosphere is consistent with enhanced, but still low, conductivities reported for hydrous olivine and orthopyroxene at levels of water reported for Kaapvaal xenoliths. The electrical lithosphere around the Kimberley and Premier diamond mines is thinner than the maximum craton thickness found between Kimberley and Johannesburg/Pretoria. The mantle beneath the Bushveld Complex is highly conducting at depths around 60 km. Possible explanations for these high conductivities include graphite or sulphide and/or iron metals associated with the Bushveld magmatic event. We suggest that one of these conductive phases (most likely melt-related sulphides) could electrically connect iron-rich garnets in a garnet-rich eclogitic composition associated with a relict subduction slab.

A 1400 km-long, 2-D magnetotelluric (MT) profile, consisting of 69 sites at 20 km intervals, across the western part of the Archaean Kaapvaal Craton, the Proterozoic Rehoboth Terrane and the Late Prot erozoic/Early Phanerozoic Ghanzi-Chobe/Damara Belt reveals significant lateral heterogeneity in the electrical resistivity structure of the southern African lithosphere. The lithospheric structures of the Rehoboth Terrane and Ghanzi-Chobe/ Damara Belt have not been imaged previously by geophysical methods. Temperature is the primary control on the resistivity of mantle minerals, and the MT derived lithospheric thicknesses therefore provide a very reasonable proxy for the "thermal" thickness of the lithosphere (i.e., the thickness defined by the intersection of a conductive geotherm with the mantle adiabat), allowing approximate present-day geotherms to be calculated. The work indicates the following present-day average lithospheric thicknesses, to a precision of about 20 km, for each of the terranes traversed (inferred geotherms in brackets): Eastern Kimberley Block of the Kaapvaal Craton 220 km (41 mW m(-2)), Western Kimberley Block 190 km (44 mW m(-2)), Rehoboth Terrane 180 kin (45 mW m(-2)) and Ghanzi-Chobe/Damara Belt 160 km (48 mW m(-2)). A clear relationship between the electrical resistivity structure of the lithosphere and the tectonic stabilisation-age of the terrane is evident. Good agreement between the inferred present-day lithospheric geotherms and surface heat flow measurements indicate the latter are strongly controlled by variations in lithospheric thickness. A significant difference in lithospheric thickness is observed between the Eastern and Western Kimberley blocks, and is consistent with previous seismic tomography images of the Kaapvaal Craton. The present-day lithospheric thickness, and reduced depth extent into the diamond stability field, accounts for the absence of diamondiferous kimberlites in the Gibeon and Gordonia kimberlite fields in the Rehoboth Terrane. Previously published mantle xenolith P-T arrays from the Gibeon, Gordonia and Kimberley fields, however, suggest that the Rehoboth Terrane had equilibrated to a cooler conductive palaeo-geotherm (40-42 mW m(-2)), very similar to that of Eastern Kim berley Block of the Kaapvaal Craton, at some time prior to the Mesozoic eruption of the kimberlites. The timing and nature of both the thermal equilibration of the Rehoboth Terrane, and the subsequent lithospheric heating/thinning event required to account for its present-day lithospheric structure, are not well constrained. A model consisting of the penetration of heat transporting magmas into the lithosphere, with associated chemical refertilisation, at an early stage of Mesozoic thermalism appears to be the most plausible model at present to account for both the present-day lithospheric structure of the Rehoboth Terrane and an earlier, cooler palaeogeotherm. Some problems, however, remain unresolved in terms of the isostatic response of the model. Based on a compilation of xenocryst Cr/Ca-in-pyrope barometry observations, the extent of depleted mantle in the Rehoboth Terrane is found to be significantly reduced with respect to the Eastern Kimberley Block: 117 km versus 138-167 km. It appears most likely that the depletion depth in both terranes, at least in the vicinity of kimberlite eruption, is explained by refertilisation of the lower lithospheric mantle. (C) 2009 Published by Elsevier B.V.

The proposed boundaries of geological terranes in northeastern Botswana are mainly based on regional magnetic and gravity data, because there are not many outcrops available due to the thick Phanerozoic cover rocks. The extent of the Zimbabwe craton into Botswana as well as the location of the boundaries to its neighbouring mobile belts (Limpopo Belt, Magondi Mobile Belt and Ghanzi-Chobe Belt) are not very well known. Magnetotelluric (MT) profiles of the Southern African MagnetoTelluric EXperiments (SAMTEX) are present in this area and provide information about lithospheric strike directions and the resistivity distributions as well as possible locations of terrane boundaries, which verify some of the proposed terrane boundaries and suggest modification for others (e.g., Ghanzi-Chobe Belt to Magondi Mobile Belt terrane boundary and the western boundary of the Zimbabwe craton).

The Southern African Magnetotelluric Experiment (SAMTEX) is the largest ever land-based magnetotelluric (MT) project. The main objective of the project is to define the geo-electric structure across the region in order to gain a better understanding of Archean and Proterozoic tectonic processes. Only the MT profiles crossing the Rehoboth Terrane, the Neoproterozoic Ghanzi-Chobe/Damara belts (collectively termed the DMB) and the southern Angola craton are the focus of this study. One of the ways in which geo-electrical structural information is obtained is by detailed analysis of the measured impedance tensor. The Groom and Bailey decomposition technique was applied to the MT data and indicates significant depth and along-profile variations in geo-electric strike and dimensionality on all transects crossing these three tectonic units (i.e. Rehoboth Terrane, Angola craton and the DMB). The geo-electric strikes are generally parallel to the north-east trending tectonic fabric as inferred from the magnetic data, but the significant strike variations with depth are expressions of heterogeneity in the lithospheric structure. The Rehoboth terrane, south of the DMB, exhibits a strongly one dimensional (1D) to moderate two dimensional (2D) structure, with preferred strike directions in the range 200-450 20for the crust-mantle period (i.e. depth) range, indicating little crust-mantle decoupling. The DMB appears to be moderately 2D at lower crustal and upper mantle depths (10-100 s) with no consistent/preferred strike direction and significant phase differences between the conductive and resistive directions. North of the DMB and into the Angola craton there are significant variations in geo-electric strike direction and dimensionality at most sites for lower-crustal and upper mantle lithosphere. Our results further indicate that the profiles have to be divided into smaller areas having similar strike directions to allow for 2D modelling and inversion.

SAMTEX (Southern African Magnetotelluric Experiment) is a multinational project that was initiated in 2003 to study the regional-scale electrical conductivity substructure of southern Africa and to infer from it the tectonic processes involved in the formation of the southern African subcontinent. Audio-magnetotelluric (AMT) data recorded during the most recent phase of the experiment are evaluated to investigate the local-scale conductivity substructure in the Otjiwarongo and Katima Mulilo regions, where in future the installation of high-voltage direct current (HVDC) earth electrodes will commence. Both of the AMT surveys are situated close to the edge of the orogenic Damara Mobile Belt in northern and north-eastern Namibia. Previous studies using magnetotellurics (MT), magnetometer arrays and geomagnetic observatory data all provide evidence of the relatively conductive nature of the Damara Mobile Belt. The Damara Mobile Belt represents in part the collision between the Congo and Kalahari cratons during the amalgamation of South Gondwana and its high conductivity is explained by the presence of interconnected fluids and conductive materials (graphites, sulphides). In contrast, the lithospheric structure of the ancient Archaean cratons, the Congo and Kalahari, are generally found to be electrically resistive and therefore it is hypothesized that the return path of DC current, flowing along the path of least resistance between the two electrodes, is most likely to lie somewhere within or in the vicinity of the Damara Mobile Belt. To obtain a better understanding of the current flow we propose using geological information, previous results of studies of the conductivity of the Damara Mobile Belt and surrounding regions and 2D inversion results from the AMT and MT data recorded during SAMTEX in northern Botswana and Namibia, as input to a 3D DC resistivity forward modelling code, and try to predict the return path that the DC current will follow.

A 1400 km-long, 2-D magnetotelluric (MT) profile across the Archaean Kaapvaal Craton, the Proterozoic Rehoboth Terrane and the Late Proterozoic/Early Phanerozoic Ghanzi-Chobe/Damara Belt reveals significant lateral heterogeneity in the electrical resistivity structure of the southern African lithosphere. The profile indicates the following present-day average lithospheric thicknesses, to a precision of about ± 20 km, for each of the terranes traversed (inferred conductive geotherms in brackets): Eastern Kimberley Block of the Kaapvaal Craton 220 km (41 mWm-2), Western Kimberley Block 190 km (44 mWm-2), Rehoboth Terrane 180 km (45 mWm-2) and Ghanzi-Chobe/Damara Belt 160 km (48 mWm-2). Previously published mantle xenolith pressure-temperature (P-T) arrays from the Gibeon, Gordonia and Kimberley fields, however, suggest that the Rehoboth Terrane had equilibrated to a cooler conductive palaeo-geotherm (40 – 42 mWm-2 ) very similar to that of Eastern Kimberley Block of the Kaapvaal Craton, at some (unconstrained) time prior to the Mesozoic eruption of the kimberlites. A model consisting of the penetration of heat transporting magmas into the lithosphere, with associated chemical refertilisation, at an early stage of Mesozoic thermalism appears to be the most plausible model at present to account for both the present-day lithospheric structure of the Rehoboth Terrane and an earlier, cooler palaeo-geotherm. Some problems, however, remain unresolved in terms of the isostatic response of the model. Based on a compilation of xenocryst Cr/Ca-in-pyrope barometry observations, the extent of depleted mantle in the Rehoboth Terrane is found to be significantly reduced with respect to the Eastern Kimberley Block: 117 km versus 138 – 167 km. It appears most likely that the chemical depletion depth in both terranes, at least in the vicinity of kimberlite eruption, is accounted for by refertilisation of the lower lithospheric mantle.

Observations of anisotropy, when understood in terms of deformation processes, are critical to illuminating the dynamics of past and present tectonic processes. In particular anisotropy can infer how continents formed, stabilized and interacted with underlying mantle regions in the past, and how they do so today. Seismology and electromagnetic observations of anisotropy are essential if we are to understand the tectonic history of a region. Seismic anisotropy, defined from SKS arrivals, is poorly constrained in depth, whereas electrical anisotropy has inherent depth localization but lower spatial resolution. Given the limitations of both sub-disciplines a more robust characterization of anisotropy is achieved by integrating complementary datasets. Southern Africa has now two rich geophysical databases from the SASE and SAMTEX experiments that can be explored, compared and contrasted for lithospheric anisotropy. Doing so suggests a new model to explain SKS observations which is based on plausible tectonic history. The new model combines the “Silver” lithospheric anisotropy and “Vinnik” asthenospheric anisotropy models, and incorporates their differentiation using electrical anisotropy.

Southern Africa, particularly the Kaapvaal Craton, is one of the world's best natural laboratories for studying the lithospheric mantle given the wealth of xenolith and seismic data that exist for it. The Southern African Magnetotelluric Experiment (SAMTEX) was launched to complement these databases and provide further constraints on physical parameters and conditions by obtaining information about electrical conductivity variations laterally and with depth. Initially it was planned to acquire magnetotelluric data on profiles spatially coincident with the Kaapvaal Seismic Experiment, however with the addition of seven more partners to the original four through the course of the experiment, SAMTEX was enlarged from two to four phases of acquisition, and extended to cover much of Botswana and Namibia. The complete SAMTEX dataset now comprises MT data from over 730 distinct locations in an area of over one million square kilometres, making SAMTEX the largest regional-scale MT experiment conducted to date.Preliminary images of electrical resistivity and electrical resistivity anisotropy at 100km and 200km, constructed through approximate one-dimensional methods, map resistive regions spatially correlated with the Kaapvaal, Zimbabwe and Angola Cratons, and more conductive regions spatially associated with the neighbouring mobile belts and the Rehoboth Terrane. Known diamondiferous kimberlites occur primarily on the boundaries between the resistive or isotropic regions and conductive or anisotropic regions.Comparisons between the resistivity image maps and seismic velocities from models constructed through surface wave and body wave tomography show spatial correlations between high velocity regions that are resistive, and low velocity regions that are conductive. In particular, the electrical resistivity of the sub-continental lithospheric mantle of the Kaapvaal Craton is determined by its bulk parameters, so is controlled by a bulk matrix property, namely temperature, and to a lesser degree by iron content and composition, and is not controlled by contributions from interconnected conducting minor phases, such as graphite, sulphides, iron oxides, hydrous minerals, etc. This makes quantitative correlations between velocity and resistivity valid, and a robust regression between the two gives an approximate relationship of Vs [m/s]=0.045*log(resistivity [ohmm])+4.5.

The Kaapvaal Craton is one of the world’s best natural laboratories for studying the lithospheric mantle given the wealth of xenolith and seismic data that exist for it. The Southern African Magnetotelluric Experiment (SAMTEX) was launched to complement these databases and provide further constraints on physical parameters and conditions by obtaining information about electrical conductivity structures within the lithosphere. Initially, magnetotelluric data acquisition was planned on profiles spatially coincident with the Kaapvaal Seismic Experiment. However with seven more partners joining the original four through the course of the experiment, SAMTEX was enlarged from two to four phases of acquisition, and extended northwards to cover much of Botswana and Namibia. The complete SAMTEX dataset now comprises MT data from over 730 distinct locations in an area of over one million square kilometres, making SAMTEX the largest regional-scale MT experiment conducted to date. Preliminary images of electrical resistivity and electrical resistivity anisotropy at 100 km and 200 km, constructed through approximate one-dimensional methods, map resistive regions spatially correlated with the Kaapvaal, Zimbabwe and Angola Cratons, and more conductive regions spatially associated with the neighbouring mobile belts and the Rehoboth Terrane. Known diamondiferous kimberlites occur primarily on the boundaries between the resistive or isotropic regions and conductive or anisotropic regions.

Electrical anisotropy in southern Africa, inferred from the analysis of magnetotelluric (MT) data recorded as part of the Southern African MT Experiment (SAMTEX), is compared with seismic anisotropy inferred from an SKS shear-wave splitting study in the same region. Given the vastly varying penetration depths in the survey area, electrical anisotropy is derived in terms of approximate depth, rather than frequency. Electrical anisotropy directions for crustal depths (<35km) show more distinct variability than those for upper mantle depths, and, not surprisingly, appear to be strongly related to large-scale geological structures. Our results for upper lithospheric mantle depths (>45km) are not consistent with the fast-axis directions inferred from the SKS analyses. Upper mantle electrical results appear to be mostly a consequence of the geometry of large-scale geological structures and provide evidence that some crustal structures are distinct at depth, while others seem to be confined to the crust. Our results indicate that the causative region for the seismic anisotropy in the lithospheric mantle has either a correspondingly weak electrical anisotropic signature, or is more prominent at greater lithospheric depths than those we investigate here.