Probing the Southern African Lithosphere With Magnetotellurics—Part I: Model Construction.

The Southern African Magnetotelluric Experiment (SAMTEX) involved the collection of data at over 700 sites in Archean to Proterozoic southern Africa, spanning features including the Kalahari Craton, Bushveld Complex, and voluminous kimberlites. Here, we present the first 3D inversions of the full SAMTEX data set. In this paper, we focus on assessing the robustness of the 3D models by comparing two different inversion codes, jif3D and ModEM, and two different subsets of the data, one containing all acceptable data and the other containing a smaller selection of undistorted, high‐quality data. Results show that the main conductive and resistive features are imaged by all inversions, including deep resistive features in the central Kaapvaal Craton and southern Congo Craton and a lithospheric‐scale conductor beneath the Bushveld Complex. Despite this, differences exist between the jif3D and ModEM inverse models that derive mainly from the differences in regularization between the models, with jif3D producing models that are very smooth laterally and with depth, while ModEM produces models with more discrete conductive and resistive features. Analysis of the differences between these two inversions can provide a good indication of the model resolution. More minor differences are apparent between models run with different subsets of data, with the models containing all acceptable data featuring higher wavelength conductivity variations than those run with fewer stations but also demonstrating poorer data fit. Plain Language Summary: We investigate the structure of the upper 200 km of the Earth beneath southern Africa. To achieve this, we utilize an electromagnetic geophysical technique called magnetotellurics (MT) which is sensitive to variations in electrical resistivity within the Earth. To reconstruct electrical resistivity from MT measurements, we use so‐called inversion algorithms. However, the results are nonunique and a variety of different parameters have to be chosen by the user during the inversion process. In order to better understand the possible variability in our Earth models, we use different inversion algorithms and compare different strategies. This allows us to assess the reliability of our results. Based on our models and their comparison, we infer that the lithosphere, the solid outer shell of the Earth, varies in thickness below our study area and is thickest below central South Africa. In addition, we can detect remnants of past continental collisions that have been preserved for hundreds of millions of years since this part of the world was assembled from the collision of various microcontinents. Key Points: We create the first large‐scale conductivity models of southern AfricaWe compare different strategies to construct continental scale models and investigate the impact on the resultsThe main inversion features are common to all models, with most differences being due to regularization [ABSTRACT FROM AUTHOR]