High Resolution Quaternary and Neogene Reconstructions of the Southwest Indian Ridge and Rifting in Eastern Africa

Abstract:

The Southwest Indian Ridge (SWIR) south of Africa is a critical link in plate circuits between the Atlantic and Indian Ocean basins and between the Nubia and Somalia plates. Detailed reconstructions of its seafloor spreading history are challenging due to the low fidelity of its magnetic anomalies, which were mostly created at slow to ultraslow spreading rates, and gaps in data coverage for some areas of the ridge. Here, we describe the first high-resolution analysis of Quaternary/Neogene SWIR plate kinematics based on nearly 5000 identifications that we made of magnetic reversals C1n (0.78 Ma) to C6no (19.7 Ma) and ~6000 crossings of 21 fracture zones and transform faults that offset the ridge. We also outline the implications for estimates of motion between the Nubia and Somalia plates since 20 Ma across rifts in eastern Africa. Searches for the Nubia-Lwandle and Lwandle-Somalia plate boundaries north of the SWIR with our new data corroborate previous evidence for respective locations near the Andrew Bain transform fault at ~30°E and at ~50°E. Inversions of the abundant new data to find best-fitting rotations at ~1 Myr intervals since 20 Ma reveal a previously unknown, ~20% deceleration of seafloor spreading rates at 7.2±1 Ma everywhere along the SWIR. Motion since 7 Ma has remained remarkably steady and agrees within uncertainties with GPS estimates that are based on more than 100 continuous GPS sites on the Nubia, Somalia, and Antarctic plates. The consistency of the geodetic and geologic estimates validates both and also supports evidence we will describe for anomalously wide outward displacement west of ~30E. Nubia-Somalia rotations determined from our new model indicate that the two plates have undergone steady relative motion since at least 19 Ma. Our new rotation for C5n.2 predicts ~70% less opening across the East Africa rift since 11 Ma than the most recently published kinematic estimate, in better accord with at least one geologically-derived estimate for the total extension.