The rate of oceanic detachment faulting at Atlantis Bank, SW Indian Ridge

The rates of slip on oceanic detachment faults and how those rates compare to sea-floor spreading rates constitute fundamental data required to constrain how oceanic core-complexes form and their role during crustal accretion. We combine sea-surface magnetic data, with the magnetic polarity of shallow-core samples and Pb/U SHRIMP ages of igneous zircon to determine the time-averaged half-spreading rate during oceanic detachment faulting at Atlantis Bank, 100 km south of the ultraslow-spreading Southwest Indian Ridge (SWIR). The Pb/U zircon ages correlate well with the magnetic ages and so highlight that magmatic accretion and faulting were coeval for over 2 Myr, creating and exposing a > 1.5-km-thick layer of gabbro for > 35 km parallel-to-spreading. We use bivariate linear regression of distance-age data and forward modeling of magnetic anomaly data to calculate a half-spreading rate during detachment faulting of 14.1 + 1.8/− 1.5 km/Myr (95% confidence limits). When integrated with regional constraints on spreading history, we note that detachment faulting coincided with a short-lived regional increase in the full-spreading rate along the SWIR and, for the ridge segment containing Atlantis Bank, spreading was highly asymmetric with ~ 80% of plate-motion accommodated by detachment faulting. Consequently, the detachment fault effectively formed the plate-boundary at the surface in this spreading segment. Highly asymmetric spreading was confined to the spreading segment containing Atlantis Bank and to the duration of detachment faulting. So the ridge segment containing Atlantis Bank migrated northward relative to its symmetrically spreading eastern neighbour, such that the intervening non-transform discontinuity shortened. We suggest that the highly asymmetric spreading may be a characteristic feature of oceanic detachment faulting, an inference supported by more poorly constrained half-spreading rates determined at several other oceanic core-complexes.