Lithospheric "corner flow" via extensional faulting and tectonic rotation at non-volcanic, slow-spreading ridges

Large degrees (up to 90°) of tectonic rotation may be the norm at slow-spreading, non-volcanic ridges. Vertically upwelling mantle beneath all mid-ocean ridges must undergo corner flow to move horizontally with the spreading plate. Because little or no volcanic crust is produced at some slow-spreading ridges, the uppermost lithospheric mantle must undergo this rotation in the regime of localized, rather than distributed deformation. Anomalous paleomagnetic inclinations in peridotite and gabbro cores drilled near the 15-20 Fracture Zone (Mid-Atlantic Ridge, ODP Leg 209) support such large rotations, with sub-Curie-temperature rotations up to 90° (Garces et al., 2004). Here, we present two end-member tectonic mechanisms, with supporting data from Leg 209 cores and bathymetry, to show how rotation is accomplished via extensional faults and shear zones: 1) long-lived detachment faults, and 2) multiple generations of high-angle normal faults. Detachment faults accommodate rotation by having a moderate to steep dip at depth, and rotating to horizontal through a rolling hinge as the footwall is tectonically denuded. Multiple generations of high-angle normal faults accommodate large rotations in a domino fashion; early faults become inactive when rotated to inopportune slip angles, and are cut by younger high-angle faults. Thus, each generation of high-angle faults accommodates part of the total rotation. There is likely a gradation between the domino and detachment mechanisms; transition from domino to detachment faulting occurs when a single domino fault remains active at inopportune slip angles and evolves into a detachment that accommodates all corner flow for that region. In both cases, the original attitude of layering within mantle-emplaced gabbro bodies must be significantly different than present day observed attitudes; sub-horizontal bodies may have been formed sub-vertically and vice-versa. Leg 209 cores record an average major brittle fault spacing of approximately 100 m, suggesting that the width of individual rotating fault blocks may be on the order of 100-200 m. Numerous fault bounded domino slices could therefore be formed within a 10km wide axial valley, with large rotations (and commensurate extension) leading to the exposure of 1km wide shallow-dipping fault surfaces, as are seen in the 15-20 FZ region bathymetry. The region's bathymetry is dominated by irregular, low-relief ridges that were likely formed by domino faulting of lithosphere with a small elastic thickness. The region contains relatively few corrugated detachment fault domes, suggesting that domino faulting may be the normal mode of lithospheric corner flow at non-volcanic ridges.