Strain transfer between disconnected, propagating rifts in Afar

TitleStrain transfer between disconnected, propagating rifts in Afar
Publication TypeJournal Article
Year of Publication2001
AuthorsManighetti I, Tapponnier P, Courtillot V, Gallet Y, Jacques E, Gillot P-Y
JournalJournal of Geophysical Research-Solid Earth
Volume106
Pagination13613-13665
Date PublishedJul
ISBN Number0148-0227
Accession NumberWOS:000169906100024
Abstract

We showed before that both the Aden and Red Sea plate boundaries are currently rifting and propagating along two distinct paths into Afar through the opening of a series of disconnected, propagating rifts. Here we use new geochronological, tectonic, and paleomagnetic data that we acquired mostly in the southeastern part of Afar to examine the geometry, kinematics, and time-space evolution of faulting related to strain transfer processes. It appears that transfer of strain is accommodated by a bookshelf faulting mechanism wherever rifts or plate boundaries happen to overlap without being connected. This mechanism implies the rotation about a vertical axis of small rigid blocks along rift-parallel faults that are shown to slip with a left-lateral component, which is as important as their normal component of slip (rates of similar to2-3 mm/yr). By contrast, where rifts do not overlap, either a classic transform fault (Maskali) or an oblique transfer zone (Mak'arrasou) kinematically connects them. The length of the Aden-Red Sea overlap has increased in the last similar to0.9 Myr, as the Aden plate boundary propagated northward into Afar. As a consequence, the first-order blocks that we identify within the overlap did not all rotate during the same time-span nor by the same amounts. Similarly, the major faults that bound them did not necessarily initiate and grow as their neighboring faults did. Despite these variations in strain distribution and kinematics, the overlap kept accommodating a constant amount of strain (7 to 15% of the extension amount imposed by plate driving forces), which remained distributed on a limited number (seven or eight) of major faults, each one having slipped at constant rates (similar to3 and 2 mm/yr for vertical and lateral rates, respectively). The fault propagation rates and the block rotation rates that we either measure or deduce are so fast (30-130 mm/yr and 15-38 degrees /Myr, respectively) that they imply that strain transfer processes are transient, as has been shown to be the case for the processes of tearing, rift propagation, and strain jumps in Afar.