How do faults grow in magmatic rifts? LiDAR and InSAR observations of the Dabbahu rift segment, Afar, Ethiopia

Dyke intrusions and normal faulting play an important role during continental break-up but little is known about how the normal faults develop. Direct evidence of dyke-induced faulting is limited by the lengthy repeat times between individual rifting episodes, the small amount of subaerial rift zones and until recently the technical ability to record small surface changes across large areas. The most recent (2005-2010) rifting episode at the Dabbahu rift segment, Afar, Ethiopia provided a unique opportunity to study dyke-induced fault growth. The combination of new high-resolution topographic LiDAR data and interferometric synthetic aperture radar (InSAR) data provides information of cumulative as well as incremental fault throw. In this thesis I use high-resolution LiDAR data of the Dabbahu rift segment to reveal a dense network of short fault segments (>3400) at various stages of fault linkage set in flood basalt plains. I develop and present a semi-automatic algorithm that extracts throw along surface fault traces from the high-resolution LiDAR DEM. The largest amount of throw (~80 m) is found on faults towards the east of the rift segment. At the central Ado’Ale volcanic edifice predominant bookshelf faulting is evident which might be an indication of a lateral shift of the dykes towards the east. I use the throw data to derive a strain field for the rift. Faults record ~140 m of extension, implying extensive resurfacing. I derived displacement data from two LiDAR surveys and InSAR data, for two separate dyke intrusions. Both data sets show that faults are re-activated in a broad, 3-4 km wide, asymmetric zone parallel to the dyke induced subsidence with the majority of the new throw being accumulated on 1-2 large west-dipping fault structures in the east. The incremental displacement-length, d − L, data presented here is the first quantitative study of accumulation of new fault throw across an entire rift segment. Incremental throw across linkage zones suggest two types of behaviour once fault linkage is complete. 1) Individual fault segments maintain the ability to slip independently. This was previously only observed during analogue modelling. 2) The connected faults act as one throughgoing fault with slip unaffected by the linkage zone. The combination of these two processes might be responsible for the commonly observed small-scale corrugation in d − L data. In contrast to published fault growth models, I present evidence that the remnant fault tip of a linkage zone does not necessarily become inactive once linkage is complete, and that linkage zones do not ‘catch up’ through accelerated throw once linkage is complete.