Self-similarity of the largest-scale segmentation of the faults: Implications for earthquake behavior

article i nfo Article history: Earthquakes are sensitive to the along-strike segmentation of the faults they break, especially in their initiation, propagation and arrest. We examine that segmentation and search whether it shows any specific properties. We focus on the largest-scale fault segmentation which controls the largest earthquakes. It is well established that major segments within faults markedly shape their surface cumulative slip-length profiles; segments appear as large slip bumps separated by narrow, pronounced slip troughs (inter-segments). We use that property to examine the distribution (location, number, length) of the major segments in 927 active normal faults in Afar (East Africa) of various lengths (0.3-65 km), cumulative slips (1-1300 m), slip rates (0.5-5 mm/yr), and ages (10 4 -10 6 yr). This is the largest fault population ever analyzed. To identify the major bumps in the slip profiles and determine their number, location and length, we analyze the profiles using both the classical Fourier transform and a space-frequency representation of the profiles, the S-transform, which is well adapted for characterizing local spectral properties. Our work reveals the fol- lowing results: irrespective of their length, 70% of the slip profiles have a triangular envelope shape, in conflict with the elastic crack concept. Irrespective of their length, the majority of the faults (at least 50-70%) have a limited number of major segments, between 2 and 5 and more commonly equal to 3-5. The largest- scale segmentation of the faults is thus self-similar and likely to be controlled by the fault mechanics. The slip deficits at the major inter-segment slip troughs tend to smooth as the faults accumulate more slip resulting in increased connection of the major segments. The faults having accumulated more slip therefore generally appear as un-segmented (10-30%). Our observations therefore show that, whatever the fault on which they initiate, large earthquakes face the same number of major segments to potentially break. The number of segments that they eventually break seems to depend on the slip history (structural maturity) of the fault.