Slip localization on multiple fault splays accommodating distributed deformation across normal fault complexities.

Features such as fault geometry and slip-rates are key inputs to assess the seismic hazard imposed by either ground motion or fault displacement. However, complexities in the geology of faults, such as relay zones and along-strike fault bends, could lead to settings characterized by high segmentation, with multiple splays arranged both along and across strike. In order to assess the seismic hazard associated with such fault sectors, it is necessary to establish whether the 3D shallow deformation is equally spread over the multiple fault splays or the activity tends to localise on specific splays. This problem is enhanced when these faults are located within urban areas, and therefore their surface expression is altered by intense anthropic activity.

Within the framework of a work on the mitigation of the fault displacement hazard associated with the Mt. Marine active normal fault (Central Italy), we have performed two paleoseismological surveys within the town of Pizzoli (about 10 km NW of L’Aquila), where the fault is expressed with several splays arranged both along and across-strike. The trenches were planned to explore (i) potential fault scarps altered by human activity, identified through aerial photographs, LiDAR and fieldwork analysis, and (ii) discontinuities in the stratigraphic record highlighted by geophysical investigations (ERT, GPR) and borehole data.

The paleoseismological surveys intercepted five fault splays arranged across-strike, three synthetic and two antithetic to the main Mt. Marine fault. The fault splays show evidence of multiple Late Pleistocene/Holocene surface-rupturing seismic events, marked by colluvial wedges and infilled fractures. Moreover, we constrained the Late Pleistocene slip-rate of the Mt. Marine fault splays by dating and correlating Late-Pleistocene paleosols found (1) outcropping in the footwall of one of the inner fault splay and (2) in a borehole located just at the hangingwall of the outermost splay.

Our results show that the fault splays exhibit different and variable activity rates, suggesting that fault activity is localized on specific fault splays through space and time with the potential to rupture simultaneously during large earthquakes. Our findings have strong implications on fault-based seismic hazard assessments, as they imply that data collected on one splay may not be representative of the behaviour of the entire fault.