DETAILED FAULT SLIP-HISTORIES BASED ON COSMOGENIC 36CL ANALYSES FROM ABRUZZO, ITALY REVEAL FAULT BEHAVIOUR OVER MULTIPLE EARTHQUAKE CYCLES

Analysis of variations in cosmogenic nuclide concentration along active bedrock fault scarps provides unique insight into fault behaviour over several thousand years and multiple earthquake cycles. Variations in slip-rate and fault interaction on hundred- to thousand- year time scales across networks of faults may be inferred based on datasets of ³⁶Cl concentration measured systematically up the scarps, parallel to the slip vector. In this presentation, we will use a robust and extensive ³⁶Cl dataset to characterise Holocene activity across a complicated network of normal faults in Abruzzo, Italy.

NE-SW directed extension in the Italian Apennines since 2-3 Ma is localised on NW-SE trending normal faults. Elevated topography in the region is likely supported dynamically by mantle convection, and buoyancy forces are the dominant control on regional extension. Faulting since the last glacial maximum (LGM) has displaced slopes preserved since the LGM, creating scarps of exposed bedrock limestone that have throws of several metres. ³⁶Cl is produced in-situ at the Earth’s surface and down to a few metres depth, primarily due to interactions between calcium and high energy cosmogenic radiation. ³⁶Cl accumulates in bedrock fault scarps as the plane is progressively exhumed by earthquakes. The geomorphology of sampling sites must be carefully constrained to ensure that exposure of the fault is only due to seismic activity and not the result of mass transport or erosional processes. We quantify the geomorphology of each site using LiDAR (terrestrial and airborne) and ground penetrating radar (GPR).

We use the ³⁶Cl data to compare the most recent fault slip history (over a few hundred to a few thousand years) with the Holocene averaged slip rate along individual fault strands. For some fault strands, these two estimates of slip rate measured over different time windows are similar, whilst for other strands, they differ. This suggests that slip may be transient across a network a faults over multiple earthquake cycles, and recent seismicity on any one fault strand may be variable compared to long-term estimates. This study highlights the need to characterise and compare fault behaviour over multiple timescales in order to anticipate future seismicity.