GNGTS 2019 - Atti del 38° Convegno Nazionale

GNGTS 2019 S essione 1.1 87 offsets across post-Last Glacial Maximum (LGM) scarps, are distributed across the entire width of the orogen, including earthquakes in 1349 AD and earlier times that were clearly located on the SW flank of the orogen causing severe damage in Rome (Cowie et al. , 2017; Beck et al. , 2018). However, at the time of writing there are no direct observations of the switching activity between faults arranged across strike, on a timescale less than the age of the LGM. The fault system in Attica, central Greece, provides an ideal opportunity to study the temporal relationships between activity on parallel sets of faults, over the required timescale, in order to provide observations of the characteristics of across-strike switching of activity. The Attica region contains sets of parallel active faults with across-strike fault spacing larger than 5 km (between 5-16 km). These faults have already revealed themselves as highly dangerous, for example during the 7 th September 1999 Mw 6.0 earthquake, which caused severe damage in the northwest part of Athens and 143 deaths due to building collapse (Pavlides et al. , 2002; Ganas et al. , 2004). Thus, studying the evolution of the slip histories for faults in Attica could provide key information on the dynamics of fault systems with parallel faults, such as the Central Apennines, and on the mitigation of the seismic hazard of the city of Athens. To do so, we present measurements of the exposure ages of three faults arranged parallel across the strike in the Attica fault system using in-situ 36 Cl cosmogenic exposure dating of fault planes tectonically-exhumed since the demise of the LGM. Detailed site characterization helped to ensure that the fault surface is exposed only through tectonic exhumation during repeated earthquakes, and it is not disturbed by geomorphological processes. The concentrations of cosmogenic 36 Cl have been modelled using a Bayesian reversible-jump Markov chain Monte Carlo approach, which iterates the slip history many tens of thousands of times, forward modelling expected 36 Cl concentrations each time, to search for the best-fit to measured 36 Cl concentrations. This allows reconstruction of the slip histories of faults to define the Holocene activity rates by constraining (1) the slip-rate on active faults, (2) the timing of the last earthquake (Telap), (3) the mean recurrence interval for maximum magnitude earthquakes (Tmean) and (4) the variability in recurrence (CV), that is the standard deviation of recurrence intervals divided by the mean recurrence interval (Pace et al. , 2016). Our modelling shows that faults have episodic behaviour, with a non-systematic alternance of periods of rapid slip accumulation (i.e. earthquake clustering) and periods of quiescence (i.e. earthquake anti-clustering). The comparison between the modelled slip histories on the different faults shows that the activity switches between faults, with earthquake clusters alternating on the different faults: earthquake clustering on one specific fault is accompanied by low rates of activity on the other faults. These results show that parallel faults spaced > 5 km interact in terms of sharing the regional strain-rate, with switching activity between faults that affects the slip-rate on a single fault. The clustered behaviour of parallel faults implies that the definition of key parameters for seismic hazard assessments (Tmean, Telap and CV) is complicated in fault systems with across-strike distributed faults. Moreover, the retrieved slip histories also highlight a difference in the ages since the three fault scarps started to be preserved. The modelling performed in Attica, located at lower latitudes compared to Central Italy, shows that for these faults exist a correlation between the age since the fault scarp started to be preserved and the elevation of the fault scarp: the lower the elevation of the fault scarp, the older is the age since the scarp started to be preserved. This implies that, at these latitudes, the elevation above the sea level seems to play a role in the dynamics of the erosion processes associated with the LGM (e.g. freeze-thaw erosion and frost-shattering). Overall, the measurements and modelling of cosmogenic 36 Cl concentrations on fault planes on three faults arranged across the strike of a fault system in Central Greece provides fundamental insights on the dynamics of normal fault systems, and on the interplay between the LGM-related erosion processes and elevation on the preservation of fault scarps. These results also imply important insights for the dynamics of faults in Central Italy, where multiple faults are arranged parallel across the strike of the fault system. We suggest that parallel faults within