Evidence for surface faulting events along the Paganica fault prior to the 6 April 2009 L'Aquila earthquake (central Italy)

We performed paleoseismological investigations at four sites across the normal Paganica fault (PF) (source of the 2009 Mw 6.3 L'Aquila earthquake), with the goal of reconstructing the rupture history and of contributing to the evaluation of the maximum event expected along the PF. We recognized five distinct surface faulting earthquakes (including the 2009) in the trenches. The age of the penultimate event is consistent with the 1461 earthquake; the third event back occurred around 1000 AD. The two oldest events have larger uncertainties and occurred in the interval 760 BC-670 AD and 2900-760 BC, respectively. The along-strike vertical displacement for each paleoevent has a limited variability consistently with the fairly homogeneous slip observed in 2009 along the northern part of the rupture. Conversely, the throws change between distinct events and range between 0.15 m in 2009 (maximum estimate) and close to 0.4 (lower bound estimate) in earlier events. These paleorecords and the high fault escarpments imply that earthquakes larger than 2009 occurred on the PF, with implications for the level of hazard. Recurrence intervals also reflect a change with time, the average interval before similar to 1000 AD is longer compared to that after this date. Two events occurred in the 2000-4000 years preceding similar to 1000 AD, while three events occurred since similar to 1000 AD. The age uncertainties affecting the interpreted events prevent the evaluation of a unique value for interevent interval; the older events appear closely spaced in time or far apart depending on the upper or lower boundary of the age interval. We tentatively assign an average interevent time of similar to 500 years for the three youngest events, whereas the time elapsed between the previous ones could be larger, in the order of 1000-2000 years. We calculate a late Pleistocene dip-slip rate for the PF of 0.2-0.4 mm/yr, consistent with 0.25-0.5 mm/yr for the early Pleistocene. Using age and throw of individual events, we calculate a similar late Holocene average dip-slip rate of similar to 0.3-0.4 mm/yr. This suggests that the portion of the PF where the 2009 continuous surface faulting occurred has fairly a constant average slip release since late Pleistocene. Finally, we discuss different rupture scenarios and alternative models of occurrence compatible with our data and their variability.