GNGTS 2023 - Atti del 41° Convegno Nazionale

Session 2.1 GNGTS 2023 New approach to constrain objective earthquake chronologies from paleoseismic fault records O. Gómez-Novell 1 , B. Pace 1 , F. Visini 2 1 Dipartimento INGEO, Università degli Studi “Gabriele d’Annunzio” di Chieti e Pescara, Chieti, Italy 2 Istituto Nazionale di Geofisica e Vulcanologia (INGV), Chieti, Italy The inclusion of fault data into seismic hazard assessments has proven to perform more realistic estimates of the hazard values, especially of the expected ground motion levels and its spatial-temporal distribution. Consequently, such practice has become increasingly popular over the last decades as the knowledge on fault activity has significantly improved. In this respect, paleoseismology is one of the most frequently used techniques as it allows to extend the earthquake catalog beyond the historical period, which is crucial in regions where the historical records are shorter than the average fault recurrences. However, paleoseismic research usually has to deal with important limitations that hamper a correct identification of the paleoearthquake history and especially their constraining in time. Among others, these concern limitations in dating the trench stratigraphic sequences (e.g., lack of suitable material or age beyond the technique resolution), the lack of continuous sedimentary records due to intermittency of the deposition, and the variability of fault slip along strike. We propose an approach to compute and constrain earthquake chronologies in faults by correlating paleoseismic trench sequences from multiple sites, especially when these are complex or poorly constrained in time. The underlying assumption is that the integration of multi-site data can allow to improve the time constraints of events. However, because paleoseismic records are usually underestimated, event correlation between sites is unrestricted, meaning that an event in one site can participate in more than one event correlation in another site if these are time compatible. The approach is designed to be as objective as possible and, for this reason, it solely relies on the unmodelled trench numerical dates limiting the event horizons as the inputs. The algorithm of the approach uses these inputs to compute the earthquake chronologies following a probabilistic workflow as we detail. First, event occurrences in each trench are modelled as probability density functions (PDFs) by randomly sampling time values of numerical date pairs. Then, all event PDFs from all sites are averaged in a mean curve representing the overall seismic occurrences in the studied fault for the period investigated. The peaks in this curve are considered as indicative of event occurrences and are therefore automatically detected by the algorithm. Lastly, the event PDFs intersecting each peak are extracted and multiplied into a product PDF per