ALBERT

Description: Seismic hazard assessment is a critical but challenging issue for modern societies. A key parameter to be estimated is the recurrence interval of damaging earthquakes. This requires the establishment of earthquake records long enough to be relevant, i . e . far longer than historical observations. We study how lake sediments can be used for this purpose and explore conditions that enable lake sediments to record earthquakes. This was achieved (i) through the compilation of eight lake-sediment sequences from the European Alps to reconstruct chronicles of mass-movement deposits and (ii) through the comparison of these chronicles with the well-documented earthquake history. This allowed 24 occurrences of mass movements to be identified, of which 21 were most probably triggered by an earthquake. However, the number of earthquake-induced deposits varies between lakes of a same region, suggesting variable thresholds of the lake sequences to record earthquake shaking. These thresholds have been quantified by linking the mass-movement occurrences in a single lake to both intensity and distance of the triggering earthquakes. This method offers a quantitative approach to estimate locations and intensities of past earthquake epicenters. Finally, we explored which lake characteristics could explain the various sensitivities. Our results suggest that sedimentation rate should be larger than 0.5 mm.yr −1 so that a given lake records earthquakes in moderately active seismo-tectonic regions. We also postulate that an increasing sedimentation rate may imply an increasing sensitivity to earthquake shaking. Hence, further paleoseismological studies should control carefully that no significant change in sedimentation rates occurs within a record, which could falsify the assessment of earthquake recurrence-intervals.