Local seismic hazard assessment in explosive volcanic settings by 1d and 2d numerical analyses

In this study, we face the problem of local seismic response in explosive volcanic setting by using an integrated geological-geophysical-geotechnical approach in the test area of Stracciacappa maar (Sabatini Volcanic District, central Italy). Our aim is to understand if the horizontal and vertical chaotic heterogeneity typical of the volcanic deposits influence site response. The Stracciacappa maar is an active eruptive centre characterised by a crater of about 1 km in diameter and a crater floor of about 30-40 m below the rim (De Rita et al. 1983; Sottili et al. 2012). The ring is mainly composed by the pyroclastic succession belonging to the last phreatomagmatic activity. This pyroclastic succession generally bends outward the rim with low angle dip (10-20°). It consists of at least 25 metres thick alternation of decimetre- to centimetre-thick layers of fine-medium ash and small lapilli. The crater depression is filled by epiclastic debris deposits and by recent and present-day lacustrine muds. The epiclastic debris deposits, of alluvial and delta origin, consist of alternation of cm-thick reworked fine-grained and coarse-grained volcaniclastic material, dipping with low-angle (1s) motions, compared to tectonic events of equivalent magnitude (Jousset and Douglas 2007); the unscaled recording at Bronte Station (BNT in ITACA database, http://itaca.mi.ingv.it) of the ML=4.4 October 27, 2002 event was employed; 2) a high magnitude far-field “tectonic” event (tectonic scenario), whose reference spectrum was built with Ground Motion Prediction Equations (Ambraseys et al., 2005) assuming M=6.5 and distance of 70 km. These conditions are compatible with seismogenic sources located in central Apennines of Italy. Three unscaled recordings of events characterised by magnitude and distance in the range of 6-7 and 60-90 km, respectively, were extracted from ITACA database (http://itaca.mi.ingv.it), matching on average the reference spectrum Two subsoil models have been considered: a detailed model (based on distribution of the lithotypes unravelled by the geological survey) and a simplified one (obtained by grouping interfingering lithotypes resting below the lacustrine silty clays). The result show that the two models have similar response in all range of the interesting period (0.1-1.0s); the damping properties of soft clays and sands deposits in the upper meters reduce the difference in the seismic response at the surface of both models. The results suggest the possibility to simplify the heterogeneous distribution of deposits in this volcanic context for assessment of seismic response purposes. Finally, we carried out both linear and equivalent linear analyses in one-dimensional and bi-dimensional conditions, in order to investigate the bidimensional effects and the role of nonlinearity on the seismic response,. In linear case the behaviour of soils was assumed linear visco-elastic with small strain damping ratio values D0; amplification factors higher than 10 were reached at 2 Hz at soft clays surface in correspondence of the centre and western edge of the maar in bi-dimensional analysis, whereas the 2D/1D ratios were in the order of 2-3 around 2 Hz with maximum values at the maar edges. In the nonlinear analyses, the maximum amplifications dropped below 10 and the 2D effects (i.e., 2D/1D ratios) were generally lower than 2 in the whole range of frequency.