High resolution ambient noise tomography of the Southwestern Alps and the Ligurian margin. Issue 2 (21st October 2019)

• Record Type: Journal Article
• Authors: Guerin, Gauthier; Rivet, Diane; Deschamps, Anne; Larroque, Christophe; Mordret, Aurélien; Dessa, Jean-Xavier; Martin, Xavier
• Is Part Of: Geophysical journal international. Volume 220:Issue 2(2020)
• Page Start: 806
• Page End: 820
• Publication Date: 2019-10-21
• Subjects: Europe -- interferometry -- Crustal imaging -- Seismic noise -- Seismic tomography; Geophysics -- Periodicals; 550
• DOI: 10.1093/gji/ggz477
• Languages: English
• ISSNs: 0956-540X

Abstract

SUMMARY: The Southwestern Alps and the Ligurian margin is a region of moderate seismicity with a high rate of small to moderate events. Identifying the active faults in this very densely populated region is critical to better assess the hazard and mitigate the risk. An accurate 3-D velocity model of the shallow to middle crust is a fundamental step to better locate the seismicity, and hence, the faults from which it originates. We performed ambient noise surface-wave tomography based on all available continuous seismological data from the French and Italian permanent networks (RESIF, INGV, RSNI), and current and past temporary experiments (AlpArray, CASSAT, SISVAR, RISVAL). In addition to these available data, we deployed three more stations to improve the spatial resolution in a region with sparse seismic station coverage. Overall, we used 55 inland seismic stations, 5 oceans bottom seismometers and 2 offshore cabled site/sensors. Data span the 2014–2018 time period. Time series from all available components were cross-correlated to reconstruct both Rayleigh and Love-wave Green's functions. For each station-pair Rayleigh and Love group velocity dispersion curves were semi-automatically picked using a frequency–time analysis. Then we regionalize these group velocities to build 2-D Rayleigh and Love velocity-maps between 1.5 and 9 s period. Using a two-step inversion, we estimate the best 3-D shear wave velocity model. The first step is based on a Neighbourhood Algorithm to recover the best three layers' velocity model at each cell of the model. We then use this three-layer model as a starting model in a perturbational method based on finite elements. At periods up to 5 s, the spatial variation of the velocity is well correlated with the effective geology of the area. Lower velocities are observed in areas where the sedimentary cover is thicker, such as the Var and Paillon valley near Nice, or in the subalpine domain in the northwestern part of the region. Higher velocities are retrieved in areas where massifs are present, such as the Argentera-Mercantour massifs in the northeastern, or the Esterel massif in the southwestern part of the region.