Characterizing ocean-bottom DAS data for Ambient Noise Interferometry offshore Catania, Sicily

The North Alfeo Fault System is a major strike-slip structure in the Ionian Sea offshore Catania, Sicily. An important element of earthquake rupture generated on such faults is understanding the role of damage zones. To achieve this, accurate high resolution subsurface imaging of such structures is essential. Ambient noise interferometry has become well established for land-based seismic arrays and more recently showing potential in reaching higher resolutions using distributed acoustic sensing (DAS) data. DAS applied on submarine infrastructures such as existing telecommunication cables appears promising for using such techniques, as it is highly sensitive to Scholte waves that propagate along the seafloor interface.

Offshore Catania, the North Alfeo fault is crossed five times by fibre optic cables, once by the 29km INFN-LNS MEOC cable and four times by the 6km FOCUS cable, a branch extension of the former. We investigated DAS strain-rate data recorded at metric channel spacing, analyzing both cable segments. In this setting, the ultimate goal is a surface wave tomography of the fault zone to constrain shear-wave velocity structure and damage zone geometry. We therefore aim to address several open methodological challenges using ocean-bottom DAS data for the reliable retrieval of Green’s functions from ambient seismic noise cross-correlations.

First, spectral and statistical characterization are conducted to identify dominant signal types including sea gravity waves, Scholte waves, earthquake signals, volcanic tremor, and boat signals, along with their frequency content. Although signal quality variations along the cable may arise from several factors, we focus on suspended cable sections, which exhibit high spatial signal similarity due to coherent ringing effects. Accordingly, a Pearson correlation matrix computed over short time windows (e.g., 1hr and 15mins) reveals that 11.8% and 4.6% of the FOCUS and MEOC cables respectively, are suspended based on ambient noise signals, with slightly higher but spatially consistent estimates from earthquake signals.  These suspended sections are primarily located on the continental slope and fault crossings in case of the FOCUS cable. A standard short-term average/long-term average (STA/LTA) method is then applied to detect all non-stationary events, such as earthquakes, whose frequency band overlaps with that of ambient seismic noise. As an initial step, the analysis is conducted for a single day (15 November, 2023) where up to 13 more events are identified that are not reported in terrestrial catalogues. These time windows are subsequently removed to retain only stationary ambient noise for cross-correlation analysis. Finally, cross-correlation using a reference channel is performed to obtain a baseline understanding of the wavefield structure. The resulting cross-correlation functions are analyzed to assess signal quality and their temporal evolution.