High resolution seismic data, we collected in the Ionian sea, reveal large submarine landslide deposits offshore from Mt. Etna (Italy), spatially consistent with the eastern flank collapse of this volcano. A large debris-avalanche deposit, we relate to the Valle del Bove scar, displays long offshore run-outs (till 20 km) and a volume of a few tens of cubic kilometres (16–21 km3). Other landslide deposits are also imaged, in particular a striking unique record of the relative timing of multiple submarine large slump events.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 In May 2005, we collected ∼480 km of high resolution 2D multichannel seismic profiles in the Ionian Sea, offshore from Mt. Etna. The “Istituto Nazionale di Geofisica e Vulcanologia” contracted with FUGRO OCEANSISMICA S.p.A. to conduct the survey. A subset of these profiles (about 360 km) is shown in Figure 1a. The seismic streamer utilized was a Litton Industries quick coupler with 96 traces and a group interval of 12.5 m, for a total active streamer length of 1200 m. The streamer was connected to a TTS-II digital data recording system with a sampling interval of 1 ms and a record length of 3 s. The near offset trace was recorded 72 m from the vessel common reference point, with an offset of 51.2 m between the source and this trace. The raw seismic data were recorded in SEG-D format on type 3490 tape cartridges. The seismic source was an airgun array of total size of 120 inch3 operated at a pressure of 2,000 psi. The shooting interval was 12.5 m with time accuracy of ± 0.05 ms. The source and streamer were towed approximately 2.5 m below sea level. The major steps of the processing sequence are reported in an auxiliary file. The P wave interval velocity along profiles was estimated by applying the Dix equation, with the RMS stacking velocity and ranges from 1,700 to 2,200 ms−1 as input parameters (see auxiliary file).
 The inland digital elevation map (DEM) of Mt. Etna, with a 10 m contour interval, was constructed from input data (contour line and spot heights) derived from 1998 stereophotograms at a 1:35,000 scale [Favalli and Pareschi, 2004]. The sea-bottom DEM comes from a bathymetric survey completed in 1996–1997 by the Istituto Idrografico della Marina IIM (the Italian Navy). The original bathymetric digital data were kindly provided to Istituto Nazionale di Geofisica e Vulcanologia by IIM (Genoa-Italy). The IIM surveys were performed using Monobeam ATLAS KRUPP DESO 20 and DESO 25 (IIM) echosounders along parallel routes as perpendicular as possible to the coastlines, with line spacing varying from 250 m to 25 m according to distance from the coast. A Multibeam ELAC Echosounder, with 100% coverage of the sea bottom, was employed by IIM beyond the 500 m bathymetric line. All data are expressed in WGS84-UTM coordinates.
3. Data Description
 From monobeam data collected by IIM, we observe an offshore bulging area, characterized by an hummocky morphology, long promontories and scars (Figures 1a and 1b). To the north, the bulge is confined by a pronounced, slightly bent prominent ridge (R in Figure 1a), a few hundred metres high, 20 km long. To the south, it is loosely confined by a narrow, well marked E-W sea canyon (The Catania Canyon in Figures 1a and 1c), occurring along the offshore extension of the InSAR interferometric features IF1 and IF2 that accomplished the eastward volcano spreading [Froger et al., 2001; Lundgren et al., 2004].
 South of the Catania Canyon, a marine terrace (MT in Figure 1c) extends offshore for about 6 km with an average slope of about 1°; the terrace is eastward bordered by the Malta Escarpment. Topographic profiles across the bulging zone show irregular slopes with numerous humps and convex features (Figure 1b). In contrast, profiles across the Malta Escarpment reveal a uniform and gentle slope (6–7°) dissected by unevenly spaced channels down to a depth of 2,000 m (Figure 1b).
 In the bulging area, our seismic survey images several seismic facies that are very similar to facies interpreted as landslide deposits imaged offshore from the Hawaiian volcanoes [Leslie et al., 2002] and in the Middle American trench [von Huene et al., 2004], characterized by highly chaotic/incoherent signals and hummocky sequences of reflectors (Figures 2a–2d). The used term landslide deposits refers to submarine mass-wasting deposit including both debris avalanche and slump deposits [e.g., Moore et al., 1989].
 The dominant frequency of the reflected signal throughout sections was about 90 Hz. In this acoustic framework the vertical resolution limit is 5–7 m, while the horizontal limit is about 200 m at 2.0 s two way travel time (TWTT) [Yilmaz, 1987].
 Four main seismic facies have been recognized within the landslide deposits. The LU-1 facies (Figures 2a, 2b, and 2c) is characterised by frequency of about 75 Hz and variable amplitude of reflectors. It has the overall shape of complex mound with blocks with chaotic and contorted reflections. Some blocks show hummocky reflections. Blocks have a length ranging between 465–750 m with average of 575 m and a thickness in the range 80–110 ms with an average of 95 ms (TWTT). The lower boundary is generally downlap over an acoustic basement (Figures 2a and 2c). LU-1 facies mainly occurs on flat areas, or mild slopes. The LU-2 facies (Figures 2b and Figure S1 of the auxiliary material) has frequency peak in the range 85–95 Hz and low to medium amplitude. It consists of complex mound with slumps and fan complex compounds with internal wavy, hummocky and chaotic reflections. The mound has a length of about 3.9–4.2 km and a thickness in the range 35–165 ms with an average of 105 ms (TWTT). The LU-3 facies (Figures 2b and 2d) has frequency distribution in the range 85 and 105 Hz and low amplitude. There are mounds with slumps and blocks with hummocky clinoform reflections separated by lateral structural truncations. The slump blocks length ranges from 250 to 675 m with average of 495 m, they have a thickness in the range 70–165 ms with an average of 110 ms (TWTT). The downlap lower boundary has disturbed and wavy reflections; somewhere there are sub-parallel and wavy reflections in apparent basal concordance. LU-3 facies mainly occurs along slopes. The LU-4 seismic facies (Figures 2b and S1 of the auxiliary material) is characterised by about 70 Hz frequencies and high amplitude and wavy to hummocky clinoforms reflections. Internal portions have disturbed and chaotic reflections. The upper boundary has toplap and structural or erosional truncation. The high amplitude strong reflectors in LU-4 facies dip landward (to the West) and are separated by zones about 300–250 ms (TWTT) thick, about 4 km in length, with wavy to chaotic reflections (Figures 1b and S1 of the auxiliary material).
4. Data Interpretation
 Offshore from Mt. Etna, between ridge R and the Catania Canyon (Figure 1a), a mass wasting deposit covers the area (NDEP in Figure 1c). It is mainly constituted by inter-connected LU-1 and LU-3. NDEP, we interpret as a debris avalanche deposit, is generally spatially continuous along the seismic lines. Locally it can disappear or have very reduced thickness, probably for lobes or local steep slopes (Figure 1d). Moreover, some difficulty in discriminating debris can depend also on the relative orientations of seismic line and flow direction. NDEP reaches a maximum distance of 20 km offshore (35 km from the VDB scar on Mt. Etna), a maximum width of 20 km, and a maximum depth of 2000 m b.s.l.; 20 km offshore, deposits are locally up to about 400 ms (TWTT) thick. The total area covered by NDEP is about 250 km2. In NDEP, apparent thicknesses, measured on single seismic profiles, range up to 420 ms with an average of 75 ms (TWTT). Using a P wave interval velocity range of 1700–2200 ms−1, the mass volume of NDEP has been estimated in the range 16–21 km3. Landslide deposits are proximal to the volcaniclastic and alluvium coastal deposits yet identified by Del Negro and Napoli,  and to the Holocene Chiancone Conglomerate and Milo avalanche deposits (Figure 1a), in turn correlated to the VDB scar [Calvari et al., 1998, 2004; Calvari and Groppelli, 1996; Del Negro and Napoli, 2002]. Along the coast, north of the Catania Canyon, marine terrace is absent possibly destroyed by the debris avalanche (Figure 1c). From the above considerations, we infer that the observed landslide deposits originated from the Holocene VDB collapse, also if we cannot exclude multiple events/multiple sources. At present, the VDB scar has a volume of approximately 6 km3, if planar triangular facets are used to connect the upper edges of its southern and northern ridges [Favalli and Pareschi, 2004]; the volume at least doubles if a conic surface is used and it further on increases if eruptive centres (as detected downslope in the Holocene debris deposits [Calvari et al. ) extruded from the volcano flanks. In addition, during the Holocene, the VDB scar was partially filled by erupted lava and tephra, quantified to be hundreds of millions of cubic metres in the last centuries, 0.5 km3 in the last century [Romano and Sturiale, 1982; Branca and Del Carlo, 2004]. Other points to consider in the volume balance are: i) erosion along the debris avalanche path, ii) changes in in-situ lavas/landslide deposit densities; iii) lost material.
 We associate the seismic facies LU-2 (Figures 2b and S1 of the auxiliary material), offshore Acireale, to run-out debris flow/debris avalanche type deposits. We highlight the possibility that this event was related to the collapse of the Pleistocene linear volcanic edifice of Acireale [Corsaro et al., 2002].
 LU-4 facies (underling facies LU-2, Figures 2b and S1 of the auxiliary material) supports the occurrence of a series of slump events. The LU-4 facies can be interpreted as a wedge of pounded sediments that have been back-rotated for the occurrence of deep-seated slumps sliding away along a volcanic flank. This in contrast to Kilauea volcano, where relatively steady state back-rotation appeared to be occurring [Morgan et al., 2000]. Further investigation on lithostatigraphy and geometry will be addressed to fully constraint the genesis and the geologic significance of the LU-4.
 South of the Catania Canyon, other landslide deposits occur (SDEP in Figure 1a), imaged as LU-3 facies. These deposits are older that the marine terrace fronting the coast and are spatially separated from the bulging area in front of Mt. Etna by the striking submarine Catania Canyon (Figure 1a). The few seismic sections available in the area apparently show the canyon slopes are not mantled by landslide deposits. However, we outline the presence of hummocky material offshore the canyon mouth (HAM in Figure 1). Moreover, as already mentioned, we observe that the canyon axis is aligned to inland tracks along which the volcano slides [Froger et al., 2001; Lundgren et al., 2004] (Figure 1a). A possible explanation is that landslide deposits mantled the canyon slopes, removed by recursive slide/slump events related to Mt. Etna sliding. Anyway the sources and the triggers of the HAM and SDEP have to be fully investigated.
 In an auxiliary file, attached to this paper, the latitude and longitude (reference datum: WGS 84), sea bottom depth (m) and upper (facies LU-1 and LU-3) landslide deposit thickness (m) for points along the analyzed seismic lines are reported; thickness are calculated using a 2000 ms−1 velocity.
 High-resolution seismic data show that failure events from Mt. Etna volcano and/or more ancient eruptive centres recur, as testified by the long run-out debris avalanche and slump offshore deposits. Some explanations of the triggering events have been proposed, also if dating of the submarine deposits and more detailed seismic data interpretation as well as more detailed bathymetric surveys are required for further constraining failure dynamics, mass-wasting deposit distribution and the past and present tsunamigenic potential of the area. In any case, imaged landslide deposits from Mt. Etna invoke for reconsideration the tsunami risk in Eastern Mediterranean (and we are just working on it). Let you think that the recent 2002 tsunami at Stromboli, Sieberg-Ambraseys intensity 5 [Maramai et al., 2005], originated from a landslide with a volume of ‘only’ 0.026 km3 [Calvari et al., 2005], just about 103 times smaller than those we detected offshore Mt. Etna.
 The hazard related to surprise tsunamis is not-negligible too, due to the offshore mass wasting deposits dissected by active faults. Possible local slump/slide events can be triggered for example by the shaking effects of a strong earthquake, being the area highly seismogenetic.
 We thank Roberto Francese and Andy Harris for discussions. We acknowledge G. Cristofalo and the team by FUGRO OCEANSISMICA S.p.A. for their professionalism. We also thank David Karatson and two unknown reviewers for their precious revisions.