Gianluca Pagnoni

Large, deepwater slope failures: Implications for landslide-generated tsunamis

Deepwater landslides are often underestimated as potential tsunami triggers. The North Gorringe avalanche (NGA) is a large (~80 km3 and 35 km runout) newly discovered and deepwater (2900 m to 5100 m depth) mass failure located at the northern fl ank of Gorringe Bank on the southwest Iberian margin. Steep slopes and pervasive fracturing are suggested as the main preconditioning factors for the NGA, while an earthquake is the most likely trigger mechanism. Near-fi eld tsunami simulations show that a mass failure similar to the NGA could generate a wave >15 m high that would hit the south Portuguese coasts in ~30 min. This suggests that deepwater landslides require more attention in geo-hazard assessment models of southern Europe, as well as, at a global scale, in seismically active margins.

Comment on “On the cause of the 1908 Messina tsunami, southern Italy” by Andrea Billi et al.

[1] The December 28th 1908 Messina Earthquake has been ranked as one of the most destructive events of the last centuries [Davison, 1936] The damages produced by ground shaking were aggravated by the effects of a remarkable tsunami, with up to 11 m of run-up height, that followed the earthquake [Omori, 1909; Baratta, 1910; Tinti et al., 2004]. The location of the causative fault is still a matter of debate [e.g., Argnani et al., 2009], the modelling of the associated tsunami [Tinti and Armigliato, 2003] allowing uncertainty. [2] Billi et al. [2008] have recently proposed that the tsunami that stroke the coast of the Strait in December 1908 originated from a large submarine landslide (20 km) located offshoreGiardini-Naxos. The hypothesis is based on the study of tsunami arrival times [Platania, 1909; Baratta, 1910] and is supported, according to Billi et al. [2008], by inspection of: a multibeam morphobathymetry [Marani et al., 2004] and of a crustal-scale seismic profile [Scrocca et al., 2004]. [3] The arguments of Billi et al. [2008], however, have weak points that cast doubt on their interpretation. But most importantly, data collected in the last few years by the authors of this comment (A. Argnani and F. Chiocci), and work on tsunami modelling (S. Tinti and his group) cast further doubt on the proposed hypothesis that a large submarine landslide that was caused by the 1908 earthquake is located offshore Giardini-Naxos. These issues will be discussed in the following.

The 1977 Gioia Tauro Harbour (South Tyrrhenian Sea, Italy) Landslide-Tsunami: Numerical Simulation

On July 12th, 1977, waves 5 m high hit the harbour of Gioia Tauro, provoking relevant damages on facilities and infrastructures, but fortunately no casualties. The source was supposed to be a submarine mass failure, ranging 5 million m3 and occurring in very shallow water at the heads of a canyon just in front of the port. The sliding mass cut a cable at 600 m depth, meaning that it travelled more than 15 km far from the source. In this work we reconstruct the hypothesized mass failure, compute the slide dynamics and simulate the generated tsunami through numerical modelling. The results obtained fit well with the observations, strengthening the hypothesis of such a landslide as the source for the 1977 tsunami. This event can be considered as paradigmatic of a category of coastal hazards: small submarine landslides occurring close to populated coasts may pose considerable risk, even if only at a local scale.

Numerical Simulation of the BIG’95 Debris Flow and of the Generated Tsunami

Underwater mass movements along continental margins constitute a relevant threat for coastal communities. In the Mediterranean Sea this issue is particularly critical, both for the concentration of potential sources and for the density of population living along the coast. One of the biggest event recently mapped by submarine geophysical surveys, named BIG’95, is located in the Balearic Sea, between the homonymous islands and the Eastern coast of Spain: it occurred around 11 kyrs BP, involving a volume of around 26 km3 over an area of 2,200 km2, from 200 m depth down to 1,600 m and over: the looser fraction of the mass was driven along the Valencia Channel, NE-ward, more than 110 km far from the source area. The BIG’95 event is here simulated through numerical codes, adopting some reasonable simplifications on the slide morphology and dynamics. The tsunamigenic impulse is then evaluated and the propagation of the wave simulated over a computational domain covering the Western Mediterranean Sea. Results show that an event of these characteristics can generate catastrophic tsunami along the coasts of the Balearic Sea, reaching also North Africa, Corsica and Sardinia coasts with relevant waves.

Reconstruction and Tsunami Modeling of a Submarine Landslide on the Ionian Margin of Calabria (Mediterranean Sea)

The Ionian margins of Calabria are affected by repeated sediment failures, recorded by slide scars at seabed and stacked slide deposits. We present a reconstruction of the geometry and dynamics of one of the largest seabed features, the Assi failure on the relatively steep slope off southern Calabria, and use it as input to numerical modeling to evaluate the potential tsunamigenic hazard. The Assi failure is up to 6 km wide and at least 18 km long, and involved the displacement of ca. 2 km3 of sediment, inferred to have taken place within the last 4,000 years in two main phases. The first and larger phase is used as input to the tsunami modeling, on the assumption that the slide moved in a single step as a coherent mass of 1.85 km3, in order to evaluate the most disruptive possible consequences. The results indicate that within 8 minutes, waves just over 1 m in height affect the southern Calabrian coast between Monasterace and Roccella Jonica, where their capacity to cause damage could be amplified in small harbours. This shows that tsunamis represent a hazard for Ionian coastal areas, and calls for accurate monitoring and further study.

The Double Landslide‐Induced Tsunami

The 2002 crisis of Stromboli culminated on December 30 in a series of mass failures detached from the Sciara del Fuoco, with two main landslides, one submarine followed about 7 min later by a second subaerial. These landslides caused two distinct tsunamis that were seen by most people in the island as a unique event. The double tsunami was strongly damaging, destroying several houses in the waterfront at Ficogrande, Punta Lena, and Scari localities in the northeastern coast of Stromboli. The waves affected also Panarea and were observed in the northern Sicily coast and even in Campania, but with minor effects. There are no direct instrumental records of these tsunamis. What we know resides on (1) observations and quantification of the impact of the waves on the coast, collected in a number of postevent field surveys; (2) interviews of eyewitnesses and a collection of tsunami images (photos and videos) taken by observers; and (3) on results of numerical simulations. In this paper, we propose a critical reconstruction of the events where all the available pieces of information are recomposed to form a coherent and consistent mosaic.

Investigations on the Possible Source of the 2002 Landslide Tsunami in Rhodes, Greece, Through Numerical Techniques

The island of Rhodes (Aegean Sea, Greece) has been repeatedly hit by tsunamis in the past due to the numerous tsunamigenic sources present in the area, most of which are seismic. Here an investigation is made on the most recent event that occurred on March 24, 2002 in the northeast of the island: unexpected waves affected a 2 km long coastal segment, overtopping part of the sea-wall (3–4 m high) that runs along the littoral road a few km away from the city of Rhodes. Data on the tsunami are poor. Due to the lack of evidence of seismological or meteorological causes, the hypothesis of a landslide source for the tsunami is here explored by means of numerical codes implementing models both for the landslide dynamics and the tsunami propagation. The reconstructed failing mass is located at about 100 m sea depth with a volume of 30 million m3. A sensitivity analysis is further run by varying the source volume, to evaluate the effects on the tsunami impact.

Chapter 15 – Geoethical and Social Aspects of Warning for Low-Frequency and Large-Impact Events like Tsunamis

Real-time warning of large-impact events like tsunamis, that may attack the nearest coast soon after the generation, implies that there is no time to collect enough data before tsunami arrival to provide accurate estimates of the wave height and tsunami duration, which may result in issuing inaccurate alarm messages to the population. Assessments of tsunami hazard and analyses of tsunami risk, that are at the basis of long-term prevention policies, can be affected by large uncertainties as well, since they suffer from systematic scarcity of real observations because large tsunamis are very infrequent events. This paper discusses the role of geoscientists in providing predictions and the related uncertainties. It is stressed that through academic education geoscientists are formed more to improve their understanding of processes and the quantification of uncertainties, but are often unprepared to communicate their results in a way useful for society. Filling this gap is crucial for improving the way geosciences and society handle natural hazards and devise effective means of protection.