Tsunami hazard and risk evaluation in the Gulf of Naples: State of - - PowerPoint PPT Presentation

Tsunami hazard and risk evaluation in the Gulf of Naples: State of the art and perspectives

• S. Tinti, F. Zaniboni, A. Armigliato and G. Pagnoni

University of Bologna, Dept. of Physics, Sector of Geophysics, Bologna, ITALY

COST Conference on “Urban Habitat Constructions Under Catastrophic Events” 16-18 September 2010 Naples (Italy)

PRESENTATION OUTLINE

• Brief overview of the tsunami history of the Gulf of Naples
• Tsunami hazard assessment through numerical simulations of

historical/scenario tsunamis generated in the near field (Vesuvius, Ischia) and in the far field (Stromboli)

• Future perspectives: tsunami vulnerability and risk assessment

HISTORICAL TSUNAMIS IN THE MEDITERRANEAN

Catalogue produced as a deliverable of the EC FP6 TRANSFER Project (http://www.transferproject.eu/). Dimension and colour of the symbols indicate different values of tsunami intensity expressed in the Sieberg-Ambraseys scale.

TSUNAMIS IN THE GULF OF NAPLES

ISCHIA CAPRI NAPLES VESUVIUS (6 events) 1760

2 low-reliability events

QUERYING THE DATABASE FOR EVENT PARAMETERS

QUERYING THE DATABASE FOR EVENT DESCRIPTION

Vesuvius is not the only cause of tsunamis in the Gulf of Naples! Near field: Ischia (landslides) Far field: Stromboli (landslides)

TSUNAMI HAZARD ASSOCIATED WITH VESUVIUS PYROCLASTIC FLOWS Reference case: 1631 eruption and tsunami

According to Rosi et al. (1993), in the harbour of Naples waves of 2-5 m in amplitude were observed. From the TRANSFER tsunami catalogue, we know that the sea was seen to withdraw and then to inundate the coast of the Gulf for three times: the largest retreat (about 1 mile) was reported in Sorrento.

FORCING

From Tinti et al. (2003) based on the hypothesis that the tsunami was due to the light component of the pyroclastic flow (see also Esposti Ongaro et al. (2002)). 2 min 5 min 10 min Dynamic pressure (kPa) cm

Time (min)

Forcing is assumed to be a pressure impulse propagating seaward with constant radial velocity over a sector centered on the Vesuvius crater. Pulse duration assumed constant. Pulse radial speed: 15 m/s.

TSUNAMI PROPAGATION FIELDS

5 min 10 min 20 min 30 min Field of maximum tsunami energy (kJ/m2) in the first 30 minutes after the onset. cm Computed by means of the numerical FE code UBO- TSUFE, developed and maintained by TRT- DFUNIBO and implementing the non-linear shallow water equations (see Tinti et al., 1994).

MAXIMUM WATER ELEVATION ALONG THE COAST

1-Miseno, 2-Pozzuoli, 3-Bagnoli, 4-Posillipo, 5-Napoli, 6-Torre del Greco, 7-Torre Annunziata, 8-Castellammare, 9-Sorrento, 10-Positano

20 40 60 80 100 120

• 0.4
• 0.2

0.0 0.2 0.4 0.6

10 9 8 7 6 5 4 3 2 1

20 40 60 80 100 120

• 0.4
• 0.2

0.0 0.2 0.4

10 9 8 7 6 5 4 3 2 1

SENSITIVITY ANALYSIS

Distance along the coast (km) Distance along the coast (km) Extreme water elevations (m) Extreme water elevations (m)

VARYING PULSE SPEED VARYING PULSE DURATION

V = 10 m/s V = 15 m/s V = 20 m/s V = 25 m/s 1.5 min 3.0 min 6.0 min

TSUNAMI HAZARD ASSOCIATED WITH CATASTROPHIC LANDSLIDES AT ISCHIA

Ischia Debris Avalanche (IDA)

Chiocci F.L., de Alteriis G. (2006). The Ischia debris avalanche: first clear submarine evidence in the Mediterranean of a volcanic island prehistorical collapse, Terra Nova, 18, 202-209.

• Series of surveys around Ischia from 1998 to

2004 evidenced submarine deposits all around the island (N, W and S)

• IDA as a worst case scenario
• The largest deposit was located south of

Ischia

• covering an area of more than 250 km2
• up to 50 km far from the coast
• thickness ranging 5-30 m
• volume of 1.5-3 km3
• dating some ky BP

LANDSLIDE NUMERICAL MODEL (UBO-BLOCK)

• Mass divided into a “chain” of

blocks that:

• conserve their volume
• can interact
• can’t separate
• Centre of mass model

Tinti S., Bortolucci E., Vannini C., 1997. A block-based theoretical model suited to gravitational sliding, Natural Hazards, 16, 1-28.

Motion equation aik = Gik - Rik + Fik

i time step k block

Gik gravity and bottom friction Rik resistance term Fik internal interaction term

• Numerical resolution of motion equation
• The simulation is stopped when mass

velocity goes below a predefined threshold

• r when the slide exits the computational

domain

LANDSLIDE SIMULATION

• MODEL INPUTS
• sliding surface
• top of sliding mass
• predefined centre of mass trajectory

(red line)

• lateral slide boundaries

Chiocci F.L., de Alteriis G. (2006)

LANDSLIDE SIMULATION

initial slide thickness Volume ~ 3.7 km3

Distance along the profile (m) Height (m)

Time step: 1 second Number of blocks: 10 Subaerial friction coefficient (ma): 0.09 Submarine friction coefficient (mw): 0.02 Density: 2000 kg/m3 Final simulation time: 710 s

LANDSLIDE SIMULATION

LANDSLIDE SIMULATION

Landslide profile

• Mass deposits between 900 and 1000 m b.s.l.
• runout of about 30 km

LANDSLIDE SIMULATION

VELOCITY AND FROUDE NUMBER Fr = Vhor √ g·h

COMPUTED TSUNAMI PROPAGATION FIELDS (UBO-TSUFE code)

MAXIMUM AND MINIMUM WATER ELEVATIONS OVER THE COMPUTATIONAL DOMAIN

TSUNAMI HAZARD ASSOCIATED WITH CATASTROPHIC LANDSLIDES AT STROMBOLI

December 30th 2002 Tsunamis

Photos taken during the January 2003 post-event field survey

• Caused by 2 landslides in Sciara del

Fuoco

• The first submarine, the second sub

aerial

• Relatively small-sized, volume ~ 106-

107 m3

• Waves reaching 10 m runup,

affecting all the coasts of the island

FAR-FIELD TSUNAMI EFFECTS OBSERVATIONS AFTER THE 30 DECEMBER 2002 EVENTS

From Maramai et al. (2005) For example, at Marina di Camerota (point 5) a sea withdrawal

• f about 5–12 m was
• bserved.

In the zones adjacent to the harbour, seawater penetrated inland for about 60–70 m

THE HOLOCENE SCIARA DEL FUOCO COLLAPSE

See e.g. Rosi et al. (2000) Tinti et al. (2000, 2003) Tibaldi (2001)

• occurred about 5 ky BP
• volume estimated in 1 km3
• landslide simulated by

means of UBO-BLOCK2D (25 blocks, 0.5 s time step)

• final simulated run-out

distance: 15 km

Stromboli Holocene landslide scenario: Snapshots of the tsunami propagation in the southern Tyrrhenian

(computed with a finite difference code by Dr. Kenji Satake based on TRT-DFUNIBO landslide simulation)

Maximum positive elevation field (inundation) Maximum negative elevation field (retreat)

Extreme water elevation fields computed in the Naples harbour area for the Stromboli Holocene landslide scenario

(computed by means of UBO-TSUFD code in the frame of the DPC “Progetto Porti” coordinated by EUCENTRE)

FUTURE PERSPECTIVES: TSUNAMI VULNERABILITY AND RISK ASSESSMENT

The tsunami risk estimation and mitigation problem has been tackled in a systematic way only in very recent years. In general terms, the risk estimation is the result of the combination of detailed inundation maps coming from the hazard analysis, and of the vulnerability assessment. One of the approaches existing for the second aspect has been developed by the EU-funded SCHEMA project: it distinguishes between primary (type and material) and secondary (ground, age, foundation, orientation, etc.) criteria for buildings, and it adopts a building damage matrix, basically depending on building type and water inundation depth.

Vulnerability classification established in SCHEMA

Vulnerability classification for selected areas around Catania

Building damage matrix adopted by SCHEMA

CONCLUSIONS

The tsunami hazard assessment for the Gulf of Naples has been performed so far exclusively on the base of the scenario technique. Three main different sources of tsunami hazard have been identified in the volcanic activity of Vesuvius and on the possible occurrence of large to catastrophic landslides in correspondence with Ischia and Stromboli. The main threat appears to be related to landslides in Ischia, with the IDA scenario forecasting tsunami waves impacting the Gulf within 10-15 minutes and with maximum elevations in the order of 20 m. Our numerical simulations must be improved as regards the tsunami generation by pyroclastic flows. Tsunami vulnerability and risk assessment is a topic that has only recently started to be tackled. We have briefly described a method that have been proposed by the EC SCHEMA Project, based on the assumption of proper vulnerability and damage matrices, to be crossed with the expected maximum inundation in a given site provided by numerical simulations. We have shown a preliminary application to the city of Catania.

alberto.armigliato@unibo.it http://www.transferproject.eu/ http://www.schemaproject.net/

TSUNAMI MODEL – COMPUTATIONAL GRID

Km Km

TSUNAMI SIMULATION

MAXIMUM AND MINIMUM WATER ELEVATIONS ON THE COAST Ischia Procida

Southern Tyrrhenian coast

TSUNAMI SIMULATION

SYNTHETIC MARIGRAMS

CONCLUSIONS

• Ischia island represents a potential source for tsunami that can have effects also
• n the regional scale
• The definition of the IDA limit-case scenario allows us to assess the maximum

hazard connected to landslides in this area

• A subaerial-submarine scenario has been adopted for IDA
• High velocity reached (around 60m/s as a maximum)
• The generated wave reaches the Gulf of Naples in some minutes, continuing to

perturbate the basin for almost 2 hours due to tsunami energy entrapment

• Apart from some isolated peak points, Ischia is reached almost everywhere by

20-30 m wave, Procida by 10-20 and the Tyrrhenian coast by 10 m almost everywhere

• Can subaerial landslides reach the sea and be associated to the
• bserved submarine deposits?