Preliminary modeling of the tsunami triggered
by the Algiers earthquake, 21 May 2003
 
 

Hélène Hébert (CEA - DASE/LDG), 12 June 2003













  After the 21 May 2003 Algiers earthquake (CSEM-EMSC special page), sea disturbances have been observed on the southeastern coast of the Majorca and Menorca Islands, as well as in Ibiza (Baleares). Witnesses have reported waves up to 2 m high, and a mean period of 10-12 min in Majorca (see here). Tide gage records are available in Palma (Figure 1), showing a maximum water height of about 60 cm (or peak-to-trough amplitude of 1.2 meter).
 
 


Figure 1: Tide gage record in Majorca, courtesy of María Jesús Garcia (IEO)














These sea level variations can be related to a tsunami that could have been triggered either by the coseismic seafloor deformation induced by the submarine fault involved in the earthquake, or by a submarine landslide.
In a first attempt we test here a source due to a coseismic seafloor deformation, that can be estimated through a model of elastic dislocation (Okada, 1985). This initial seafloor deformation is computed (Figure 2) with the following parameters taken from Harvard CMT solution. The mean displacement on the fault plane is about 85 cm.
The first slip distributions obtained from seismological inversions show that the depth was probably shallower than the one we used here, therefore the initial seafloor displacement is probably larger than what is displayed in Figure 2.
 

Fault dimensions (km x km) 40 x 20
Strike, dip, rake
56°, 46°, 71°
Seismic moment (Nm)
0.201 1020
Rigidity (Nm2)
30 109
Epicenter
3.52°E - 36.98°N
Depth (km)
17

Figure 2: Initial seafloor deformation computed with the parameters above.
Note the small negative deformation compared to the positive one.














The initial deformation is assumed to be fully and instantaneously transmitted to the sea surface, where, through restoring gravity forces, tsunami waves begin to propagate across the sea.
Since tsunami wavelengths are much larger than the mean water depth, we assume that the long wave theory is valid. The simulation is performed using a finite difference method that solves the hydrodynamic equations of continuity and of motion, including non linear terms.
Nested bathymetric grids are used to account for the wave amplification near the coasts. Here we used data taken from the global bathymetry data deduced from altimetry (compiled by Smith and Wessel, 1997), and we build a 2' grid describing the western Mediterranean Sea, a 30'' grid for the Baleares Islands, and 10'' grids for the Majorca and Menorca Islands (Figure 3).


Figure 3: Computational domain, and location of the fine grids describing the Baleares Islands.



 
 
 
 
 
 
 
 
 
 
 

  • Results
    • The waves reach the Baleares Islands after a propagation of about 20 to 30 minutes, in agreement with the arrival in Palma roughly deduced from Figure 1. As is usual in far-field tsunami propagation, the highest energy, corresponding to the maximum water heights displayed in Figure 4, is limited to an area perpendicular to the fault strike (Hébert et al., 2001a), posing a significant threat to the Baleares Islands for this kind of fault azimut.
    An animation of the propagation is available here (.gif file, ~8 Mo, or low-quality .mpeg file, 530 ko).
     

    Figure 4: Sea surface computed after a 25 min tsunami propagation.

    Figure 5: Maximum water heights computed after a 1.5 h propagation.

     
     

    In the Baleares, the results must be considered with caution. They do not display a great amplification of the waves (Figures 6 and 7). Since the submarine slopes are quite steep, especially towards Majorca (Figure 3), the tsunami waves could have been reflected, protecting Palma from a large amplification.

    However it is worth noting that the areas of (relatively) highest computed amplification correspond to the localities where abnormal waves have been reported, in Majorca (Porto Cristo) and Menorca (Mao). Further studies should definitely use refined bathymetric data in the harbours to assess the hazard in the Baleares more precisely.
     
     
     


    Figure 6: Maximum water heights reached in Majorca.

     Figure 7: Maximum water heights reached in Menorca.

      The modelings carried out here are very preliminary, and use rough bathymetric data only, especially for the fine grids. To account for the wave amplification reported along the SE coasts of Majorca or Menorca, fine data should be added in the bathymetric grids used for the simulation.
    Anyway these modelings show that a coseismic source can explain tsunami observations in the Baleares Islands, especially along their SE coasts.
    We note that tsunami observations along the Algerian coast have not be reported so far, and attention must be paid to future witnesses accounts.
    Finally, it is also worth mentioning that a dense and good-quality tide gage network on the Mediterranean coasts could allow the implementation of inversion techniques to recover the submarine coseismic deformation, as it is performed in Japan for earthquakes located in the Nankai Trough (e.g. Tanioka and Satake, 2001 ; Baba et al., 2002).
     
     
     
     
     
     
      Baba, T., Y. Tanioka, P.R. Cummins and K. Uhira, The slip distribution of the 1946 Nankai earthquake estimated from tsunami inversion using a new plate model, Phys. Earth Plante. Int., 132, 1-3, 59-73, 2002.
    Hébert, H., P. Heinrich, F. Schindelé, and A. Piatanesi, Far-field simulation of tsunami propagation in the Pacific Ocean: impact on the Marquesas Islands (French Polynesia), J. Geophys. Res., 106, C5, 9161-9177, 2001.
    Okada, Y., Surface deformation due to shear and tensile faults in a half-space, Bull. Seismol. Soc. Am., 75, 1135-1154, 1985.
    Smith, W.H.F., and Sandwell, D.T., Global seafloor topography from satellite altimetry and ship depth soundings, Science, 277, 1956-1962, 1997.
    Tanioka, Y., and K. Satake, Detailed coseismic slip distribution of the 1944 Tonankai earthquake estimated from tsunami waveforms, Geophysical Research Letters, 28, 1075-1078, 2001.
     
     
     
      Guibourg, S., P. Heinrich and R. Roche, Numerical modeling of the 1995 Chilean tsunami. Impact on French Polynesia, Geophysical Research Letters, 24, 775-778, 1997.
    Hébert, H., P. Heinrich, F. Schindelé, and A. Piatanesi, Far-field simulation of tsunami propagation in the Pacific Ocean: impact on the Marquesas Islands (French Polynesia), J. Geophys. Res., 106, C5, 9161-9177, 2001.
    Hébert H., F. Schindelé, and P. Heinrich, Tsunami risk assessment in the Marquesas Islands (French Polynesia) through numerical modeling of recent and generic far-field events, Natural Hazards and Earth System Sciences, 1, 233-242, 2001.
    Hébert, H., A. Piatanesi, P. Heinrich, F. Schindelé, and E. A. Okal, Numerical modeling of the September 13, 1999 landslide and tsunami on Fatu Hiva Island (French Polynesia), Geophys. Res. Lett., 29, 10, doi:10.1029/2001GL01374, 2002.
    Heinrich, P., A. Mangeney, S. Guibourg, R. Roche, G. Boudon, and J.-L. Cheminée, Simulation of water waves generated by a potential debris avalanche in Montserrat, Lesser Antilles, Geophysical Research Letters, 25, 3697-3700, 1998.
    Heinrich, P., F. Schindelé, S. Guibourg and P.F. Ihmlé, Modeling of the February 1996 Peruvian tsunami, Geophysical Research Letters, 25, 2687-2690, 1998.
    Heinrich, P., R. Roche, A. Mangeney and G. Boudon, Modéliser un raz de marée créé par un volcan, La Recherche, Mars 1999, 66-71, 1999.
    Heinrich, P., S. Guibourg, A. Mangeney, and R. Roche, Numerical modeling of a landslide-generated tsunami following a potential explosion of the Montserrat Volcano, Physics and Chemistry of the Earth, 24, 16-168, 1999.
    Heinrich, P., A. Piatanesi, E. Okal, and H. Hébert, Near-field modeling of the July 17, 1998 event in Papua New Guinea, Geophysical Research Letters, 27, 3037-3040, 2000.
    Heinrich, P., A. Piatanesi, and H. Hébert, Efficiency of deep submarine landslides in producing tsunamis: the 1998 Papua New Guinea event, GJI, 145, 97-111, 2001.
    Schindelé, F., D. Reymond, E. Gaucher and E.A. Okal, Analysis and automatic processing in near-field of eight 1992-1994 tsunamigenic earthquakes: improvements towards real-time tsunami warning, Pure and Applied Geophysics, 144, 381-408, 1995.