Tsunami Resonance in Palma Bay and Harbor, Majorca Island, as Induced by the 2003 Western Mediterranean Earthquake

The tsunami induced by the May 21, 2003, Algerian Boumerdès-Zemmouri earthquake (moment magnitudeM w= 6.9) propagated across the western Mediterranean Basin, thereby causing material damages in some harbors and coastal areas. This was the case in the Balearic Islands and particularly the Palma harbor. Attempts to simulate the 2003 tsunami event found discrepancies between tsunami arrival times and wave amplitude when comparing tide gauge records with results from numerical models. To date, all published model results of the amplitude of the tsunami are underestimations, attributed to numerical limitations due to the lack of high-resolution bathymetry and poor definition of harbor geometry. Other views suggest the inappropriateness of the available seismic source parameterizations and the possible occurrence of submarine landsliding triggered by the earthquake that has not been included in the numerical simulations. In this article we present the results of a numerical study aimed at better understanding the response of the interacting Palma bay and harbor under the impact of the 2003 western Mediterranean tsunami. The transference of the tsunami energy from the generation area to the continental shelf, the bay, and the harbor has been studied and compared with the natural oscillation modes of the bay and harbor water bodies. Our work includes a sensitivity analysis of the source parameterization and the bathymetry grid size for the bay and harbor as a way to explain the discrepancies between simulations and observations. The Palma harbor tide gauge shows that energy from the tsunami concentrated mainly in periods that fitted to the natural modes of oscillations of the bay. Therefore, the significant wave amplification observed inside the harbor, mainly in its northern basin, was generated by a resonance effect induced by Palma bay. The improvement of the bathymetry grid resolution in the bay and harbor domains and the inclusion in the simulations of the exact harbor geometry and internal configuration result in a slight wave-high increment that is much below the wave height recorded in the tide gauge. Our results strongly point to a necessary revision of the tsunami seismic source parameters.