The propagation of tsunami waves over time (selected faults from Eastern Mediterranean Coast)

Dataset were employed to simulate the propagation of tsunami waves. Our study modelled 42 fault zones located in the eastern Mediterranean to identify those that could have generated an earthquake strong enough to cause flooding and destruction of the Porphyreon through tsunami wave inundation.
Numerical modelling techniques formed a key component of this study and were employed to simulate the propagation of tsunami waves based on data from contemporary seismic stations. The main controlling factors influencing the generation of earthquakes—and, indirectly, tsunamis—were considered to be fault geometry (length, width, strike, dip, depth, depth to top) and dislocation parameters (strike, rake, slip, shear modulus). Subsequently, these were utilised in the course of further analyses. We analysed the earthquakes with epicentres located in the south-eastern Mediterranean taking into the account the geological structure of the region. To facilitate tsunami wave propagation modelling, we used MIRONE software package (for details see Luis, 2007) which incorporates the TINTOL code (http://fct-gmt.ualg.pt/mirone/downloads/windows.html, accessed December 2022) for hydrodynamic processes simulation. The method used in this study follows that of Omira et al. (2009), in which the initial disturbance on the sea surface—caused by seafloor deformation— is calculated using algorithms developed by Mansinha and Smylie (1971).
The subsequent step involved modelling the deformation of the seafloor surface using the EMODnet Digital Bathymetry (DTM 2020) and the models proposed by Mansinha and Smylie (1971). This approach was selected based on its computational efficiency and its abibility to minimise the uncertainties commonly associated with input parameters. To further reduce sensitivity to poorly constrained variables the model was configured to calculate exclusively the vertical component of earthquake-induced deformation. For the tsunami propagation simulation, a time step of 1.5 seconds was used, and calculations were performed for a duration of 3,000 seconds from the wave excitation. To accomplish this, we considered several key parameters, including hypocentre depth, fault angle, dip and displacement slip, and shear modulus, all of which were based on present-day seismological datasets. Where direct information was unavailable, values were inferred from geological cross-sections and regional seismic profiles. In the absence of detailed fault geometry data, we made the following assumptions: 80 degrees for slip faults, 25 degrees for thrust faults, and 60 degrees for normal faults. Hypocentre depths for the model were obtained from the USGS Earthquake Catalogue database provided the relevant fault had exhibited seismic activity within the observational timeframe of the seismic stations. Where no direct information was available, the hypocentre was assumed to be located at the midpoint of the total fault depth. Fault dip and direction of displacement were determined using focal mechanism interpolations data available through the USGS Earthquake Catalogue. In instances where the information was lacking, the direction of displacement was estimated by drawing analogies from other tectonically similar regions worldwide. The slip parameter was defined as the minimum displacement required to achieve a specific earthquake magnitude.