Abstract

In seismic hazard assessment, reliance on Vs30 proxies and 1D shear wave velocity profiles often leads to underestimated ground motion. This is particularly evident in areas with complex geological structures, such as Greater Beirut (GB). The metropolis, situated near active seismic faults, experienced significant nearby earthquakes in 551, 1202, and 1837. It is characterized by diverse soil compositions, that vary from sandy terrains to limestone formations, demanding a detailed geotechnical model for seismic hazard studies. Our research developed a comprehensive 3D geotechnical model for GB, integrating data from around 500 boreholes, 700 geophysical measurements, refined DEM, and geological insights. The model delineates variations in bedrock elevation and geological strata, some sites exhibiting sediment depths up to 80 meters. We performed an iterative data analysis by combining the horizontal to vertical spectral ratio method (H/V measurements) with borehole data. This approach enabled us to estimate the average shear wave velocity (Vs-mean) in the sedimentary layer and the depth of the bedrock across the model. To address data gaps in southern GB, we used a Random Forest machine learning model, trained on interpolated points from Kriging in the central model part, ensuring continuous representation of sedimentary units even in data-limited areas. Ongoing work involves seismic simulations predicting ground motion amplification in Beirut. Using a 3D hexahedral mesh generated via Python code, we will conduct full 3D numerical simulations of seismic wave propagation. These simulations aim to provide insights into Beirut's seismic response, contributing to earthquake preparedness and risk mitigation. We will present preliminary results in predicting the seismic motion in Greater Beirut using SPECFEM3D, a spectral-element method software designed for 3D seismic wave propagation simulations.