Following the objectives outlined in WP1 of the S4H project, the S4H team, in collaboration with Amirhossein Mohammadi from the University of Minho and Prof. Aysegul Askan from the Middle East Technical University, has recently published a paper in the Geophysical Journal International (GJI). This publication details the researchers’ exploration of the effectiveness of incorporating stochastic ground motion simulations and advanced machine learning (ML) techniques, specifically artificial neural networks (ANN), to enhance the accuracy of seismic parameter predictions related to earthquake events.
In response to Turkey’s frequent seismic activity, this study investigates the potential of stochastic ground motion simulations to model real earthquake motions. By utilising these simulations alongside artificial neural networks (ANN), the research aims to enhance the precision of earthquake motion representation. The approach develops a comprehensive local GMM specific to Turkey, using region-specific simulated earthquake records with the ANN algorithm and considering factors such as fault mechanism, focal depth, moment magnitude, distance from the earthquake source, and shear wave velocity. With 7359 records, the study evaluates spectral ordinates, peak ground acceleration, and velocity, shedding light on the GMM’s effectiveness and uncertainties, ultimately confirming its ability to accurately replicate seismic phenomena, particularly during the February 6th, 2023, earthquakes in Turkey.
Furthermore, the researchers developed an online platform to enhance accessibility to the models for interested users (https://amirxdbx-gmm-turkey-simulation-deploy-lp47x4.streamlit.app/). The source codes are available at https://github.com/amirxdbx/GMM_Turkey_simulation.
For more details, click here. The published version of the paper can also be found at the following link:
Share this post