Three-Dimensional Shear Wave Velocity Modeling in Tabriz City Using Ensemble Neural Networks and Geospatial Interpolation

This study develops a high-resolution ۳D shear wave velocity (Vs) model for Tabriz, Iran, by integrating field measurements with advanced machine learning and geostatistical techniques. Initial Vs profiles from ۲۶ geotechnical stations were expanded to ۴۲۶ locations using ensemble neural networks, enabling detailed spatial analysis through GIS-based kriging interpolation. The model reveals strong correlations between Vs distribution and geological features, including sediment thickness and bedrock depth. Southern Tabriz exhibits higher Vs values (۶۰۰-۱,۲۰۰ m/s) due to shallow bedrock, while central and northern regions show lower velocities (۲۰۰-۵۰۰ m/s) associated with thick alluvial deposits. The methodology demonstrates the efficacy of hybrid Al-geostatistical approaches for urban seismic characterization in data-scarce environments, providing critical insights for earthquake hazard mitigation and infrastructure planning.

کلیدواژه ها:

Shear wave velocity (Vs) ، ۳D geotechnical modeling ، Ensemble neural networks ، Spatial interpolation ، Kriging