Seismic hazard assessment in Menton, France: Topographical site effect zoning considering a semi-empirical approach and a Machine Learning scheme

TL;DR

This study assesses topographic seismic amplification in Menton, France, using a semi-empirical and machine learning approach to classify amplification effects and produce a microzonation map for better seismic risk management.

Contribution

It introduces a novel combination of the Frequency-Scaled Curvature method with machine learning clustering for seismic microzonation based on topographic effects.

Findings

Five distinct amplification groups identified

Microzonation map of topographic effects created

Method improves accuracy over Eurocode coefficients

Abstract

The presence of topography influences the seismic ground motion and may result in strong amplifications, generally at the top of hills and reliefs. The increasing urbanization of hills requires an accurate estimation of these effects even in areas of moderate seismicity. The simplified coefficients provided by the Eurocodes8 do not depend on the frequency and underestimate the amplification in many situations, which justifies the development of new methods based on easily accessible data. The city of Menton, located in the southeast of France, between the Alps and the Ligurian basin, is one of the most exposed metropolitan cities. We propose a study of topographic effects applied to the Menton area. Topographic amplification is calculated, on a wide frequency band, using the Frequency-Scaled Curvature method, from a DEM and an average value of the shear wave velocity. We then propose to apply an automatic clustering approach to classify the amplification curves into five groups with similar properties. We then deduce a first microzonation map of the topographic effects in the Menton area.