Search for a command to run...
Seismic risk represents a major concern for densely populated urban areas, particularly in regions characterized by persistent volcanic and tectonic unrest. The city of Naples, southern Italy, is currently affected by an ongoing bradyseism crisis associated with the Campi Flegrei caldera, which has resulted in frequent low-to-moderate magnitude earthquakes (M 2–3+) over recent months. In this context, this study presents an integrated, data-driven framework for urban-scale earthquake risk mapping that combines probabilistic seismic hazard assessment with exposure and vulnerability modelling using convolutional neural networks (CNNs) and GIS techniques. Seismic hazard was quantified using earthquake records spanning 1990–2024 and modelled through six conditioning factors: elevation, slope, earthquake magnitude density, epicentral density, distance to epicentres, and peak ground acceleration. These spatial layers were integrated using a CNN architecture to generate a probabilistic hazard map representing the likelihood of earthquakes with magnitudes ≥3.5. Human exposure was subsequently assessed by integrating gridded population datasets with building footprints and parcel-level spatial data where available. Structural vulnerability was estimated through the fusion of land use/land cover information and recent building height data, both reclassified into susceptibility scores reflecting potential earthquake damage. The combined vulnerability index was categorized into five classes, with higher values corresponding to dense urban areas and taller building stock. The final seismic risk map was produced by integrating hazard, exposure, and vulnerability layers. Results highlight that areas characterized by high population density and intensive urban development sexhibit the highest seismic risk, consistent with observed urban patterns. The proposed methodology offers a transferable and automated approach for urban seismic risk assessment and can support risk-informed planning and disaster mitigation strategies in seismically active metropolitan regions.