Advanced Methods for Earthquake Loss Modeling

Estimating the potential losses that earthquakes inflict on the built environment is central to seismic risk management and transfer. Producing credible estimates requires moving beyond isolated, single-event analyses toward a probabilistic framework that reflects how earthquakes actually affect geographically distributed portfolios. This thesis develops a Monte Carlo simulation engine for regional earthquake loss estimation, built around three aspects that strongly influence aggregate losses yet are often simplified or neglected. Ground motion is represented through spatially correlated fields, so that the shaking experienced across distributed assets remains jointly consistent. Seismic events are modeled not in isolation but as mainshock–aftershock sequences, capturing the losses accumulated during the aftershock phase. Throughout, uncertainty is propagated systematically, from ground motion prediction to structural fragility and vulnerability. Applied to Italian exposure and calibrated on historical earthquakes, the framework produces exceedance-probability curves and standard risk metrics, offering a transparent and extensible tool for quantifying earthquake loss under realistic modeling assumptions.