A Monte Carlo simulation of error propagation in a GIS-based assessment of seismic risk

Abstract

This paper uses a GIS-based Monte Carlo simulation to gauge the sensitivity of risk assessment results to random perturbations in earthquake ground shaking intensity zone boundaries. Sensitivity to error in intensity estimation is also considered. Zonal boundaries may vary within a 1560 m corridor and still yield property loss estimates that are within five per cent of their original estimate 95 per cent of the time. Yet a five per cent error in ground shaking intensities yields an 11 to 15 per cent error in property loss estimates. The method illustrates the use of GIS in simulating error propagation and identifying strategic error reductions in risk assessments.