Improving models of spatial correlation of earthquake ground motion to enable more informed seismic hazard and risk assessments

Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2022
Thesis identifier
  • T16165
Qualification Level
Qualification Name
Department, School or Faculty
  • In an increasingly risky world, more sophisticated tools to quantify and manage seismic hazard and risk are required. In particular, the potential-loss and risk estimates of spatially distributed infrastructures and portfolios of buildings have posed major challenges to hazard and risk analysts. Indeed, the quantification of the seismic performance of these systems requires estimates of the spatial correlation of earthquake ground motions across a region. The modelling of correlation has gained increasingly importance over the past decades, because it still represents a crucial step in the catastrophe modelling process, but there remain significant uncertainties. This thesis attempts to advance the understanding of spatial correlation by critically investigating the factors and physical parameters that most affect the spatial variability of ground motions. A thorough literature review is provided along with analyses on large databases of recorded strong ground motion from previous earthquakes and ground-motion simulations. General outcomes suggest that spatial correlation properties are period-, regionally- and earthquake-dependent, so that a single stationarity and isotropic spatial correlation model, calibrated on heterogeneous databases including many regions and events, may not be suitable to describe the spatial variability of ground motions for the region of interest. Key findings are used to develop customised correlation models to be included in deterministic (scenario-based) calculations to investigate to what extend spatial correlations may affect risk estimates. Besides these, probabilistic event-based results are also presented to further progress knowledge about the integration of spatial correlation and its associated uncertainties into catastrophe models. Such outcomes may constitute a starting point for the development of an upcoming earthquake catastrophe model for Italy. Finally, the main results of this thesis can provide a primer for loss and risk modellers as well as researchers to understand, interpret and model spatial correlation for the generation of appropriate ground shaking maps and to improve hazard and risk assessments.
Advisor / supervisor
  • Pytharouli, Stella
  • Douglas, John
Resource Type