Earthquake losses registered worldwide over the last century have triggered crippling effects on the economic and social systems of wealthy and undeveloped countries alike. In the face of ever increasing impacts, earthquake loss modelling is essential for the prediction, prevention and mitigation of the adverse effect of future seismic events. Given the complex nature of the process, it is utopian to seek absolute certainty when gathering the required resources.
As an inherent property of any analytical process, uncertainty must not be 'removed' from the equation. Therefore, seismic risk assessments can only be meaningful if coupled with its accompanying analysis of uncertainty, be that aleatory or knowledge-based. In addition to the continued improvement in the characterization of random and epistemic uncertainties in seismic hazard, recent years have seen a major swing in emphasis towards the explicit inclusion of uncertainties in the performance assessment of buildings.
Building upon the link between seismic hazard and ground motion selection, these efforts have improved the treatment of record-to-record variability and/or the random nature of geometric and structural building parameters. However, several questions remain entirely unanswered or lack a deeper understanding.
Within the context of building portfolio loss estimation, this talk will explore issues relating to (1) the impact on building fragility of hazard-consistent record selection and hazard disaggregation, and (2) uncertainty characterisation in loss metrics and fragility analysis. It highlights shortcomings in the state-of-art and the implications of this for earthquake loss modelling.
For more information please contact: