Empirical Tsunami Fragility Functions Derived Using Improved Statistical Methods

Abstract eng:
Tsunami fragility functions for buildings provide a probabilistic link between tsunami intensity and building damage. They are a component of tsunami risk models, and so are vital for land-use and emergency planning, performance-based engineering, as well as human and financial loss estimation. When selecting or developing risk models, a Tsunami Intensity Measure (TIM) (e.g. inundation depth, velocity, force estimates etc) must be selected which provides the best possible representation of the damage potential of the tsunami inundation. However, there is no consensus as to which flow parameter is the most appropriate TIM to estimate fragility. This paper presents a rigorous methodology using advanced statistical methods for the selection of the optimal TIM for fragility function derivation for any given dataset. This methodology is demonstrated using a unique, detailed, disaggregated damage dataset from the 2011 Great East Japan Earthquake and Tsunami, identifying the optimum TIM for describing observed damage for the case-study locations. Several advanced statistical methods are introduced which are novel in the fields of fragility analysis: multiple imputation for treatment of incomplete data-entries, semi-parametric Generalised Additive Models, and k-fold cross-validation for model comparison and optimization. The methodology presented in this paper has application for researchers and risk modellers in the engineering, DRR and insurance industries; and the statistical methods presented have implications for fragility function derivation and selection for multiple hazards.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket