Is flow velocity important in tsunami empirical fragility modeling?

The influence of flow velocity on structural damage induced by tsunami inundation is investigated to improve empirical fragility models considering flow velocity as explanatory hazard variable in addition to inundation depth. The analysis is based on extensive tsunami damage data for the 2011 Tohoku earthquake collected by the Ministry of Land, Infrastructure, and Transportation of the Japanese Government. Multivariate tsunami fragility curves are developed through multinomial logistic regression of un-binned data. This approach facilitates the flexible development of various nested models considering inundation depth alone or inundation depth and velocity altogether. Statistical diagnostic metrics, such as the Bayesian Information Criterion, the Akaike Information Criterion, and the residual deviance, are used to determine which model improves the predictability of tsunami damage. The significance and importance of including flow velocity in the vulnerability models are assessed quantitatively by examining the influence of different spatial resolutions in elevation model and different source models. Then, the effects of coastal topography have been investigated. Numerical results show that flow velocity generally improves the fragility models, particularly for severer structural damage states, and that it is important for sites along the coast where the inundation depth is not extremely high. Coarse digital elevation model and inaccurate source models have influence on the calculated values of flow velocity and thus they affect the accuracy of fragility modeling. Finally, two different fragility models are calibrated for plain-type and ria-type coasts by reflecting differences in hydrodynamic behavior and recorded damage on the structures.