Ground Motion Estimates Updated by Building Damage

Abstract

Classical ShakeMap relies on a conditional Gaussian random field approach to earthquake shaking estimation, where the mean and covariance matrix are derived from a ground motion model, a correlation model for intra-event residuals, and the recordings from seismic monitoring stations. In urbanized areas, seismic damage to buildings represents an uncertain measure of earthquake shaking. If a seismic fragility model is available, the surveyed damage can be used to update the ShakeMap in a probabilistically consistent manner. The study shows, via ideal cases first, how much damage information can change the estimate of the shaking intensity and the related uncertainty, generally reducing it when compared to the classical ShakeMap, and how the damage-informed ShakeMap tends to get closer to the true shaking intensity. The case in which errors in damage surveys are possible is also considered. Finally, the classical and damage-updated ShakeMap are compared in the real case of the 2009 L'Aquila earthquake in central Italy, where tens of thousands of buildings were hit, leading to conclude that damage information, when available, should not be neglected for earthquake shaking estimation.