← Back to Sessions

Aggregate Behavior Analysis of Ground Motion Distributions and Their Effects on Loss Estimation

Description: 

The U.S. Geological Survey’s (USGS) ShakeMap product provides estimates of ground motions in areas directly affected by an earthquake. It does this by combining estimated ground motions, their uncertainties, recorded and observed ground motions, and a spatial correlation model. We investigated the uncertainties on the output hazard and loss estimation arising from the assumed input uncertain distribution parameters (mean and correlation). Our focus in this work is on investigating the aggregate behaviors of shaking (the relative proportions of grid cells having different shaking intensities) when we include the uncertainty in ground motions. We employed a Monte Carlo approach to model the uncertainties’ effects; wherein we generated numerous spatially correlated realizations of ground motions using a combination of circulant embedding and kriging. For each realization, we analyzed the relative number of grid cells having different shaking intensities and investigate the distribution of these relative counts across numerous realizations for various shaking intensities. A key outcome of our investigation is that the relative counts of ground motions given by the mean shaking grid underestimate the number of grid cells realized at the higher shaking intensities. That is, any given realization would mostly likely have a larger number of high shaking grid cells than the mean estimate). Hence, overall larger populations exposed to higher shaking. This in turn (on average) leads to higher loss estimates when shaking uncertainty is included. Understanding these aggregate shaking behaviors is important when incorporating shaking uncertainty into loss and hazard models that rely on aggregated shaking estimates.

Session: ShakeMap-related Research, Development, Operations, Applications and Uses

Type: Oral

Date: 4/19/2023

Presentation Time: 04:30 PM (local time)

Presenting Author: Davis T. Engler

Student Presenter: Yes

Invited Presentation: Yes

Authors