Spatiotemporal measurement and econometric modeling of resilience dimensions

Loading...
Thumbnail Image
Date
2018
Journal Title
Journal ISSN
Volume Title
Publisher
University of Alabama Libraries
Abstract

This dissertation presents a systems-based empirical modeling framework based on three dimensions of resilience described by the theoretical resilience curve to determine disaster, community, and natural environment factors which affect the resilience dimension measurements. The three dimensions include robustness, or ability to retain system functionality when subjected to an extreme event, rapidity, or the time that it takes to recover system functionality, and functionality change, the difference in pre-event functionality and the functionality at a discrete time interval. The resilience models are created using econometric frameworks and built using spatiotemporal data sets, with data collected from public resources and acquired through deep learning procedures conducted on publicly available imagery. The Resilience Robustness Model (RRoM) is constructed with a comparison of discrete choice model frameworks to find correlation between measured damage states and community characteristics and to validate a transferable model structure. The Resilience Rapidity Model (RRaM) is constructed using an ordered probit model to determine likelihood of recovery within specified time intervals. The functionality change model is constructed with a binary discrete choice model to determine likelihood of recovery at a specific time interval. The model framework was tested on the system of community buildings affected by the 2011 Tuscaloosa Tornado. In addition to the modeling framework, a methodology to rapidly collect, store, disseminate, and analyze post-event and longitudinal data is presented. These tools are developed to facilitate the creation of deep learning algorithms to automatically measure damage states and recovery patterns as well as create variables to enhance the resilience measurement models. The synthesis of enhanced data collection and information extraction fueling empirical resilience models will aid decision-makers in the prioritization of resilience goals in community planning.

Description
Electronic Thesis or Dissertation
Keywords
Civil engineering
Citation