Development of Global Social Vulnerability Model for Earthquakes

Abstract

Although earthquakes are one of the most devastating natural hazards that affect humanity, very few attempts have been made to consider the human elements that put populations at risk. Most researchers focus on the assessment of seismic hazard and the physical risk of a given area (i.e. potential for losses of life and infrastructure). But more integrated approaches that address both physical risk and social/human characteristics that place people at risk are needed to assess earthquake risk in a manner that is robust and holistic. Measurement of the concept of social vulnerability is one of the methods of addressing the human element of earthquake risk, which can be defined as characteristics or qualities within social systems that create the potential for loss or harm. There is no agreed-upon framework and established sets of data, however, to measure social vulnerability to earthquakes. It is within this context that the objective of this thesis is to take steps toward addressing the aforementioned area of opportunity by focusing on the human component of earthquake risk worldwide. The latter is accomplished by following a step-by-step methodology for producing composite indices representing the social vulnerability of countries to earthquakes within three topical areas (human impact potential, economic vulnerability, and recovery and reconstruction potential) that includes the development of a “wish list” of variables, missing data imputation, correlation analysis, and global as well as geographically weighted regression analyses for variable validation and selection. Three research questions guide this work: I) What metrics may provide the best comparative assessment of vulnerability to earthquakes from a societal perspective? II) To what extent do these metrics predict measurable outcomes from earthquakes, including property losses, casualties, and displacement? III) How does the predictability of metrics for measuring social vulnerability to earthquakes vary by world region? The results show that a number of different indicators can be used to assess social vulnerability considering the human impact, economic vulnerability and recovery potential, and these indicators vary across space. Using earthquake losses, fatalities, homelessness and total population impacted by earthquakes as a dependent variable in the regression analyses, the results also demonstrate that some of the variables such as housing and commercial building density have higher associations with earthquake impacts than other variables.