Indices of Social Vulnerability to Hazards: Model Uncertainty and Sensitivity
Advisor: Dr. Susan L. Cutter
Social vulnerability indices have emerged over the past decade as an approach to quantitatively measure the social dimensions of natural hazards vulnerability. Validation of the metrics with external reference data has posed a persistent challenge in large part because social vulnerability is not directly observable. This research applies global uncertainty and sensitivity analyses to internally validate the methods used in the most common social vulnerability index designs. Global uncertainty analysis is performed to assess the robustness of index ranks when reasonable alternative index configurations are modeled. Global sensitivity analysis is employed to evaluate which index construction decisions have the greatest influence on the output rank variability. The research is guided by the following questions:
- What is the uncertainty associated with social vulnerability index ranks?
- What is the spatial relationship between social vulnerability and uncertainty?
- Which modeling decisions contribute the most to variability in index ranks?
- Do the uncertainty and sensitivity analyses findings vary by study area location?
All of the evaluated social vulnerability indices were imbued with a high level of epistemic uncertainty, which results from subjective decisions made during each phase of index construction. This uncertainty was represented cartographically on output social vulnerability maps. Spatial autocorrelation between social vulnerability and epistemic uncertainty was found to be low. The most common index designs were replicated, with hierarchical models found to be the most robust, and inductive models the least.
The variation in deductive vulnerability index ranks is most sensitive to the choice of transformation method, hierarchical models to the selection of weighting scheme, and inductive indices to the indicator set and scale of analysis. The findings demonstrate why global uncertainty and sensitivity analyses should be a standard element in social vulnerability index development. Specific recommendations for each stage of index construction are provided so that the next generation of social vulnerability indices can be developed with a greater degree of robustness and transparency.Dissertation