Where scale matters for social vulnerability indices: a multiyear analysis of the United States

Abstract

Social vulnerability indices have found widespread acceptance and application in planning and risk management. However, though a valuable tool for addressing equity concerns, prior studies have demonstrated shortcomings of such indices in terms of their reliability and validity. While index uncertainty associated with level of data resolution has been thoroughly examined, the impact of varying geographic extents has received less attention, and index consistency across the entire United States and over time has not been previously studied. In this study I examine variability associated with changing data resolution and geographic extent for four years of social vulnerability indices for the United States, quantifying inconsistencies across 32 index variants, and helping to explain why these inconsistencies occur. I find that changes in data resolution and extent lead to vulnerability designation switching in nearly 22 % of index variant comparisons between 2019 and 2021. In addition, at least 50 % of block groups are subject to at least one instance of vulnerability designation switching when comparing across all index variants. The results suggest that index inconsistencies are partly the result of spatial non-stationarity in vulnerability processes and confirm that prior findings for case studies apply across the entire U.S. In addition, I find that rural and tribal areas are prone to higher variability in index values, and identify13,872 block groups where social vulnerability metrics are relatively consistent. Overall, this study reinforces the evidence suggesting that social vulnerability indices constructed with different scales of data can produce inconsistent patterns in relative vulnerability measures and supports calls for caution in their application for decision-making.

Kate Nelson

Associate Professor, SCALes PI

My research interests include landscape diversity, agricultural adaptation, strategic retreat, vulnerability assessment, and scaling relationships.