A Florida urban heat risk index: Assessing weighting and aggregation approaches

Abstract

Extreme heat is a burgeoning public health concern facing cities, thus composite indicators (CI) are useful identifying at-risk populations and directing resources. Nonetheless, heat risk assessments regularly overlook fundamental CI components like weighting, aggregation, and exploring links to other models. This analysis examined how weighting and aggregation alter CI scores, associated spatial distribution, and performance. Models were validated with heat-related mortality data. The Florida urban heat risk index (FUHRI) demonstrated how weighting and aggregation influence model outputs. Applying uncommon statistical weights performed better than traditional approaches, and multiplicative aggregation outperformed conventional additive aggregation. Results support further exploring non-traditional methods to enhance extreme heat risk assessment while underscoring importance of comparing multiple CI. Despite FUHRI score and performance variation, there was often agreement about at-risk geographies; however, consistency does not guarantee accuracy identifying areas needing adaptation and mitigation. Decision makers should use caution interpreting CI unless validation analysis demonstrates adequate performance, and even then, models must be continuously refined using the most recent public health data and desirable methods available.