PNNL

Abstract

Extreme heat is intensifying worldwide, yet estimates of heat hazard and exposure inequalities depend on both the heat metric and how extreme days are defined. Using summer 2022 in the Mediterranean, we quantify population heat exposure with four metrics—land surface temperature (LST), air temperature (Ta), heat index (HI), and wet-bulb globe temperature (WBGT)—under absolute (fixed-value) and relative (anomaly-based) thresholds. Under absolute thresholds, total heat exposure differs by over two orders of magnitude across metrics (31.3 billion person-days for Ta vs 0.3 billion for HI). Geographic hotspots also diverge: WBGT concentrates in humid coastal North Africa (e.g., the Nile Delta), whereas Ta and LST are more widespread. Under relative thresholds, exposure totals converge and crossmetric hotspot agreement increases (e.g., the Ta–WBGT top-tercile overlap increases from 10.7% to 29.0%), shifting hotspots toward densely populated southern Europe. Crucially, the exposure–deprivation relationship also reverses across threshold frameworks: absolute thresholds concentrate exposure in more deprived North Africa and the Middle East, whereas relative thresholds shift the burden toward less-deprived but populous European cities. This sensitivity is decision-relevant: WBGT-based city rankings are almost completely reordered when switching threshold frameworks (Spearman ? = 0.01). Threshold choice thus shapes hotspot identification, supporting routine dual reporting of absolute and relative exposures for distinct heat-action objectives.