RESUMO
With rising global temperatures, cities increasingly need to identify populations or areas that are vulnerable to urban heat waves; however, vulnerability assessments may run into ecological fallacy if data from different scales are misconstrued as equivalent. We assess the heat vulnerability of 1983 residents in Vienna by measuring heat impacts, exposure, sensitivity and adaptive capacity with mirrored indicators in the mapping paradigm (i.e. census tract data referring to the geographic regions where these residents live) and the surveying paradigm (i.e. survey data referring to the residents' individual households). Results obtained in both paradigms diverge substantially: meteorological indicators of hot days and tropical nights are virtually unrelated to self-reported heat strain. Meteorological indicators are explained by mapping indicators (R2 of 15-40 %), but mostly not by surveying indicators. Vice versa, experienced heat stress and subjective heat burden are mostly unassociated with mapping indicators but are partially explained by surveying indicators (R2 of 2-4 %). The results suggest that the two paradigms do not capture the same components of vulnerability; this challenges whether studies conducted in the respective paradigms can complement and cross-validate each other. Policy interventions should first define which heat vulnerability outcome they target and then apply the paradigm that best captures the specific drivers of this outcome.
