Disparate social impacts by satellite and self-report: Extreme precipitation and reported flood impacts in Pernambuco, Brazil

• In Event: 41065 - Population and Environment

Abstract

How are impacts of extreme precipitation events socially stratified? Using the case of extreme-precipitation-driven floods in Pernambuco, Brazil in 2022, we demonstrate that the answer varies considerably depending on data source. We integrate satellite-derived precipitation data from the Climate Hazards group Infrared Precipitation with Stations (CHIRPS) with household disruption data from the Demographic Consequences of Zika and COVID-19 Crises (DZC) Survey. We demonstrate that extreme precipitation exposures display disparate impacts by race, religion, and maternal education. However, these disparities largely dissipate after adjusting for city-level location, highlighting the dominance of geospatial distribution differences across population subgroups in driving these disparities. In contrast, household-survey-reported flood impacts remain strongly associated with economic status and children in the household after adjusting for \textit{both} geographic location \textit{and} measured precipitation, highlighting the importance of household reports for revealing highly localized vulnerabilities. Further analyses demonstrate a sharper dose-response relationship between precipitation and reported impact among the poorest households and implicate highly localized infrastructures in the discrepancy. Results suggest that measured-exposure-based risk assessments may systematically underestimate the disproportionate impact of climate disasters on marginalized populations subgroups, and that equity-oriented preparedness requires attention to both their spatial distributions and the highly localized infrastructures to which they have access.

Authors

• Mohamad AlAbbas, University of Pennsylvania

• Leticia Marteleto, University of Pennsylvania

• Emily Carroll Hannum, University of Pennsylvania