SHEDIS: Linking Subnational Hazard and Exposure information with DISaster impact records for international risk analysis

Understanding the drivers of disaster outcomes and identifying hotspots of social vulnerability requires datasets that integrate societal impacts, physical hazards, and human exposure. However, widely used international disaster databases, such as the Emergency Events Database (EM-DAT), often lack detailed information on hazard characteristics and population exposure. This limits their utility for comprehensive risk assessments and interdisciplinary research.

We present SHEDIS, an open-access family of datasets addressing this gap by linking disaster impact records from EM-DAT with subnational data on hazard metrics, human exposure, and disaster locations. The first module, SHEDIS-Temperature, focuses on temperature-related disasters occurring from 1979 to 2018, encompassing 382 events across 2,836 subnational locations in 71 countries. This dataset provides high-resolution hazard metrics derived from 0.1°, 3-hourly meteorological data, including absolute indicators such as apparent temperature (accounting for humidity and wind) and percentile-based thresholds to identify extreme temperature events. Population exposure is quantified using annually interpolated population maps, with metrics such as person-days of exposure to hazardous temperatures. Outputs are aggregated at both the impact record-level and administrative unit-level, offering flexibility for varied analytical needs.

Future expansions of SHEDIS will incorporate additional hazard types, further supporting global-scale risk assessments and practical applications. By providing detailed, subnational hazard and exposure data linked to disaster impacts, SHEDIS enables more nuanced analyses to advance international disaster science, inform resilience strategies, and contribute to disaster risk reduction.