Flooding is a major global hazard affecting thousands every year, however existing flood databases are limited in their capacity to support flood prediction models and disaster risk management professionals due to data fragmentation, inconsistent inclusion thresholds, limited spatial information, and biased data sources such as news reports. This study addresses these limitations through two primary objectives, first by creating a global flood database for the period following the launch of the Surface Water and Ocean Topography Satellite (SWOT) mission (December 2022 - February 2026), and second by analyzing bias in news reporting of flood events in socially vulnerable areas in the contiguous United States. The integrated database combines flood events reported by EM-DAT, GDACS, Copernicus EMS, and USGS. The resulting database contains 4847 flood events; the highest concentration of floods is found in the U.S due to the high density of USGS stream gage data. The assessment of reporting bias in socially vulnerable communities in the U.S. used an LLM to pair flood events with flood reports, and a negative relationship was found between social vulnerability and flood reporting. This result suggests that flood events in socially vulnerable communities are underreported, and that databases built solely using news reports of flood events likely underrepresent these communities.
Katherine Fear (Tue,) studied this question.