Every year, flooding causes more than $32 billion in damages to American communities. In the upcoming years, experts predict that figure will rise as climate change causes storm events to become more severe and unpredictable: According to forecasts, by 2050, the risk of flooding will rise by over 26%. Metropolitan areas with higher populations of Black, Indigenous, and People of Colour (BIPOC) people are disproportionately affected by flooding due to socioeconomic disparities in flood risk mitigation.
In this context, interpretable machine learning is used by Nadja Veigel and associates to enhance comprehension of the effectiveness of flood resilience tactics. 400 behavioral and socioeconomic factors that impact disaster response and mitigation were chosen by the team to construct the machine learning model. They included top-down policy initiatives like community-level policies as well as bottom-up, household-level initiatives like insurance or property upgrades.
The National Flood Insurance Programme and the American Community Survey of the United States Census Bureau provided open-access flood insurance data to the authors.
The findings demonstrated that following significant floods, the majority of households purchase flood insurance. Therefore, the likelihood of uninsured residents is higher in communities that are not frequently or heavily exposed to damaging flooding. The authors also point out that low awareness of flood history in urban areas due to high population turnover impedes mitigation and readiness initiatives. Adoption of insurance is also generally lower in urban areas.
On the other hand, proactive measures such as the Community Rating System (CRS) of the National Flood Insurance Programme are implemented at the local level. By reducing insurance premiums for communities that implement mitigation and floodplain management strategies, the CRS promotes the adoption of insurance. The authors propose that by focusing on underprivileged, at-risk communities, the system could more successfully address the inequality associated with flood risk.
The study confirms earlier findings that communities at risk are more likely to experience flooding on a regular basis and could gain from greater resilience. Based on data, top-down strategies—like the CRS—provide proactive flood solutions that aid in addressing structural disparities in risk. While flood insurance is still a vital risk management tactic, it is frequently a reactive measure that offers little assistance on its own when combined with neighborhood-level initiatives.
