A sustainability-based framework for assessing dam failure impacts: Integrating AI-assisted indicator extraction and application to the Fundão case

Dam failures generate long-lasting environmental, social, and economic consequences that challenge sustainable recovery and governance. Existing assessment approaches typically focus on single domains or short-term impacts, limiting their ability to capture interdependencies and long-term effects. This study develops a Sustainability-Based Dam Failure Impact Framework (SDFIF) to systematically evaluate short- and long-term consequences across environmental, social, and economic pillars. The framework integrates a PRISMA-guided literature review (88 sources), AI-assisted indicator extraction, and hierarchical classification, resulting in 328 validated indicators. The framework was applied to the 2015 Fundão tailings dam failure, revealing that social and economic impacts (normalized scores >0.9) persist beyond partial ecological recovery, indicating sustained livelihood disruption, cultural loss, and community vulnerability. These findings highlight the importance of incorporating temporal persistence and cross-domain interactions in post-disaster assessment. The SDFIF provides a transparent and adaptable tool to support recovery monitoring, regulatory evaluation, policy development, and sustainability-oriented risk management in dam governance.