Motivation

Reservoir sedimentation is an urgent global challenge that directly threatens long-term water security. While dams and reservoirs provide essential services—such as freshwater supply, hydropower generation, irrigation, and flood protection—their storage capacity is steadily eroded as trapped sediment accumulates. Unlike structural deterioration, which unfolds over many decades, sedimentation can rapidly reduce a reservoir’s operational lifespan, disrupt hydropower efficiency, increase flood risks, and hinder downstream sediment delivery critical for sustaining ecosystems and deltas. Recent studies indicate that global reservoir capacity has already peaked, and per capita storage is in decline, underscoring the widespread nature of this problem. Despite its global significance, reliable assessment of reservoir sedimentation remains limited. Direct field-based methods are accurate but costly and difficult to scale, while indirect modeling and remote sensing approaches often rely on sparse or highly localized data. Until recently, the lack of a unified, open-source global dataset on sedimentation prevented the development of scalable, data-driven solutions. To address this gap, RECLAIM builds upon the newly available Global Reservoir Inventory of Lost Storage by Sedimentation (GRILSS) dataset, providing the first open-source Python package designed to analyze, model, and predict reservoir sedimentation loss using machine-learning approaches. By making sedimentation assessment more accessible and scalable, RECLAIM supports sustainable reservoir management and informed water infrastructure planning.