A decision support tool for rubble stabilization on coral reefs

Loose, persistently unstable rubble beds are increasingly common on coral reefs due to climate-driven disturbances and human impacts. These rubble beds often act as bottlenecks to reef recovery and typically do not stabilize without intervention. Rubble stabilization is used globally to restore degraded reefs, but there has been limited synthesis of its effectiveness across methods and environments. A growing need exists to translate current knowledge into practical guidance to support management and improve restoration outcomes. Bayesian Belief Networks (BBNs), useful for modeling complex systems with substantial uncertainty, were applied in this study to support decision-making in rubble stabilization. The model integrates expert knowledge and global data on restoration outcomes over time and in various environments. This study compares coral cover benefits of stabilization methods, including Reef Bags, flat meshes and grids, elevated frames, and solid structures, with and without coral outplants. Benefits were defined as the difference in coral cover between restoration and control sites, attributable to outplanting or natural recruitment. All methods led to increased benefits over time, though outcomes varied in magnitude and timing. Methods that included outplants generally produced earlier gains in coral cover, but benefits were density-dependent and longer-term outcomes remain uncertain. Higher coral cover benefits were associated with smaller rubble pieces, steeper slopes, and stronger hydrodynamic forces, provided stabilization structures were well anchored and maintained, which is often challenging in practice. Improved long-term monitoring, consistent terminology, and data sharing will be critical to strengthen model reliability and support decision-making in the global reef community.