Multi-scale collapse of coral cover under climate change

Projecting ecosystem trajectories under future climates is critical for conservation planning, yet remains constrained by uncertainty arising from limited data, ecological complexity, and biological and environmental variability. Variability, when disentangled from uncertainty, offers critical insights into population and community dynamics. For example, enhanced vital rates (growth, survival, fecundity) can enable faster population recovery, while environmental variability provides spatial refuges and ecological niches. Here, we integrated vital rate data from ~10,000 coral colonies across environmental gradients on the Great Barrier Reef (GBR) into a metacommunity model to project coral trajectories to 2100. Variability in coral-cover trajectories was explored as a function of (i) vital rates, (ii) spatial scales (from within-reefs to across regions), and (iii) future climate scenarios (represented by three Shared Socioeconomic Pathways, SSPs, informed by an ensemble of downscaled Global Climate Models, GCMs). Climate scenarios explained the greatest variance in coral-cover trajectories, with all spatial scales (across sites, reefs and regions) contributing substantial additional variance. Coral cover declined over the century under all emissions scenarios except SSP1-2.6 (warming <2°C), despite modelling increasing population thermal tolerance through natural selection. This decline was accompanied by a drastic loss of variability in coral-cover trajectories across all spatial scales, signalling reduced scope for natural adaptation, traditional management and restoration. By the end of the century, variability across GCMs also declined, leading to higher certainty about long-term reef state, i.e. coral collapse, than certainty about the near-term, should global emissions follow moderate-to-high pathways.