As with many taxa, the distribution of reef-building corals is expected to move poleward in a warmer ocean. High-latitude seas will never support the diversity of tropical reefs, but could act as refugia for some coral species. Identifying these refugia is therefore a conservation priority. Here, we present the first fully process-based, eco-evolutionary model for coral range shifts, with global coverage at 8km resolution. The model uses an empirical fitness landscape based on SST and pH (from an ensemble of CMIP6 models) and benthic light intensity, permits colonisation, and is coupled to a global larval dispersal model.
Despite being given no explicit information on the distribution of coral reefs, the model reproduces modern range limits and recent coral cover trends. In agreement with observations, our simulations suggest that increases in coral abundance at some high-latitude coral communities are largely due to the proliferation of existing local species, rather than tropical coral range expansion. Specialist high-latitude coral species may locally benefit from future warming but, in contrast to predictions from many habitat suitability models, significant tropical coral range expansion is unlikely. High-latitude seas are therefore unlikely to buffer tropical corals against the catastrophic loss of diversity predicted over the coming century.