Emergent Spatial Patterns Can Indicate Upcoming Regime Shifts in a Realistic Model of Coral Community

Abstract

Increased stress on coastal ecosystems, such as coralreefs, seagrasses, kelp forests, and other habitats, can make themshift toward degraded, often algae-dominated or barren communi-ties. This has already occurred in many places around the world,calling for new approaches to identify where such regime shiftsmay be triggered. Theoretical work predicts that the spatial struc-ture of habitat-forming species should exhibit changes prior to re-gime shifts, such as an increase in spatial autocorrelation. However,extending this theory to marine systems requires theoretical modelsconnectingfield-supported ecological mechanisms to data and spatialpatterns at relevant scales. To do so, we built a spatially explicit modelof subtropical coral communities based on experiments and long-termdatasets from Rapa Nui (Easter Island, Chile), to test whether spatialindicators could signal upcoming regime shifts in coral communi-ties. Spatial indicators anticipated degradation of coral communi-ties following increases in frequency of bleaching events or coralmortality. However, they were generally unable to signal shifts thatfollowed herbivore loss, a widespread and well-researched source ofdegradation, likely because herbivory, despite being critical for themaintenance of corals, had comparatively little effect on their self-organization. Informative trends were found under both equilib-rium and nonequilibrium conditions but were determined by thetype of direct neighbor interactions between corals, which remainrelatively poorly documented. These inconsistencies show that whilethis approach is promising, its application to marine systems willrequire detailed information about the type of stressor andfill-ing current gaps in our knowledge of interactions at play in coralcommunities.