Recent attention has been placed on the idea that dramatic community shifts represent shifts between alternative stable states (ASS), though robust evidence is limited. We tested predictions from ASS theory on two Panamanian reefs by quantifying positive feedback processes that could produce ASS. Using in situ assays of algal consumption, growth and nutrient limitation we evaluated two mechanisms (coral facilitation of herbivory and algal facilitation of nutrient availability) for potential positive feedbacks with community composition. Then, using those mechanisms, we developed a spatially explicit cellular automaton model of competition for space between coral and algae to evaluate the role of positive feedbacks in driving ASS dynamics.
Results/Conclusions
From the empirical tests we found that in a coral-dominated state herbivory rates were higher, algal growth lower, and nutrient limitation stronger than in an algal-dominated state. These results showed a positive feedback between community composition and herbivory as well as algal growth. Within the model the presence of positive feedback mechanisms alone drove divergence between community states and generated hysteresis by producing separate basins of attraction around coral-dominated and algal-dominated community states under a range of intermediate environmental conditions. Each state reached an abundance where it persisted with only minor fluctuations indicating positive feedbacks can be sufficient to stabilize each state. In addition, with positive feedbacks, coral and algae in the model would segregate to separate regions of the model space producing sharp boundaries. This accorded with previous theory and observational surveys of benthic community composition that showed reefs were segregated into distinct patches dominated by either coral or algae. As our model results show that positive feedbacks can be sufficient to produce and stabilize alternate community states, identifying such mechanisms should be considered a valuable initial diagnostic criterion for identifying possible ASS dynamics.