An Experimental Test of Multi-species Markov Models: Are Barnacles Long-term Facilitators of Mussel Bed Recovery?

Abstract

Predicting responses of complex ecosystems to environmental impacts is challenging because of the web of species interactions and their potential to generate indirect effects. Multispecies Markov chain models provide one potential predictive framework for complex ecosystems. A previously parameterized Markov model of a rocky intertidal community predicted little effect of removing acorn barnacles from the system on sessile species composition. In contrast, short-term observations of recovery following disturbance suggest that acorn barnacles play a key facilitation role in mussel-bed recovery. In a 6-yr experiment in which balanoid acorn barnacles [Balanus glandula Darwin, 1854 and Semibalanus cariosus (Pallas, 1788)] were chronically and selectively removed, final species composition did not differ significantly from that of model predictions (90.3% of mean composition explained, P > 0.5), and the mussel Mytilus californianus Conrad, 1837 was able to attain its typical dominance. These results demonstrate that multispecies Markov chain models can generate reliable predictions and that acorn-barnacle facilitation does not have a substantial effect on long-term community composition in this system.