Abstract
Resilience-based management aims to promote or protect processes and species that underpin an ecosystem's capacity to withstand and recover from disturbance. The management of ecological processes is a developing field that requires reliable indicators that can be monitored over time. Herbivory is a key ecological process on coral reefs, and pooling herbivorous fishes into functional groups based on their feeding mode is increasingly used as it may quantify herbivory in ways that indicate resilience. Here we evaluate whether the biomass estimates of these herbivore functional groups are good predictors of reef benthic assemblages, using data from 240 sites from five island groups in American Samoa. Using an information theoretic approach, we assembled a candidate set of linear and nonlinear models to identify the relations between benthic cover and total herbivore and non-herbivore biomass and the biomass of the aforementioned functional groups. For each benthic substrate type considered (encrusting algae, fleshy macroalgae, hard coral and turf algae), the biomass of herbivorous fishes were important explanatory variables in predicting benthic cover, whereas biomass of all fishes combined generally was not. Also, in all four cases, variation in cover was best explained by the biomass of specific functional groups rather than by all herbivores combined. Specifically: 1) macroalgal and turf algal cover decreased with increasing biomass of 'grazers/detritivores'; and 2) cover of encrusting algae increased with increasing biomass of 'grazers/detritivores' and browsers. Furthermore, hard coral cover increased with the biomass of large excavators/bio-eroders (made up of large-bodied parrotfishes). Collectively, these findings emphasize the link between herbivorous fishes and the benthic community and demonstrate support for the use of functional groups of herbivores as indicators for resilience-based monitoring.