Describing herbivorous coral reef fish assemblages across multiple spatial scales

Background/Question/Methods

Describing and predicting ecological communities depends on the scale at which the community is observed. Recent advances in technology permitting the collection of large-scale, georeferenced data underwater now enable us to assess the factors driving community state at different scales in subtidal marine systems. Coral reefs have been shifting from coral-dominance to macroalgal-dominance in many locations, thus it is important to understand how changes in reef state influence the structure of the overlying fish community relative to static factors, such as geographic location. We used novel methodology to survey coral reef fish assemblages and associated benthic communities in the lagoons of Moorea, French Polynesia. We layered these data with a suite of other predictor variables that include physical metrics calculated from bathymetry and geographic variables, such as distance offshore, to construct a robust set of predictor variables to model herbivorous fish biomass, a group of ecological and economic importance in this system. The unprecedented scale of these observations, ~225,000 m2 distributed around the island, permit us to assess how predictable reef herbivore communities are at six different spatial scales and to determine how the most important variables used in these predictions varied based on the scale observed.

Results/Conclusions

We found that as the spatial scale increased (ranging from 50 m2 to 1500 m2 survey units), the deviance explained more than doubled, indicating that reef herbivores become more predictable at larger spatial scales. Additionally, we found that the most important variables in these predictions typically changed from those describing the geographic location of the surveys to those describing the availability of physical reef structure as the scale increased. Reef state metrics (i.e., coral vs. macroalgal cover) were important variables across scales. These results provide key insights into how herbivorous reef fishes are structured across spatial scales that are not adequately captured using traditional techniques. Understanding the relative impacts of these different predictor variables, which include metrics that are impervious to changes in reef state (such as geographic location) and those that can change due to continued anthropogenic impacts (such as the cover of biological substrate), will provide greater insight into how the distribution and abundance of herbivorous fishes may change as live coral continues to decline globally due to anthropogenic impacts such as climate change and overfishing.