Spatial variation in keystone predation on foundation species drives local and landscape- scale diversity patterns

Christopher D. G. Harley, Amelia Virginia Hesketh, Alyssa-Lois Gehman, Christopher Neufeld

The proposition that keystone predators increase biodiversity has been criticized for failing to consider the species facilitated by, rather than outcompeted by, dominant prey. Here, we revisit the relationship between the original keystone species, Pisaster ochraceus, and the diversity of rocky intertidal communities inclusive of competitively dominant prey, competitively
subordinate space-holders, and the many taxa facilitated by both. Using a system of beach-bounded rocky headlands, we document landscape-scale variation in predation pressure, with lower Pisaster densities and lower mid-intertidal mussel consumption rates along landward segments of headlands where low intertidal rocky refugia for sea stars were unavailable due to
the higher level of the sandy beach surface. Landward (low predation) segments hosted expansive mid-intertidal beds of the competitively dominant mussel Mytilus californianus, which were rare or absent on seaward (high predation) segments. Compared to consistent dominance of M. californianus on landward segments, seaward segments were dominated by a patchwork of habitat-forming barnacle and seaweed species, which resulted in higher within-plot (alpha) and among-plot (beta) diversity of primary space-holders. Of these various biogenic habitat types, mussel beds hosted the greatest number of unique taxa. However, the total richness was similar between landward and seaward segments because pronounced among-plot variation in the identity of competitively inferior foundation species led to higher beta diversity for the entire community where predation was high and mussels were rare. When M. californianus beds were compared to beds of the barnacle Semibalanus cariosus, which often dominated high-Pisaster areas, we found that both epizoic and total richness were higher for S. cariosus beds at the plot level (alpha diversity), but that beta diversity was higher within M. californianus beds when all associated species were considered. Our results suggest that relationships between keystone predators and biodiversity depend on the spatial scale studied and the identity and diversity of foundation species that replace competitive dominants. Because many taxa may preferentially or even uniquely associate with either competitively dominant or competitively subordinate habitat-forming species, total biodiversity will be maximized not “with” or “without” a keystone predator, but when predation pressure varies across a landscape.