SPECIES INTERACTION STRENGTH: TESTING MODEL PREDICTIONS ALONG AN UPWELLING GRADIENT

Abstract

A recent model predicts that species interactions in benthic marine communities vary predictably with upwelling regimes. To test this model, we studied thePisaster–Mytilusinteraction at 14 rocky intertidal sites distributed among three oceanographic regions along a 1300‐km stretch of the U.S. West Coast. Regions included an intermittent‐upwelling region (northern), a persistent‐upwelling region (central), and a region of weak and infrequent upwelling (southern). We quantified predation rates by the sea starPisaster ochraceuson its main preyMytilus californianusby transplanting mussels into the sea star's low‐zone foraging range and comparing the rate of mussel loss in +Pisasterplots to those in −Pisasterplots. To evaluate the relation between predation rates and key ecological processes and conditions, we quantified phytoplankton concentration and rates of mussel recruitment, mussel growth, mussel abundance, and sea star abundance.Predictions of the model are expressed as responses of predator and prey abundance, and species interaction strength (per capita and per population or total impact at the population level). As predicted by theory,per capitapredation rates were independent of upwelling regime, with no variation with region. Contrary to expectation however, per‐population predation rates were similar between intermittent‐ and persistent‐upwelling regions but were greater under strong upwelling than under weak upwelling conditions. The greatest variation inper‐populationpredation rates was at the level of site within region. Also contrary to theory, average abundances of prey (mussel cover) and predators (sea stars) were similar among oceanographic regions and varied mostly at the site level.As expected from theory, predation rate was high where sea star density was high, a condition that often coincided with a high food supply (phytoplankton) for filter feeders, including larvae, and high recruitment. With the exception of two sites having dense sea star populations and thus high predation, low values of either or both were associated with low predation, suggesting that the supply of prey often depended on conditions that favored subsidies of both phytoplankton and new larvae to prey populations. The occurrence of high predator density and high predation at sites of low inputs of particulate food and propagules suggests that understanding sea star life history is a key to a fuller understanding of variation in predation on a coastal scale. Evidence suggests that often sporadic recruitment of sea stars along the coast is balanced by great longevity, which tends to even out predation impact on coastal intertidal communities.