Contrasting effects of giant kelp on dynamics of surfperch populations

The effect of giant kelp, Macrocystis pyrifera, on the population dynamics of two temperate reef fishes, striped surfperch (Embiotoca lateralis) and black surfperch (E. jacksoni), was examined. Based on an understanding of how particular reef resources influence abundances of the surfperch and of the effect of giant kelp on those resources, we anticipated that Macrocystis would adversely affect populations of striped surfperch but would enhance those of black surfperch. The natural establishment of giant kelp at sites at Santa Cruz Island, California, resulted in the predicted dynamical responses of surfperch. Abundances of striped surfperch declined rapidly when and where dense forests of giant kelp appeared, but showed little change where Macrocystis was continuously absent over the 8 y period of study. Abundances of adult black surperch, which increased following the appearance of giant kelp, were lagged by >1 y because the dynamical response involved enhanced local recruitment. No change in abundance of black surfperch populations was evident at areas without giant kelp.The mechanism by which giant kelp altered the dynamics of the surfperch involved modification of the assemblage of understory algae used by surfperch as foraging microhabitat. Foliose algae (including Gelidium robustum) were much reduced and turf was greatly enhanced following the appearance of Macrocystis; these two benthic substrata are the favored foraging microhabitat for striped surfperch and black surfperch respectively. Populations of both surfperch species tracked temporal changes in the local availability of their favored foraging microhabitat. Thus, while neither species used Macrocystis directly, temporal and spatial variation in giant kelp indirectly influenced the dynamics of these fishes by altering their foraging base. These results indicate that the dynamics of striped surfperch and black surfperch were governed to a large degree by density-dependent consumer-resource interactions. The present work underscores the predictive value that arises from a knowledge of the mechanisms by which processes operate.