Development and application of mass-balanced ecological network models for kelp forest ecosystems

California kelp forests are highly productive and species rich ecosystems. However, ecosystem-wide consequences of fishing higher tropic levels (fishes) and the effect of climate on primary producers such as the giant kelp, Macrocystis pyrifera, are not well understood. I develop and apply mass-balanced ecological network models, Ecopath with Ecosim, to explore separately how fishing and the dynamics of giant kelp influence ecosystem functions (e.g., species interactions, biomass dynamics), structure (e.g., the distribution of biomass density among species or species groups) and their dynamics. Faced with the difficulty of synthesizing information required to construct these models, I develop and apply an online database (http://kelpforest.ucsc.edu/) to facilitate the accessibility of such information. It is the first online database designed specifically to inform development of ecological network models. To explore ecosystem-wide effects of fishing in giant kelp forests, I examine (i) the extent to which changes in species interactions and biomass of nodes caused by fishing extend across the ecological network, how these changes vary with (ii) levels of fishing mortality, (iii) fishing of six different species of fishes, and (iv) when all six species are fished simultaneously. Results suggest that fished species differ markedly in the extent to which species interactions and biomass densities are altered across the ecosystem and these responses vary with different levels of fishing mortality. I also used the models to predict ecosystem-wide responses to different dynamics of giant kelp biomass. I test the hypotheses that different scenarios of dynamics of giant kelp biomass will influence (i) total network biomass, (ii) distribution of biomass density across nodes, (iii) temporal variation in biomass density of nodes, and (iv) how this variation differs among trophic levels. Results suggest that both the mean and the variability of giant kelp biomass alter the direction and magnitude of change in total network biomass. Variation is greater for lower trophic levels. Although all inferences of these models are based solely on trophic interactions, they illustrate the value of ecosystem models to generate hypotheses and predictions of ecosystem responses to one or more changes in kelp forests