Abstract
Individual metabolism, predator-prey relationships, and the role of biodiversity are major factors underlying the dynamics of food webs and their response to environmental variability. Despite their crucial, complementary and interacting influences, they are usually not considered simultaneously in current marine ecosystem models. In an attempt to fill this gap and determine if these factors and their interaction are sufficient to allow realistic community structure and dynamics to emerge, we formulate a mathematical model of the size-structured dynamics of marine communities which integrates mechanistically individual, population and community levels. The model represents the transfer of energy generated in both time and size by an infinite number of interacting fish species spanning from very small to very large species. It is based on standard individual level assumptions of the Dynamic Energy Budget theory (DEB) as well as important ecological processes such as opportunistic size-based predation and competition for food. Resting on the inter-specific body-size scaling relationships of the DEB theory, the diversity of life-history traits (i.e. biodiversity) is explicitly integrated. The stationary solutions of the model as well as the transient solutions arising when environmental signals (e.g. variability of primary production and temperature) propagate through the ecosystem are studied using numerical simulations. It is shown that in the absence of density-dependent feedback processes, the model exhibits unstable oscillations. Density-dependent schooling probability and schooling-dependent predatory and disease mortalities are proposed to be important stabilizing factors allowing stationary solutions to be reached. At the community level, the shape and slope of the obtained quasi-linear stationary spectrum matches well with empirical studies. When oscillations of primary production are simulated, the model predicts that the variability propagates along the spectrum in a given frequency-dependent size range before decreasing for larger sizes. At the species level, the simulations show that small and large species dominate the community successively (small species being more abundant at small sizes and large species being more abundant at large sizes) and that the total biomass of a species decreases with its maximal size which again corroborates empirical studies. Our results indicate that the simultaneous consideration of individual growth and reproduction, size-structured trophic interactions, the diversity of life-history traits and a density-dependent stabilizing process allow realistic community structure and dynamics to emerge without any arbitrary prescription. As a logical consequence of our model construction and a basis for future studies, we define the function Φ as the relative contribution of each species to the total biomass of the ecosystem, for any given size. We argue that this function is a measure of the functional role of biodiversity characterizing the impact of the structure of the community (its species composition) on its function (the relative proportions of losses, dissipation and biological work).