River food webs inhabitant watersheds with branching (or dendritic) geometries. Recent modeling and empirical work has suggested strong effects of branching geometries on processes such as population persistence, genetic structure, and patterns of species diversity. In addition, actual river networks show influences of branching structure on population dynamics and mechanisms driving community assembly, with different mechanisms sometimes operating in different areas of the network. However, most work that explicitly considers branching network structure has focused on within taxa or trophic-level dynamics. In contrast, food webs in rivers are known to be strongly influenced by processes that may interact with network structure and position. We explore how dynamics of species-rich food webs are influenced by the branching nature of river networks. Simulated food webs are constructed using a size-based “niche model” paired with consumer-resource models where vital rates are determined by empirical allometric relationships. Species within the food web are then allowed to disperse among habitat patches that are arranged in a branching structure. Dispersal is modeled to vary among species in a number of ways, including randomly, allometrically, and with competition-colonization tradeoffs. Dispersal is also examined across gradients of with-in versus out-of-network movement constraints.
Our models are capable of producing a wide array of dynamics, including stability, dramatic fluctuations, and extinctions. Branching network structure and dispersal mode influences food web stability, yet intrinsic features of the food web have a strong interaction effect with these. Food webs with certain intrinsically destabilizing structures show greater sensitivity to influence by spatial effects, with specific network geometries and dispersal modes largely playing a secondary role in shaping dynamics. Species-rich food webs with destabilizing energy flow patterns can be stabilized regionally by dispersal; dendritic structure and dispersal traits can enhance this asynchrony. Small, highly stable food webs show significantly less network or dispersal influence. We conclude by discussing the potential of mechanistic food web models to inform management of freshwater systems under global environmental change.