Dynamical effects of weak trophic interactions in a stochastic food web simulation

Abstract

Network models are traditional in community ecology. For example, they provide a rich analytical toolkit to put higher predators into a multispecies context. Better understanding their top-down effects and the potential bottom-up control on them would be of key importance for predictive ecosystem management. Food web architecture may be used to predict community dynamics, but it is an old question how reliable are the studies considering only static information. A general and intuitive assumption is that stronger links (with larger weights) mediate stronger effects. We study this statement and use an illustrative case study. We investigate the trophic structure of the Prince William Sound food web in terms of biomass flows, and study its simulated dynamics in a stochastic modelling framework. We aim to understand bottom-up effects of preys on consumers: we focus on the fluctuations of top predator populations, following disturbance on their prey. Several disturbance regimes are studied and compared. Food web structure and link weight generally predict well the average impacts of preys on top-predators, with larger flows mediating stronger effects. Most exceptions appear for weak links: these are less predictable, some of them can be surprisingly important.