Color and degree of interspecific synchrony of environmental noise affect the variability of complex ecological networks

The fundamental debate on complexity–stability relationships in ecosystems is becoming more important in the face of dramatic contemporary changes. Most research using network approaches to address this challenge has focused on predator–prey interactions and static networks. We describe a stochastic ecological network simulation model that combines predator–prey interactions with mutualism, parasitism and basal species competition. Network topology is based on a generalized niche model, while temporal dynamics are specified using bioenergetic models with noise terms. We use the model to explore the effects of non-predator–prey interactions, color of environmental noise, and the degree of synchronization of environmental noise across species on the coefficients of variation of total ecosystem energy content and Shannon entropy over time. In regression trees derived by binary recursive partitioning, the presence or absence of synchrony gave the greatest difference in the means of data points for both system energy and Shannon entropy, followed in turn by white versus colored noise and pink versus red and black noise. The effects are present despite the large variation in the Monte Carlo simulations reflecting the variability of real ecosystems. Non-predator–prey interactions explained relatively small proportions of the total deviance, and each had different effect signs depending on the presence of the other interaction types and on whether energy or entropy was measured. These results underscore the importance of modeling more realistic colors of environmental noise in understanding and predicting the dynamics of food webs and ecological communities.