Distributed control in ecological networks

Understanding how "control" is exercised in ecological systems, even giving a more appropriate definition or meaning to the word "control" in this context, is an important theoretical issue and would increase our ability to manage ecosystems. Conventionally, in food web ecology, the distinction is drawn between bottom-up and top-down control. In that literature, the bottom-up hypothesis asserts that the primary producers are the source of system regulation and the top-down hypothesis states that keystone species at a higher trophic level can regulate the system. However, we know that in reality control of system behavior is much more complex and distributed than this dichotomy would suggest. Indeed, there is an urgent need for a succinct, yet more complete and comprehensive conceptual framework for thinking about control and for deriving insights into what governs ecosystem organization. In an ecosystem, each element contributes to the overall flow-storage pattern observed in the system through its interactions with the other elements; in this sense, control is distributed among the system elements. Those pair-wise system interactions can be identified and quantified using network analysis. Since the network analysis methodology accounts for both the input (recipient-oriented) and output (donor-oriented) influences from each element, it is possible to use this methodology to move beyond the simple top-down and bottom-up perspective of control. Here, I connect the network analysis methodology to traditional control theory, reintroduce a network-based control parameter using flow analysis and extend the methodology to network storage analysis. Model ecosystems are constructed and used to investigate these properties. (C) 2004 Elsevier B.V. All rights reserved.