Network Measures of Ecosystem Complexity

We argue that the networks that can be constructed to represent ecosystems may inform us about the open-endedness of the evolutionary systems that underlie their dynamics. By adopting this approach we circumvent problems that arise from looking for open-endedness at the level of the organism, the more usual approach. We then examine various measures of ecosystem (niche web) complexity and propose a new informationtheoretic approach, Chaitin-Shannon Web Complexity. We compare its behaviour to that of the more common measures in ecology, in the light of common intuitions about complexity over a set of test networks and real ecosystem trophic webs. We also examine how the various measures accommodate intuitions about the complexity of these networks.