Topological keystone species in ecological interaction networks:: Considering link quality and non-trophic effects

There is increasing evidence that non-trophic interspecific interactions play an at least as important role in community dynamics as trophic relationships. More and more studies on pollination, mutualism and facilitation are published but these effects are interpreted more like alternative explanations than being synthesized with results of trophic analyses. Here, we construct and analyze the interaction web of the well-studied Chesapeake Bay mesohaline ecosystem. By interaction web we mean a food web completed by a carefully selected set of non-trophic links. We quantify the interaction structure of the web and the positional importance of nodes by different network indices. We perform the suitable analyses for different variants of the network: combinations of direction, sign and weights, as well as considering also non-trophic links result in a set of webs of different information content. We also create a semi-quantitative variant of the web, in which only the order of magnitude of the mass flows are considered. The appropriate network indices for each web variant are calculated and compared. Finally, however our paper is primarily of methodological nature, we present some findings about the fish community of the Bay. We suggest that the multiple techniques presented here, adapted even from social network analysis, can help field conservation efforts by suggesting optimal preferences for data collection. (c) 2006 Elsevier B.V. All rights reserved.