Sampling effects and the robustness of quantitative and qualitative food-web descriptors

Food-web descriptors serve as a means for among-web comparisons that are necessary for the discovery of regularities in respect to food-web structure. Qualitative descriptors were however found to be highly sensitive to varying levels of sampling effort. To circumvent these shortcomings, quantitative counterparts were proposed which take the magnitude of trophic interaction between species into consideration. For 14 properties we examined the performance with increasing sampling effort of a qualitative, an unweighted quantitative (giving the same weight to each taxon), and a weighted quantitative version (weighing each taxon by the amount of incoming and outgoing flows). The evaluation of 10 extensively documented quantitative webs formed the basis for this analysis. The quantitative versions were found to be much more robust against variable sampling effort. This increase in accuracy is accomplished at the cost of a slight decrease in precision as compared to the qualitative properties. Conversely, the quantitative descriptors also proved less sensitive to differences in evenness in the distribution of link magnitude. By more adequately incorporating the information inherent to quantitative food-web compilations, quantitative descriptors are able to better represent the web, and are thus more suitable for the elucidation of general trends in food-web structure. (C) 2003 Elsevier Ltd. All rights reserved.