Out‐of‐sample predictions from plant–insect food webs: robustness to missing and erroneous trophic interaction records

摘要: With increasing biotic introductions, there is a great need for predictive tools to anticipate which new trophic interactions will develop and not. Phylogenetic constraint of in both native novel food webs can make some predictable. However, many are sparsely sampled, or may include inaccurate interactions. In such cases, it unclear whether modeling methods still useful We ran bootstrap simulations host-use models on Lepidoptera-plant data set remove records add erroneous order observe the effect missing prediction with plants. found that model was robust large amount interaction records, but lost power addition relatively few records. The loss due inaccuracy estimating phylogenetic distance between hosts. Removal proportionally their encounter frequency field had little power. Host-use have immediate value predicting from large, sampled databases