A statistical social network model for consumption data in trophic food webs

Abstract

We adapt existing statistical modeling techniques for social net- works to study consumption data observed in trophic food webs. Thesedatadescribethefeedingvolume(non-negative)amongor- ganisms grouped into nodes, called trophic species, that form the foodweb.Modelcomplexityarisesduetotheextensiveamountof zerosinthedata,aseachnodeinthewebispredator/preytoonly asmallnumberofothertrophicspecies.Manyofthezerosarere- gardedasstructural(non-random)inthecontextoffeedingbehav- ior.Thepresenceofbasalpreyandtoppredatornodes(thosewho never consume and those who are never consumed, with proba- bility 1) creates additional complexity to the statistical modeling. Wedevelopaspecialstatisticalsocialnetworkmodeltoaccountfor suchnetworkfeatures.Themodelisappliedtotwoempiricalfood webs;focusisonthewebforwhichthepopulationsizeofsealsis ofconcerntovariouscommercialfisheries