Simple prediction of interaction strengths in complex food webs

Eric L. Berlow, Jennifer A. Dunne, Neo D. Martinez, Philip B. Stark, Richard J. Williams, Ulrich Brose

Research output: Contribution to journalArticlepeer-review

224 Scopus citations

Abstract

Darwin's classic image of an "entangled bank" of interdependencies among species has long suggested that it is difficult to predict how the loss of one species affects the abundance of others. We show that for dynamical models of realistically structured ecological networks in which pair-wise consumer-resource interactions allometrically scale to the 3/4 power - as suggested by metabolic theory - the effect of losing one species on another can be predicted well by simple functions of variables easily observed in nature. By systematically removing individual species from 600 networks ranging from 10-30 species, we analyzed how the strength of 254,032 possible pair-wise species interactions depended on 90 stochastically varied species, link, and network attributes. We found that the interaction strength between a pair of species is predicted well by simple functions of the two species' biomasses and the body mass of the species removed. On average, prediction accuracy increases with network size, suggesting that greater web complexity simplifies predicting interaction strengths. Applied to field data, our model successfully predicts interactions dominated by trophic effects and illuminates the sign and magnitude of important nontrophic interactions.

Original languageEnglish (US)
Pages (from-to)187-191
Number of pages5
JournalProceedings of the National Academy of Sciences of the United States of America
Volume106
Issue number1
DOIs
StatePublished - Jan 6 2009

Keywords

  • Body size
  • Ecological networks
  • Species extinctions
  • Species interaction strengths
  • Systems theory

ASJC Scopus subject areas

  • General

Fingerprint Dive into the research topics of 'Simple prediction of interaction strengths in complex food webs'. Together they form a unique fingerprint.

Cite this