• food webs;
  • fuzzy cognitive maps;
  • fuzzy logic;
  • interaction strength;
  • kokako;
  • loop analysis


  • 1
    Quantitative predictions of the effects of perturbations on communities of interacting species have often proved to be difficult. However, if precise predictions are not a requirement then qualitative models of community dynamics offer an alternative method for predicting species responses to perturbations.
  • 2
    We used two qualitative modelling approaches to predict the effects of predator control on the fledging rate of an endangered New Zealand bird, the North Island kokako. The first approach was based on loop analysis and provided predictions on the probable direction of change in species abundances to single species perturbations. The second approach, ‘fuzzy interaction webs’, used fuzzy logic in the framework of a fuzzy cognitive map to provide predictions on the probable magnitude of change in species abundances to perturbations.
  • 3
    Using both methods, we predicted the qualitative change in the equilibrium fledging rates of kokako under various regimes of single- and multispecies predator control (ship rats, brushtail possums and stoats). Single species control was insufficient to lift the fledging rate from ‘low’ to ‘moderate’. However, simultaneous control of both ship rats and possums had the greatest influence on the fledging rates compared with any other combination as a result of the additional indirect effect of ship rat control on stoat abundance.
  • 4
    We propose qualitative modelling of community dynamics as a method suitable for predicting the effects of perturbations in complex ecological communities that can encapsulate diverse sources of knowledge about food web interactions. We believe that these methods are a useful set of heuristic tools that can be used to propose testable hypotheses about ecosystem functioning that can complement existing statistical and quantitative modelling approaches.