Tailoring interaction network types to answer different ecological questions

Abstract

Ecological interaction networks are an important tool for describing species interactions, but many network approaches are available, each with their strengths and weaknesses. In this Perspective, we describe how interaction networks can be differentiated in two main ways: by the extent of node aggregation (how species are lumped into groups) and by the type of information contained in links (potential versus realized interactions). We discuss the ecological questions that each network type can address, how measurements from different types of network should be interpreted, and their relative advantages. Networks with nodes aggregated to functional groups are suitable when focusing on ecosystem-level processes and ecosystem functions. Species-level networks provide information about the assembly of ecological communities or about how abiotic and biotic drivers influence species persistence. Networks with potential links are particularly useful for understanding ecological redundancy or for long-term or large-scale studies, where all potential interactions are likely to be realized. Networks of realized interactions provide access to finer mechanisms of the interplay between abiotic and biotic factors in determining ecological interactions. Identifying the advantages and limitations of different interaction networks will aid methodological decision making and increase the utility and applicability of ecological networks in biodiversity and conservation research programmes. This Perspective provides guidance about how to select the most appropriate type of interaction network for different research questions in ecology. Networks can vary according to the extent of node aggregation and the information that links represent; each kind of network formulation has advantages in different contexts.