Testing a method for quantifying structural redundancy in ecological communities

In ecological communities, some taxa contribute to ecosystem functioning in highly similar ways. When numerous taxa within a community contribute similarly, this might suggest a degree of functional redundancy, where the loss of one species may not impair ecosystem functioning, providing other functionally similar species can compensate for the loss.

Multidimensional ordination is commonly used to analyse patterns in ecological communities over space and time, being most intuitively visualised using multidimensional scaling techniques. The patterns produced by these techniques can therefore be thought of as a community's structure. Research has shown that it is typical for multiple taxa in a community to contribute to this structure interchangeably, representing a different form of redundancy, structural redundancy, where only a small subset of taxa is required to recreate the overall community pattern. Further to this, there can be multiple mutually exclusive subsets of taxa within a community that can reproduce the full‐community pattern.

In this study, we examine the proposed method to quantify structural redundancy within ecological communities. The intention behind the development of this method was to explore any potential link between the statistical concept of structural redundancy and concepts of ecosystem function, such as functional redundancy and compensation. Notionally, a high degree of structural redundancy within a community data set might be indicative of the potential for functional redundancy.

We perform a series of tests on multiple freshwater macroinvertebrate community data sets to discover what, if any, characteristics of a community strongly influence the amount of structural redundancy present. Our tests ultimately determined that the method is extremely sensitive to sample size. As such, we demonstrate that this approach is unable to provide reliable information regarding the potential for functional redundancy in ecological communities.

This was an unexpected result, which motivated a further preliminary simulation revealing that the underlying correlation procedure is also subject to the same sensitivity. This analytical technique is widely used in community ecology, especially in the analysis of biota–environment relationships, and is implemented in major statistical software packages.