A simulation tool to scrutinise the behaviour of functional diversity metrics. Issue 1 (14th August 2017)

• Record Type: Journal Article
• Authors: McPherson, Jana M.; Yeager, Lauren A.; Baum, Julia K.
• Editors: Price, Samantha
• Is Part Of: Methods in ecology and evolution. Volume 9:Issue 1(2018)
• Page Start: 200
• Page End: 206
• Publication Date: 2017-08-14
• Subjects: abundance distributions -- artificial ecological communities -- functional diversity -- functional redundancy -- index reliability -- R code -- simulations -- site‐by‐species abundance matrix -- trait distributions -- traits matrix; Ecology -- Periodicals; Evolution -- Periodicals; 577
• DOI: 10.1111/2041-210X.12855
• Languages: English
• ISSNs: 2041-210X

Abstract

Abstract: Many indices have been proposed to measure functional diversity and its four distinct dimensions: functional richness, evenness, divergence and redundancy. Identifying indices that reliably measure the functional diversity dimension(s) of interest requires careful testing of how each index responds to species' traits and abundance distributions. In the absence of a convenient simulation tool, tests with artificial data have to date explored only a limited number of scenarios or have altered trait and abundance distributions only indirectly based on principles of evolution and community assembly. We provide simul.comms, an R function that allows users to test the efficacy of functional diversity indices by easily creating artificial species communities with user‐specified abundance and trait distributions for continuous, ordinal and categorical traits. To illustrate the function's utility, we examine the performance of R, a recently published abundance‐sensitive index for functional redundancy. We use two approaches to designing simulation tests for this example analysis. The first uses simul.comms to create six separate sets of artificial communities to qualitatively assess how R responds to predictable changes in functional redundancy. The second uses simul.comms to independently alter seven community composition parameters, whose influence on R is then examined quantitatively via effect sizes in linear regression. Our analyses indicate that R broadly mirrors expected changes in functional redundancy and predictably responds to changes in community composition parameters. R appears, however, to primarily reflect trait distributions, responding minimally to variance in abundance and counter‐intuitively to abundance range. Further refinement of tools to measure functional redundancy may therefore be desirable. The R tool we provide should assist with refining functional diversity measures, a critical step towards improving our ability to understand and mitigate the impacts of biodiversity loss on ecosystem functioning. Because simul.comms simply produces two linked matrices, a species‐by‐traits matrix and a site‐by‐species abundance matrix, it may be equally valuable in exploring questions and analytical approaches in other areas of community ecology.