Similarity analyses in restoration ecology and how to improve their utility

Abstract

Use of multivariate and nonparametric statistical analyses such as similarity percentages analysis has increased in the past decade within restoration studies. While very useful to compare community composition of restored habitats to reference areas, the ease in which these analyses can be applied, coupled with their power, can result in interpretation errors that could have negative ramifications upon restoration projects. Primarily, similarity measures are often used without stipulating similarity or dissimilarity targets. Despite these drawbacks, the benefits of these methods far outweigh the risks, especially if practitioners take care to specify their community similarity targets a priori and base these targets upon a representative range of reference conditions. However, restoration practitioners should remain focused on fulfilling the objectives of ecological restoration, by ensuring the development of functional habitat that is informed, but not dictated by, statistical analyses. As such, practitioners should take steps to ensure that meaningful univariate trends (e.g. an important species at risk or invasive species) and the overall functionality of the habitat should not be neglected, in favor of these methods of data analysis.

Structuring forces of intertidal infaunal communities on the northern coast of British Columbia, Canada: assessing the relative importance of top-down, bottom-up, middle-out, and abiotic variables

“…Since multivariate analyses can obscure key taxa-specific relationships (Spasojevic and Suding 2012;Murray et al 2014;Gerwing and Hawkes 2021), a PERMANCOVA (which also works well as a univariate analysis) was conducted for selected infaunal taxa. This selection was conducted via similarity percentages analyses (SIMPER; Clarke 1993), which identify taxa contributing the most to observed community variation between sites and rounds.…”

Ecosystem state assessment after more than 100 years since planting for dune consolidation

“…We suggest that plant characteristics can be fruitfully used in rehabilitation assessment as they value the taxonomic and functional species diversity at the community scale. Yet, we stress the importance of an accurate selection of the comparison sites, possibly encompassing different potential rehabilitation outcomes of the target site (Gerwing & Hawkes 2021).…”

“…Our study corroborates the strategy of comparing multiple reference sites to evaluate rehabilitation success and ecosystems state (Gallego‐Fernández et al 2011; Jaouadi et al 2017). Although, few studies in coastal environments have used more than one site, likely because suitable reference sites are difficult to find (Lithgow et al 2013; Gerwing & Hawkes 2021). As expected, our results using different ecological indicators at the site level brought us to similar conclusions of a previous study in the same rehabilitated site using vegetation‐plot data (Landi et al 2012), namely that the rehabilitated site was the most similar to the least disturbed one.…”

Restoration, reclamation, and rehabilitation: on the need for, and positing a definition of, ecological reclamation

“…Admittedly, such a distinction is a matter of degree, as no project realistically attempts to create habitat that is perfectly identical to surrounding reference conditions; nor would that be consistent with the SER Standards. We posit that ecological restoration and ecological reclamation could be differentiated by similarity targets between reference and restored/ reclaimed ecosystems (Gerwing & Hawkes 2021). Ecological restoration seeks the highest attainable recovery of target native reference ecosystems; conversely, ecological reclamation aims for a lower similarity target that may or may not fall within the concept of ecological restoration.…”