Cross-continental analysis of coastal biodiversity change
Abstract
Whereas the anthropogenic impact on marine biodiversity is undebated, the quantification and prediction of this change is not trivial. Simple traditional measures of biodiversity (e.g., richness, diversity indices) do not capture the magnitude and direction of changes in species or functional composition. In this paper, we apply recently developed methods for measuring biodiversity turnover to time-series data of four broad taxonomic groups from two coastal regions: the southern North Sea (Germany) and the South African coast. Both areas share geomorphological features and ecosystem types, allowing for a critical assessment of the most informative metrics of biodiversity change across organism groups. We found little evidence for directional trends in univariate metrics of diversity for either the effective number of taxa or the amount of richness change. However, turnover in composition was high (on average nearly 30% of identities when addressing presence or absence of species) and even higher when taking the relative dominance of species into account. This turnover accumulated over time at similar rates across regions and organism groups. We conclude that biodiversity metrics responsive to turnover provide a more accurate reflection of community change relative to conventional metrics (absolute richness or relative abundance) and are spatially broadly applicable.
Methods
Data comprised community composition assessments for major organism groups in two coastal regions (Germany, South Africa), including phytoplankton and zooplankton (grouped collectively as ‘plankton’ hereafter), benthic macroinvertebrates, fish and coastal birds. Site descriptions for the sampling locations of these regions are expanded upon in the respective article. Each dataset reports on the presence and abundance of species (or in some cases, a coarser taxonomic level, but hereafter ‘species’ are referred to) in coastal marine and estuarine ecosystems. We pooled the data to yearly averages per species and site across all samplings. Datasets were checked for consistent naming of species and reporting, but otherwise used as reported.
We used the approach suggested by Hillebrand et al. (2018, as cited in the articles) to quantify the amount of change in species composition. All analyses were performed in R, we used the vegan package to calculate taxon richness and effective number of species (ENS) as measures of standing diversity, where ENS is the standardized species diversity measure assuming equal abundance in the community.For turnover, we used the presence-absence based Jaccard index to calculate the species exchange ratio based on richness (SERr), where 0 is no exchange of species identities, and 1 a complete overturn. We used Wishart’s dissimilarity as a measure of abundance-based species exchange ratio (SERa), which is based on Simpson’s index of dominance, a feature shared with ENS. SERa also ranges between 0 and 1, with 0 indicating that species identity and relative abundance remains unchanged, but 1 indicating a complete exchange of species or a change in dominance structure.
Usage notes
Two data sets are provided
year.csv
For each sampling station (StationID), identifying the organism group analysed (organism) and region (region), we report the annual ENS and species richness (S) for each year (year_start) as well as the change in richness, SERr and SERa to the following year x+1.
all.csv
To analyse how biodiversity change accumulates over time, we calculated SERa and SERr for all combinations of years, i.e., between any year x and any consecutive year y. For each sampling station (StationID), identifying the organism group analysed (organism) and region (region), we give the starting year (year_start = x) and the temporal distance between y and x (dist), and the measures for SERa and SERr as well as the corresponding change in species richness.