Long-term monitoring programs are fundamental to detecting changes in ecosystem health and understanding ecological processes. In the current context of increasing anthropogenic threats on marine ecosystems, understanding the dynamics and response of communities becomes essential. We used data collected over 14 years in the REBENT benthic coastal invertebrates monitoring program, at a regional scale in the North-East Atlantic, covering a total of 26 sites and 979 taxa. Four distinct habitats were studied: two biogenic habitats associated with foundation species in the intertidal and subtidal zones and two bare sedimentary habitats in the same respective tidal zones. We used Community Trajectory Analysis, a statistical approach that allows for quantitative measures and comparisons of temporal trajectories of ecosystems. We compared observed community trajectories to trajectories simulated under a non-directional null model in order to better understand the dynamics of the communities, their potential drivers, and the role of the studied habitats in these dynamics. Despite strong differences in the community compositions between sites and habitats, the communities followed non-directional dynamics during the 14 years monitored, which suggested stability at the regional scale. However, the shape, size, and direction of the trajectories of benthic communities were more similar within than among habitats, also suggesting the influence of the nature of the habitat on community dynamics. Results showed a higher variability in community composition in the first years of the monitoring in the intertidal bare habitat and confirmed the role of biogenic habitats in maintaining temporal stability. They also highlighted the need to apprehend the role of transient and rare species and the scale of observation in temporal beta diversity analyses. Finally, our study confirmed the usefulness of Community Trajectory Analysis to link observed trajectory patterns to fundamental ecological processes.

Benthic communities have been monitored yearly since 2003 along the coast of Brittany (France) within the REBENT program (http://www.rebent.org). We focused on four habitats and their associated endofaunal communities: intertidal seagrass beds (only Zostera marina beds are monitored within the REBENT), intertidal sandy beaches, subtidal maerl beds and subtidal soft sediments (respectively referred to as intertidal biogenic habitat (IBIO), intertidal bare habitat (IBAR), subtidal biogenic habitat (SBIO) and subtidal bare habitat (SBAR) in the manuscript).

At each site, three faunal samples were taken at each of three fixed sampling points distributed 200 m apart (0.03 m² cores in the intertidal and 0.1 m² Smith-McIntyre grabs in the subtidal), except for Pierre Noire (8) where up to 10 grabs were taken at the sampling site.

Sampling was performed between the end of February and the beginning of May before recruitment of most species occurs in the region. In the laboratory, specimens were sorted, counted and identified to the lowest possible taxonomic level (usually species). Since the acquisition and identification of specimens were not carried out by the same people over the years of the monitoring program, we proceeded to a taxonomic homogenization: each recorded taxon was scrutinized by experts in benthic taxonomy and their names were checked thanks to the World Register of Marine Species (WoRMS Editorial Board 2021) to ensure a consistent taxonomic resolution. In order to minimize the prevalence – and potential effect – of missing data as much as possible, we only selected sites with complete time series and with at least 3 core or grab samples in any particular year. Samples were pooled to estimate abundances at the site level. In the end, this led to a selection of 26 sites monitored from 2005 to 2018 while keeping a spatial resolution covering the coasts of Brittany and encompassing most of the environmental settings found in this region. Of these 26 sites, 5 were in IBAR, 7 in IBIO, 8 in SBAR and 6 in SBIO . We conducted the analyses at the habitat level or within the same tidal level because it may not be relevant to run analyses including two different tidal levels since sampling gears differ between intertidal and subtidal sites.

More details can be found in the article.

All analyses were conducted with the R programming language version 4.1.2. (R Core Team, 2021. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/).