Shifts in biodiversity and physical structure of seagrass beds across 5 decades at Carriacou, Grenadines

and 1969 or 2016, than between 1969 and 2016 (S3 Table). Ordination analysis (non-metric MDS) shows a broad cluster of stations in 1994 characterized by the absence (except at one site) of sea urchins (Tripneustes ventricosus group), in contrast to stations in both 1969 and 2016 where sea urchins occurred at 7 out of 17 stations, generally at moderate to high frequen- cies (Fig 7). Correlations between ordinations for benthic assemblages and frequencies of occurrence for functional groups were highest for seagrasses followed by the various macroal- gal groups, and lowest for sponges. Extent of the seagrass-free inshore zone and erosional scarps on transects The extent of the seagrass-free inshore zone (SFI) increased at all 6 HWE sites between 1969 and 2016, with a more than 2-fold increase in the average (from 16.9 to 35.9 m). In 2 of the 3 HWE transects where the SFI was measured in 1994 and SFI increased between 1969 and 2016 (T6, T7), most of this increase occurred between 1994 and 2016 (Fig 8, S2 Fig); at T9, SFI decreased between 1994 and 2016. At LWE sites, changes in SFI were generally smaller and more variable between 1969 and 2016: SFI increased measurably at one site (T2, from 0 to 14 m), decreased at another (T14, from 27 to 17 m), and was effectively stable (~ 4–5 m) at the remaining 2 sites (T3, T10), resulting in a minimal change on average (from 8.6 to 10.2 m). 2-way ANOVA showed a significant effect of year (greater in 2016) on SFI (F1,16 = 8.0, P = 0.012) but no effect of wave exposure (F1,16 = 1.2, P = 0.289) or interaction between year

and wave exposure (F1,16 = 2.7, P = 0.119). Given the unbalanced design and high variability between transects at both LWE and HWE sites, we conducted Welch 2-sample t-tests to com- pare SFI between wave exposure levels in 1969 and 2016. This showed a highly significant effect of wave exposure on SFI in 2016 (t7.2 = -3.53, P = 0.009) but no effect in 1969 (t4.8 = -1.2, P = 0.287). No scarps were observed on transects at LWE sites in 1969 and 1994 but 1–2 scarps were present at 3 of the 4 sites in 2016. In 2016 scarps were present at on all HWE transects, except T5 where the seagrass bed had been scoured down to the underlying bed rock (Fig 9, S2 Fig). Fig 6. Taxon richness and diversity of the seagrass bed community. Mean + SE richness and diversity (Shannon Index) at stations (n = 17) in 1969, 1994, 2016. https://doi.org/10.1371/journal.pone.0306897.g006 PLOS ONE Seagrass bed biodiversity and structure and across 5 decades PLOS ONE | https://doi.org/10.1371/journal.pone.0306897 August 1, 2024 12 / 25