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

was adjusted to mean low water (MLW). The presence of different seagrass species, and of sea- grass-free bottom, were recorded at points (distances) along transects where changes in sea- grass composition occurred. The recorded substratum cover type was deemed constant from each observation point until the point where there was a change in type. Total cover per sub- stratum type was calculated as the sum of the distances (in m) represented by that type along a transect to 50 m offshore, common to all transects (and thus restricted to nearshore areas only). Cover values were converted to percent of the total distance (50 m) for analysis. We counted erosional scarps occurring on the inshore margin of seagrass beds and associ- ated with “blowouts” within the seagrass beds [35] along transects, including those beyond 50 m. Only actively eroding scarps (i.e. those with underlying seagrass bed sediments exposed) were counted; older, overgrown scarps were not included in these counts. We also measured the extent of the seagrass-free inshore zone (SFI) as the distance along a transect (in m) from the shore at MLW to the first occurrence of seagrass. This zone often was the subtidal exten- sion of a sandy beach; in some areas it contained scattered rubble, boulders or sparse macroal- gae but no coherent seagrass patch or bed. We limited our analysis of these erosional features to 80 m offshore, common to all transects. In 1994, due to logistical constraints, the full set of transect measurements was limited to 4

of the 10 transects (T6, T7, T9, T14); scarps however were counted on all 10 transects. In 2016, the full set of transect measurements was repeated on 9 the 10 transects. The remaining tran- sect (T9) was incorrectly positioned in 2016; the correct site was sampled in February 2019 and the observations included with the 2016 data. Based on observed wave conditions in the field, we classified 6 transects (T1, T5 –T9) as HWE (High Wave Energy) sites, and 4 transects T2, T3, T10 and T14 as LWE (Low Wave Energy) sites in 1969. Relative Exposure Index was calculated after [42, 43] for the shore posi- tion on each of the 10 transects in 2016 (2019 for T9) based on local wind data [44] and limit- ing ocean fetch to 200 km [45]. There was an order of magnitude difference in mean REI between HWE sites (T1, T5 –T9: 5327) and LWE sites (T2, T3, T10, T14: 421) in 2016 (S1 Fig), consistent with our qualitative classification. The minimum REI recorded at 1 HWE site (T1) was equivalent to the maximum recorded at 1 LWE Site (T2). However, our observations of wave conditions at T1 indicated it was a high wave energy site, which was not captured in the REI due to the reduction in fetch from surrounding islands. This HWE designation was sup- ported by the occurrence of coarser, better sorted sediments at T1 than T2 based on samples in 1969 [35]. Based on studies of reef sediment transport and distribution off the east coast of Carriacou, Watering Bay (where T2 is located) has been characterized as a low wave energy environment [46, 47]. Stations

In addition to line transect sites, in 1969, 24 sampling stations (10 x 10 m grids) were located at sites chosen to be representative of seagrass beds around Carriacou, as viewed in the aerial photographs (e.g., fringing bed, offshore patch, lagoonal patch, cobble bank). Of these 24 sta- tions, 17 were resampled in both 1994 and 2016 and form the basis of our spatial-temporal comparisons (Fig 1, S1 Table). At each station, 12 quadrats (0.25 m2) were randomly posi- tioned within the sampling grid. Presence of all readily visible macroflora (seagrasses, sea- weeds) and macrofauna (sponges, corals, echinoderms, conch) was recorded in each quadrat. In 1969 and 1994, macroalgal specimens that were not clearly identifiable in the field were col- lected for subsequent description and identification. Macroalgae were identified by reference to Taylor [48]. A collection of 14 sponges in 1969 was identified by Prof. George Hechtel (State University of New York) in 1971. In 2016, photographs of all species were taken with an PLOS ONE Seagrass bed biodiversity and structure and across 5 decades PLOS ONE | https://doi.org/10.1371/journal.pone.0306897 August 1, 2024 5 / 25