Methods reported in paper in MEPS.

Six sites at 7–9 m depth were established in 1992 using randomly selected coordinates restricted to hard substrata on fringing reefs between Cabritte Horn and White Point, St. John. Four of these consist of coral communities on predominantly igneous substrata (White Point, west Little Lameshur Bay, East Tektite, and Cabritte Horn), and two on a carbonate framework (Europa Bay and Neptune’s Table). Sites were annually surveyed, and each was photographically sampled along a permanently marked transect. The transects were 20 m long from 1992 to 1999, and in 2000 were extended to 40 m because the implementation of digital photography allowed more photoquadrats to be recorded on each dive than was possible using color slide film. Surveys were completed in the late spring or summer, and from 1998–2019, employed sampling between July 17 and August 27, and between May 27 and August 14 from 1992 to 1997 (summarized in Edmunds 2022). Approximately 18 photoquadrats (0.5 × 0.5 m) were recorded along each transect at each site from 1992–1999, and ~ 40 photoquadrats were recorded annually thereafter. Photoquadrats were randomly positioned each year along the transects, and were recorded with lighting provided by twin strobes (Nikonos SB 105) attached to a variety of cameras. The cameras were mounted on a framer that supported them with their focal plane parallel to the quadrat, and the framer was positioned approximately orthogonal to the benthos at all sites. Large coral colonies and igneous boulders occasionally resulted in non-orthogonal placement, but such cases were rare on the study reefs. Initially a 35 mm Nikonos V camera fitted with color slide film was used, and digital photography was used from 2001 with resolutions increasing from 3–36 MP over 19 yrs. The color slides were digitized (4000 dpi) for analyses.

Photoquadrats first were analyzed for percentage cover of benthic space holders using CPCe software, and then CoralNet software when it became available, in both cases with 200 random points overlaid on each image and annotated manually. Second, the photoquadrats were analyzed using ImageJ software to quantify the abundance and size of coral colonies. Colonies were defined as autonomous areas of tissue, and were scored when they were enclosed by the photoquadrats, or if more than half of the area of colonies that were roughly circular in outline was within the photoquadrat. The size of colonies fully enclosed in the photoquadrats was calculated as the mean of the maximum and minimum diameters, assuming they were circular, and in cases of partial enclosure, colony size was estimated from a single diameter assuming they were circular. With these procedures, colonies as small as ~ 5 mm diameter were resolved, and they were identified to the lowest taxonomic level possible, counted, and their sizes recorded (mm).

Some colonies could not be identified to species across the colony size range encountered, particularly for the smallest colonies. Some taxa, therefore, were pooled to genus (e.g., Orbicella annularis, O. franksi, and O. faveolata), to create a consensus list of 30 taxa for which density and size are reported. At each site, coral density was calculated using photoquadrats as statistical replicates, and coral sizes were determined using coral colonies as statistical replicates. To describe overall changes in coral communities at the six sites, density and sizes of coral colonies were pooled among taxa and mean values calculated by site for the 28 years from 1992 to 2019.

Problems/Issues:
Some irregular number of quadrats were samples at some sites in some years.

Funding note: The most recent funding for this time series was provided by NSF award OCE-2019992 for project "RUI: Pattern and process in four decades of change on Caribbean reefs." The "Project" and "Funding Sources" sections of this page list also include past awards that directly funded this dataset.

897571_v1_coral-sizes.csv