Coral Faunas of the Florida Keys: A Report on the Keyswide Coral Reef Expedition

Aronson, Richard; Murdoch, Thaddeus

Coral reef communities have declined along the Florida reef tract, but the magnitude and causes of that degradation remain controversial. Part of the problem stems from the inappropriate extrapolation of small-scale parte s observed on individual reefs to much larger scales. The Keyswide Coral Reef Expedition was conceived as a multidisciplinary effort to conduct a hypothesis-driven, synoptic survey of the biota of the entire Florida reef tract. To achieve a better understanding of patterns of variation among regions of the reef tract, variation among reefs within regions, and variation within reefs, we videographically sampled replicate transects on 20 reefs at 13 -19m depth from Biscayne Bay to the Dry Tortugas. Of the 20 reefs sampled, 16 were characterized by spur-and-groove habitat in this depth range. Three of the 16 spur-and-groove sites-- Carysfort, Tennessee and Western Sambo Reefs--were each sampled in three areas ("subreefs") these subreefs, or reef subsections, were separated by <1 km. Coral cover (from point counts), species diversity (H'; also from point counts), and the number of coral colonies were estimate for each video transect. A nested Analysis of Variance (ANOVA) design was used to compare the variation among Carysfort, Tennessee and Western Sambo Reefs to the variation am g subreefs within those reefs. For both coral cover and H', this analysis revealed low variation among subreefs within reefs but significant variation among reefs. A nested ANOVA on the coral colony count data revealed significant variation at both the among-subreef and among-reef levels. Cluster analysis of multivariate coral cover data (point counts partitioned by species) showed that subreefs within reefs were more similar to each other than they were to subreefs from different reefs. Cluster analysis of multivariate coral colony count data yielded similar results. A second series of nested ANOVAs compared variation among reefs ( 1 0-km scale) to variation among regions of the reef tract (50 km scale), again based on coral cover, H' and coral colony counts. Variation in these univariate measures was again significant at the among-reef scale. Significant inter-regional variation was also detected. That variation became non-significant for coral cover and H', but not for coal colony counts, when the isolated reefs of the Dry Tortugas region were removed from the analysis. Thus, variation among subreefs adequately predicted variation within the reef as a whole, but variation among reefs within regions outweighed variation among regions. Cluster analyses of multivariate coral cover, coral colony count and species presence/absence data supported these conclusions. When the four hardground sites were included in these cluster analyses they generally clustered together, indicating that substratum type influences the composition of the coral assemblage living in this depth range. The results suggest that there are no geographic gradients in coral cover or the composition of coral assemblages in the habit and depth range studied. Coral cover and assemblage composition may be related to e distances reefs from passes between the Keys; community structure in this depth range may be affected by inimical waters from Florida Bay. Thus, reef community composition is probably determined by the interaction of multiple factors that act at the scale of the individual reef and at larger scales.