Coral community calcification (G, kg CaCO₃ m^-2 yr^-1) by site and year was calculated based on colony sizes that were measured as described above and used to calculate G summed by photoquadrat (i.e., photoquadrats were treated as statistical replicates). In the formulation applied herein, G provides an estimate of gross coral calcification, as we do not consider any source of dissolution and bioerosion. G is not equivalent to NCC because dissolution/bioerosion is not considered, and we do not evaluate calcification by taxa other than corals. First, the mean diameters of each coral colony in each quadrat (i.e., the mean colony diameter in planar view), were used to estimate the 3-dimensional colony size (i.e., the actual size of living tissue) based on taxon-specific rugosity for Caribbean corals (Equation 1, see supplemental file "Coral Calcification Equations"). Actual tissue cover was required to accurately estimate colony-level calcification, because coral species differ in their 3-D shapes, which affects G depending on colony morphology.

See Equations in supplemental pdf file "Coral Calcification Equations".

The annual calcification (kg colony-1 yr-1) of corals in each photoquadrat was calculated by summing the CPi values by taxon. Values were standardized to a square meter to be consistent with a commonly used unit of G, kg CaCO3 m-2 yr-1, which can be converted to other units through stoichiometry (e.g., 1 kg CaCO3 m-2 yr-1 = 27.4 mmol CaCO3 m-2 d-1). In our calculations we assume that coral calcification rates are uniform with colony size (i.e., isometric). Work in Moorea has shown that some colony calcification rates increase rapidly up until colonies are ~10–15 cm diameter, but our assumption of isometry is currently necessary for Caribbean corals as the data necessary to parameterize the relationship between colony size and coral calcification currently is lacking for the corals in this region. However, given that nearly all colonies in our study are small (averaging < 6 cm) and remain small due to high demographic turnover, the use of taxon-specific calcification (i.e., independent of colony size) rate is less problematic than would be the case in Moorea. Our assumption of isometric coral calcification may render a conservative bias in our calculations of coral G. Using photoquadrats as statistical replicates, coral G (pooled among taxa) was compared among sites and years (Hypothesis 2) using a mixed effects, two way ANOVA in which site was a random effect, year was a fixed effect, and G was log(x +1) transformed to meet the assumptions of the statistical test. G was also averaged by site and year using values pooled among taxa and also by taxon, and the values were qualitatively compared over time for each site. They were also used to assess whether the taxa contributing to coral G varied within- or among- sites as a function of coral community composition and colony size.

Location: South coast of St. John, US Virgin Islands. 18.31644, -64.724528. Research conducted during annual field expeditions to research lab: The Virgin Islands Ecological Research Station.