In May 2015 and 2016, skeletal cores were collected from 39 colonies of Siderastrea siderea and 31 colonies of Pseudodiploria strigosa at four pairs of inner-outer reef sites spanning the Florida Keys Reef Tract. All cores were extracted from the vertical growth axis of each coral colony using a CS Unitec model 2 1335 0010 hydraulic drill affixed with hollow extension rods and a 5 cm diameter wet diamond core bit. Collected cores were then stored in capped PVC tubes filled with 100% EtOH and transported to the University of North Carolina at Chapel Hill where they were air dried in preparation for sclerochronology development. To assess coral skeletal growth histories, all cores were scanned using X-ray computed tomography (CT) on a Siemens Biograph CT scanner at the Biomedical Research Imaging Center, University of North Carolina at Chapel Hill. Coral cores were oriented lengthwise in rows of 4 to 5 on the scanning table, and equipment parameters were set to 120 kV, 250 mAs and 0.6 mm slice thickness with images reconstructed at 0.1 mm increments using the H70h “Very Sharp Spine” window.

All images were exported from the CT scanner as DICOM files, which were then 3-dimensionally reconstructed using the open-access Horos v2.0.2 medical image viewing software. High- and low-density bands were visualized using a 10-mm thick ‘Mean’ projection oriented as a rectangular prism through the center of each core. All boundaries between semiannual density bands were delineated manually and three sets of linear transects were drawn down the length of the cores using the Region of Interest (ROI) tool in Horos. Note, each set of transects was drawn within the exothecal space between corallite walls in order to standardize density measurements and to avoid aberrant density spikes in areas where the transect may otherwise have crossed a high-density corallite wall. Density and calcification measurements are therefore lower than would be expected if all features of the skeletal architecture were taken into account. Additionally, it has been shown that individual colonies may vary in their timing of high- and low-density band deposition due to intraspecific differences in tissue thickness and morphology (Barnes and Lough, 1996; Carricart-Ganivet, 2013). Thus, to approximate a consistent time standard between cores, we begin all chronologies at the top of the first fully deposited density band beneath the band of terminal growth. By-pixel density measurements were extracted from linear transects and average density was calculated for each semiannual high- and low-density band. Following DeCarlo et al (2015), nine coral standards of known density were included in every scanning session to convert density measurements from CT Hounsfield units to g cm-3. Average density of each standard was assessed in Hounsfield units using Horos and a standard curve was created for all cores scanned in the corresponding session. Linear extension (cm) was measured in Horos as the width of each annual density band couplet, and calcification (g cm-2 yr-1) was calculated as the product of density and linear extension.

Note, because cores were collected in subsequent years, the most recent year of growth (2015) was not included for cores collected in 2016 in order to keep the beginning of chronologies uniform throughout.

Analysis was based on previously published methods in Castillo et al (2011), Castillo et al (2012).