Coral core extraction

Cores were extracted by SCUBA divers using a pneumatic core drill [1] in 2009 or a hydraulic drill (Chicago Pneumatic COR 5 in 2012 or CS Unitec model 2 1335 0010, 3.8 HP) in 2015, both equipped with a 5 cm diameter diamond-tipped core bit [1] Backreef S. siderea cores collected in 2015 were collected using a pneumatic drill with a 2.5 cm diameter diamond-tipped core bit due to permitting restrictions. All cores were extracted from colonies that appeared healthy (i.e., no bleaching, abnormalities, scarring, or disease) near the center of each colony. Cores were extracted parallel to the growth axis of each colony and spanned the entire height of the colony, with the exception of the backreef S. siderea cores collected in 2015 that ranged from 10-50 cm—spanning only the upper portion of the colony. Overall, core lengths ranged from 10 cm to > 1 m. After extracting each core, a concrete plug was inserted into the drilled hole and sealed with Z-spar underwater epoxy to prevent bioerosion. Epoxy was only placed on the skeleton and the concrete to avoid damage to the living tissue surrounding the hole. Cores were rinsed in ethanol, stored in PVC tubes for transport, and transported to the University of North Carolina at Chapel Hill for analysis. Collection permits were obtained from the Belize Fisheries Department and all cores were collected and transported pursuant to local, federal, and international regulations.

Coral CT Procedures

Coral cores collected in 2009 and 2012 were CT-scanned on a Siemens Somatom Definition AS (120 kV, 300 mAs, 0.6 mm slice thickness) [2-4] at Wake Radiology Chapel Hill in 2013 using methods modified from Carilli et al. [3] and De’ath et al. [5]. Briefly, whole (i.e., unslabbed) cores were CT scanned with the growth axis oriented perpendicular to the length of the CT table. The resulting CT scans were uploaded to the DICOM viewing program Osirix for further analysis following methods modified from Carilli et al. [3]. Transects were drawn parallel to the core growth axis using the “length” tool in Osirix. and within the exothecal space between corallite walls in order to standardize density measurements between transects and cores. Transects were performed in triplicate for each length of the core in order to establish an average, exported to XML, and read into the program RUNNINGCORALGUI, which identified the local density extrema (in Hounsfield units) of the data in each XML file. The locations of these local extrema were then quantified via pixel counting, with halfway points between local extrema defining the boundaries of low and high-density bands. The number of pixels between these halfway points and the average density in Hounsfield units was quantified for the set of pixels between the halfway points. The linear extension of each seasonal light and dark band was then quantified from the total length of the line tool data in pixels, which was then converted to cm.

Coral cores from 2015 were CT scanned on a Siemens Biograph mCT (120 kV, 250 mAs, 0.6 mm slice thickness) at UNC Biomedical Research Imaging Center (BRIC). CT images were reconstructed at 0.1 mm increments using the H70h “very sharp spine” window. All images were exported from the scanner as DICOM files, which were then read into the medical image viewer Horos v2.0.2 (open-source version of Osirix). Semiannual density bands were visualized using a 10-mm thick “Mean” projection oriented through the center of the core. In place of RUNNINGCORALGUI, all boundaries between high- and low-density bands were delineated manually and three sets of linear transects were drawn down the length of the cores using the ROI tool in Horos. Each set of transects was drawn within the exothecal space between corallite walls in order to standardize density measurements between cores and avoid abnormal density spikes in areas where the transect crossed a high-density corallite wall. By-pixel density measurements were then extracted from linear transects and average density was calculated for each semiannual density band. Linear extension (cm) was measured in Horos as the width of each density band, and calcification (g/cm2) was calculated as the product of average density and linear extension.

Coral core density standardization

Nine coral standards were used for density calibration. These standards were pieces of various coral species from the Caribbean that had the same width as the coral cores. Volume and mass of these standards were calculated with calipers and a Mettler Toledo XPE205 analytical balance. Real-world density for each standard was calculated as mass (determined by Mettler Toledo XPE205 analytical balance) divided by volume. The nine internal density standards were scanned along with the cores at least once per scanning session (3-4 scans were completed during each 1-2 hour scanning session). A standard curve was developed for each scanning session that related Houndsfield density (measured from CT scan) to actual coral density (g/cm3), similar to DeCarlo et al. [6].

1. Castillo et al (2011)
2. Saenger et al (2009)
3. Carilli et al (2012)
4. Cantin et al (2010)
5. De'ath G. et al (2009)
6. DeCarlo et al (2015)