How to quantify and minimise error in coral skeletal density estimates using X-ray µCT

X-ray micro-computed tomography (µCT) is widely used in environmental and ecological studies to visualise the complex skeletons of marine calcifying organisms. However, the potential to quantify skeletal density in marine calcifiers is complicated because µCT greyscale intensities are influenced by many factors. Here we constrain sources of error in coral skeletal densitometry and show how errors can be quantified, minimised and reported. We establish the accuracy and precision of density estimates by analysing a population of massive Porites spp. coral that vary in bulk average skeletal density (1.07–1.57 g cm$^{−3}$), size (4–25 cm in diameter), mass (87.9–2176.5 g) and morphology (41 specimens scanned in 81 individual µCT acquisitions). The population mean offset between the µCT-derived density and actual values was 0.2 ± 4.6% (⁠$\overline{x} ± 1$ SD, $n = 41$). Following acquisition-specific corrections based on bulk density offsets, equivalent internal regions of low and high density within repeatedly scanned corals under different settings varied by only ±0.8% (1 SD). We present a workflow for µCT laboratories to measure density, and minimise and report errors. Application of these recommendations will ultimately improve data comparison of coral skeletal growth assessments. The workflow recommendations are transferable to the µCT densitometry analysis of other marine calcifying organisms.

• View at Publisher

Free Full Text ( Final Version , 3mb )