Quantifying coral morphology

Coral morphology has important implications across scales, from differences in physiology, to the environments they are found, through to their role as ecosystem engineers. However, quantifying morphology across taxa is difficult, and so morphological variation is typically captured via coarse growth form categories (e.g. arborescent and massive). In this study, we develop an approach for quantifying coral morphology by identifying continuous three-dimensional shape variables. To do so, we contrast six variables estimated from 152 laser scans of coral colonies that ranged across seven growth form categories and three orders of magnitude of size. We found that 88% of the variation in shape was captured by two principal components. The main component was variation in volume compactness (cf. convexity), and the second component was a trade-off between surface complexity and top-heaviness. Variation in volume compactness also limited variation along the second axis, where surface complexity and top-heaviness ranged more freely when compactness was low. Traditional growth form categories occupied distinct regions within this morphospace; however, these regions overlapped due to scaling of shape variables with colony size. Nonetheless, with four of the shape variables we were able to predict traditional growth form categories with 70 to 95% accuracy, suggesting that the continuous variables captured most of the qualitative variations implied by these growth forms. Distilling coral morphology into continuous variables that capture shape variation will allow for better tests of the mechanisms that govern coral biology, ecology and ecosystem services such as reef building and provision of habitat.