Computational protocol: Population structure of the hydrocoral Millepora platyphylla in habitats experiencing different flow regimes in Moorea, French Polynesia

Similar protocols

Protocol publication

[…] All M. platyphylla colonies were georeferenced by determining their position along the transect-line (0 to 300 m) and straight-line distance from both sides of the transect (0 to 10 m). From these measures, each colony was mapped with x and y coordinates, from which the distribution index (DI) and mean neighborhood distance (ND) were calculated using the spdep package [] in R []. The DI is based on Ripley’s method [] and calculated for each transect to determine whether colonies were having a contagious (DI > 1), random (DI ≈ 1) or homogenous (DI < 1) pattern of distribution []. The mean distance to each colony’s 10 nearest neighbors was estimated and the mean ND was calculated for each transect. The mean colony density (n.m-2) and cover (%) were also calculated for each transect (i.e. 3000 m2). Using these variables, variability in the spatial distribution among habitats was quantified by one-way PERMANOVA tests in PRIMER 6 software [], since assumptions of parametric testing could not be met. Pair-wise tests followed the PERMANOVA to assess the degree of similarity among habitats. In order to determine how different habitats with contrasting water regimes affect the spatial distribution of M. platyphylla, we assumed that swell wave energy exposure decreases with habitat depth and its proximity to the coastline, as demonstrated in previous studies [,]. Consequently, the density, cover, DI and ND were regressed against the mean depth and mean distance from shore estimated from the three transects within each of the five surveyed habitats and Pearson’s r coefficient was used to test for significant correlations. [...] The size-frequency distributions of M. platyphylla populations were generated from estimates of colony sizes computed from 2D photographs. Photographs were taken from above the colony and included a plate of known dimensions positioned next to each colony. For bigger colonies, pictures were taken from a larger distance, and for 5 colonies (out of the 3561) 2 photographs were required to photograph the whole colony. Each colony size, standardized as the projected surface, was then measured (in cm2) using ImageJ 1.4f software []. The size-frequency distribution for each transect was given as percentages of all colonies belonging to 10 size classes on a logarithmic scale. Data were then analyzed using basic statistical measures of size hierarchies []: the coefficient of variation (CV) and skewness (g1), indicative of the relative abundance of small and large colonies within a population. CV and g1 were computed for each habitat per transect together with standard descriptive statistics, such as 95% percentile of the mean (describes the maximum colony size reached within a population, see []) and the probability that the data are normally distributed (Kolmogorov-Smirnov test, Pnorm). Differences in size-frequency distributions among habitats were quantified using one-way PERMANOVA based on normalized abundances. Spearman’s rank coefficient and pair-wise tests followed the PERMANOVA to assess the degree of similarity among habitats. […]

Pipeline specifications