Dataset: Fish count data
Deployment: Bernardi_2012

We compared sites using agglomerative hierarchical clustering (Ward’s minimum variance method; hclust in The R Stats Package, R Core Team, 2016) based on the Bray-Curtis dissimilarity index (for benthic data) and Cao dissimilarity Index (for fish data), following the recommendations of McCune and Grace (2002). We examined the effects of anthropogenic and physical environmental factors on fish community structure using permutational multivariate analysis of variance (PERMANOVA) based on distance matrices of the fish diversity at each site. To do so, we used adonis, in the package vegan (Oksanen et al. 2016), which partitions distance matrices among potential sources of variation. We Hit linear models to these distance matrices, and evaluated the pseudo-F ratios with a permutation test. The following model, stratiHied by year to control for potential inter-annual differences, fish ~ exposure + distance + population was selected by comparing the AIC score from models with all possible combinations of the following factors related to site characteristics: exposure (lagoonal or exposed), distance (distance in kilometers from the site to the village with jurisdiction), population (number of human inhabitants of the village with jurisdiction) and index (a measure of the site’s orientation with respect to the prevailing northeast trade winds), from a single factor to all 5 factors. We also examined the relationship between fish community structure and benthic cover characteristics using PERMANOVA. The number of permutations for all of these tests was set at 999.

BCO-DMO Data Processing Notes:

-Data were converted from wide format to long format
-nd was added to all blank cells
-site, date, time, and transect columns were added to incorporate the information contained in the header of the file