2.5. Statistical analysis
R software [18] was used to perform statistical analysis. In
default setting, analysis of similarity (ANOSIM) was done
using the BrayeCurtis distance measure and canonical correspondence
analysis (CCA). Nutrients and metals were tested
for significant correlation with sampling site distribution in the
CCA ordination plot using the envfit command, with 999
permutations. Similarity percentage (SIMPER) analysis [19]
was done according to sample groups as defined after cluster
analysis through PAST software using the BrayeCurtis distance
measure [20].