2.4 Community diversity and composition
We used the eight taxonomic subsets (arthropod orders) of ASVs to conduct analyses of community diversity and composition among sampling sites, considering haplotypes (raw ASVs) and all CLs (0.5%, 1.5%, 3%, 5% and 7.5% lineages, plus that corresponding to the GMYC species delimitation. We first estimated total accumulation richness curves for multiple levels of genetic similarity (AVS, 3% and 5% CL) using the R package betapart(Baselga & Orme, 2012). Then we tested for significant differences in alpha diversity among communities by performing ANOVAs with post-hoc.kruskal (method Bonferroni).
Total beta diversity (Sorensen index, βsor ), additive turnover (Simpson index, βsim ; species replacement, without the effect of variation in richness) and nestedness (Sorensen–Simpson index, βsne ; pure richness effect) components were then estimated based on community compositions at different hierarchical levels. We used the R packagevegan (Oksanen et al., 2019) and community composition matrices to perform non-parametric multidimensional scaling ordination (NMDS) based on Sorensen similarity. Plots were created with the ordispideroption to visualise the compositional ordination of the communities among sites. We then compared arthropod communities composition between sites with an analysis of similarity (ANOSIM) test for each taxonomic group. ANOSIM is a non-parametric analogue for analyzing variance and testing multivariate differences between groups, based on a resemblance matrix and rank dissimilarity (Clarke, 1993). Plots were made in R using package ggplot2(Wickham et al., 2020).