Figure 3.
The influence of environmental factors on variation of abundance-weighted compositions evaluated at fine to broad taxonomic resolutions. The community analyses are conducted on eight distinct data sets: (a) case #1 America-Soils, (b) case #2 Arctic-Soils, (c) case #3 NYpark-Soils, (d) case #4 NYfarm-Soils, (e) case #5 TaraSur-Seawaters, (f) case #6 TaraChl-Seawaters, (g) case #7 WEC-Seawaters and (h) case #8 SECS-Seawaters. Taxonomy-based compositional variation is calculated following two types of hierarchical classification (sequence similarity: otu97 to otu70 and database annotation: species to phylum), whereas phylogeny-based compositional variation is calculated using UniFrac metric for comparison. The influence of environmental factors on composition (adjusted R2, the explained variance) is quantified by distance-based redundancy analysis (db-RDA). The red boxplots (connected by the black line along taxonomic ranks) represent the explained variance of the db-RDA results based on 100 subsamples derived from the original data sets, whereas the gray boxplots represent the results based on the randomized data sets.