I'm assessing the uniqueness of the floristic assemblage of a rainforest community, comparing it with other floristic assemblages of rf communities within a geographic region. I'm wanting to know about the pros and cons of using NDMS ordination for comparison compared with partitioning Beta Diversity in to Localised Contributions to Beta Diversity.
Ordination will tease out distributions and beta diversity will compare different assemblages, so yes, a good combination :)
They're different, but complementary, techniques. nMDS allows you to visualize relationships among samples and groups of samples in lower-dimensional space (usually 2-d). If you are doing partitioning, or any of the wealth of resemblance- or distance-based analyses (e.g. ANOSIM, PERMANOVA, etc.) then you are doing hypothesis testing on those inter-group relationships. The current 'mantra' is to test using the full resemblances (replicates) and interpret by looking at the relationships among means. There's something about that here: doi:10.1111/aec.13059
- Badges
- Science method