Discriminating woody species assemblages from National Forest Inventory data based on phylogeny in Georgia

Area of Science:

• Ecology
• Forestry
• Biodiversity Science

Background:

• Forest vegetation classification and species assemblage characterization are crucial for mapping and monitoring forest diversity.
• Beta-diversity indices quantify compositional dissimilarity between samples, aiding forest community discrimination.

Purpose of the Study:

• To evaluate unsupervised classification for National Forest Inventory data from Georgia.
• To compare cluster hierarchies using species-neutral dissimilarity (Bray-Curtis) versus phylogenetic variation-based dissimilarity (Discriminating Avalanche).

Main Methods:

• Calculated mean basal area per hectare for 87 woody species across 1059 plots.
• Quantified interspecies distances and performed unsupervised cluster analysis.
• Compared Bray-Curtis dissimilarity with Discriminating Avalanche dissimilarity incorporating phylogenetic variation.

Main Results:

• Phylogenetic variation incorporation yielded more nuanced species assemblage discrimination and increased cluster coherence.
• The number of clusters increased (23 vs. 20), mean distance within clusters decreased (0.773 vs. 0.343), and within-sum-of-squares improved (344.13 vs. 112.92).
• Clusters based on genetic variation aligned better with biogeographical units like elevation and habitats.

Conclusions:

• The effectiveness of unsupervised classification for large-scale forest inventory data hinges on the dissimilarity quantification method.
• Incorporating phylogenetic variation leads to more precise forest classification, aligning with biogeographical patterns.
• This approach enhances forest mapping, monitoring, and diversity management by leveraging the genetic signal within species assemblages.