The challenges and consequences of use of a large plot database to improve the USNVC classification of Pinus palustris savannas of the southeastern Coastal Plain

Background/Question/Methods

In 2008 the US Federal Geographic Data Committee adopted new standards for the US National Vegetation Classification (NVC) that mandate the use of a peer review process for evaluating proposed changes plus the use of plot data for documentation of the recognized vegetation types.  In cooperation with the ESA Vegetation Panel we developed a demonstration dataset and associated analysis for evaluating and documenting best practices for implementing the new standards.  For this project we focused on the fire-maintained longleaf pine (Pinus palustris) savannas of the southeastern Coastal Plain.  This focal system provides unique challenges in that the floristic diversity of the region has led to recognition of over 130 longleaf-dominated vegetation associations, despite the low topographic relief and consistent canopy dominance. We were particularly interested in the degree to which quantitative data would yield a classification consistent with the types previously described, and in developing best practices for validating and incorporating changes in the established classification.

Results/Conclusions

We identified ~1100 vegetation plots from the Carolinas, Georgia and Florida containing ~1600 species and representing the range of compositional variation of longleaf pine vegetation in those states. We followed the protocol described by Peet and Roberts (2013) to homogenize the data, develop clusters consistent in compositional variation relative to other types in the NVC, assess indicator species, and propose appropriate changes in the NVC.  We examined ~80 associations (vegetation types) and the total number recognized remained relatively constant.  Of those associations ~25% remained unchanged and were for the first time quantitatively described, ~50% had complex relationships with one or more previously recognized associations, and ~25% were entirely new associations. The complex relationships between new and established associations need to be documented by set theory relationships of the sort used in the taxon concept schema of TDWG as otherwise updating datasets containing references to NVC associations becomes intractable. Despite having a large dataset spanning a large geographic region, we faced challenges related to the requirement that compositional variation of associations be described across the geographic range of those associations. The new NVC standards represent a major step forward, but actual implementation will be a challenging and slow process.