Modeling tree species diversity in NC Piedmont forests based on forest structure

Christopher R Hakkenberg, University of North Carolina at Chapel Hill, hakkenberg@unc.edu (Presenter)
Conghe Song, Univ. of North Carolina Chapel Hill, csong@email.unc.edu

Niche models can explain a substantial proportion of the variance observed in the spatial distribution of taxonomic diversity, however they are constrained by limitations in coverage of ground data that may be cost prohibitive to measure at fine resolutions over large spatial extents. This study investigates the utility of employing a remotely sensible proxy variable, forest structure, as a predictor of tree species diversity in the North Carolina Piedmont. Results confirm that in the NC Piedmont forest structure is an effective predictor of tree species diversity, even without accounting for more proximate factors driving diversity like community type or topo-edaphic conditions. Among all structural measures, total basal area and indicators of structural complexity, like maximum height, variance in DBH and the form of the size class distribution, are the best predictors of species diversity. Two remotely-derivable structural predictors (maximum height and mean basal area) together explain 33% of variance in species richness, while a nonparametric support vector regression (SVR) model based on all 16 structural predictors accounted for 55% of variance. The data likewise confirms that indicators of structural complexity increase with stand age, as does tree species diversity. In addition to the theoretical implications, these results highlight the potential for using forest structure as a remotely-derivable proxy variable, used in conjunction with other remote inputs, to parameterize forest diversity distribution models.

Presentation Type:  Poster

Session:  General Contributions   (Tue 4:35 PM)

Associated Project(s): 

• Related Activity: Related Activity or Previously Funded CC&E Activity not listed ...details

Poster Location ID: 158