A habitat characterization and suitability model for the endangered wetland plant Lindera melissifolia in the Southeastern Coastal Plain

Background/Question/Methods

Lindera melissifolia (pondberry) is a federally endangered wetland shrub extant in the Southeastern Coastal Plain and the Lower Mississippi Alluvial Valley.  Pondberry inhabits a variety of wetland types throughout its range, including isolated depression wetlands in the Coastal Plain.  The US Fish and Wildlife Recovery Plan for pondberry cites extensive habitat loss as the primary cause of the species’ endangerment and the largest challenge to its recovery.  To better understand the habitat parameters of pondberry, we surveyed 45 populations in North Carolina, South Carolina, and Georgia and collected data on co-occurring species, forest structure, topography, hydrology, soils, and disturbance.  These data were used to characterize the habitat type of each occurrence, to determine how dominant environmental factors influence pondberry abundance, and to test the hypothesis that landscape-scale data derived from a GIS accurately reflects plot-level conditions.  These findings were used to develop a landscape scale habitat suitability model. 

Results/Conclusions

Cluster analysis was used to develop a habitat characterization of Lindera melissifolia occurrences and to assign each plot a community type based upon the International Vegetation Classification.  Environmental characteristics were analyzed using a principal components analysis which revealed that out of 26 observed features canopy cover, species richness, moss and lichen cover, pondberry abundance, depth to soil gleys, and depth to soil O horizon accounted for the most variance between plots (61.8%).  Our hypothesis that landscape level data accurately reflects on-site conditions was confirmed by a preliminary comparative analysis of 6 soils characteristics.  This analysis also identified specific soils parameters to include in the habitat model.  A preliminary habitat suitability model was developed using our plot-level environmental and community analyses in conjunction with landscape characteristics including soils, land cover, patch size, and disturbance indices.  Our research demonstrates that community analysis can be combined with pre-existing landscape scale datasets to leverage in situ findings into predictive habitat suitability models.  Such models can identify the geographic locations in which pondberry restoration or reintroduction efforts are likely to produce self-sustaining populations.  The modeling techniques demonstrated here can be applied to any rare plant taxon.