Peter Palmiotto, PhD
Department of Environmental Studies
The conservation of rare plant species needs accurate predictions of suitable habitat to estimate the potential impact of climate, land use, and ecological change on individual populations and regional distributions, and in the search for new occurrences. I developed regional (6.8 x 104 km2) fine-scale (0.1 ha) predictive ecological niche models for the three birds orchid (Triphora trianthophora (Swartz) Rydb.) and small whorled pogonia (Isotria medeoloides Pursh) Raf.), two rare upland orchids that occur in relatively high densities in northern New England. Vegetation sampling at 20 T. trianthophora sites quantified the consistent dominance of American beech (Fagus grandifolia L.), a potential ectomycorrhizal host species for this orchid. Ecological Niche Factor Analysis (ENFA) identified a more marginal and specialized niche for T. trianthophora and retained totpgraphic, climate, and land cover gradients as niche factors for both species. ENFA models filtered by land cover and substrate masks and restricted to a smaller area identified more subtle ecological factors. A Generalized Linear Model (GLM) using pseudo-absences in ENFA-predicted unsuitable habitat confirmed a precipitation-based bioclimatic optimum within the study area for I. medeoloides. Spatial contagion terms in GLMs for both species addressed residual dispersal-related structure in their local (1-10 km2) and regional (= 104 km) distributions. Results show that presence-only data for these rare plants are adequate for useful predictions of suitable habitat, with several potential applications to species conservation and management.