OOS 39-3
Species distribution model utility in native plant materials development

Wednesday, August 12, 2015: 8:40 AM
316, Baltimore Convention Center
Troy Wood, Colorado Plateau Research Station, U.S. Geological Survey, Flagstaff, AZ

The use of species disritubtion modeling derived from climate data is ubiquitous in academic research. In this talk, I will present application of this statistical approach to understanding species' ecology to informing native plant materials selection. Specifically, I ask if and when outputs of formal modelling of species' distribution is more useful than raw occurrence records in identifying population source materials across a species' range for sampling and evaluation. In addition, by aligning, modeling and common garden data, I ask whether or not climate predictors that fall out as most important in determining species range also covary most strongly with trait variation among populations.


The utility of species distribution model output can facilitate objective prioritzation of which populations within a species to sample, particularly when model ouput is linked to spatial data on historical and predicted disturbance. However, the use of raw occurrence records can yield qualitative similar results. Results from the evaluation of model output variables in predicting intraspecific trait variation are mixed and based on small sample size, thus any general conclusions await further inquiry. Overall, given that the tools and data needed to model suitable habitat of species are readily at hand, their use in plant materials development can be effective, particularly for gaining biological insight into lesser-studied species.