Abstract Detail

Systematics Section / ASPT

Smith, Stephen A. [2], Beaman, Reed S. [1], Donoghue, Michael J. [2].

Comparison of predictive distribution modeling in a Viburnum species complex from Mexico and Central America.

Viburnum (Adoxaceae), a group containing ca. 160 species, includes several distinctive species complexes in the mountains of Mexico and Central America. Our analysis focused on the V. discolor species complex, characterized by large leaves and stellate trichomes. Using georeferenced herbarium specimen data from each of the taxa in this complex, we compared three predictive modeling methods developed for species distributions: the Genetic Algorithm for Rule set Production (GARP) as implemented by Desktop GARP, BIOCLIM as implemented by DIVA-GIS, and WhyWhere, a web-based predictive modeling service that uses image processing algorithms to generate predictions. In an initial comparison, we generated distribution models using default environmental data. The GARP prediction tended to produce the most conservative geographic range, whereas the WhyWhere method produced the least conservative. An interesting result of the BIOCLIM method was over prediction of certain species ranges into the known distribution (niche) of a closely related species. This result can be used to infer niche conservatism and quantifies the ecological similarities of taxa within a clade. We are currently generating a standard set of environmental data layers to use in all three models, which will allow better comparisons among the methods.

1 - Yale University, Peabody Museum of Natural History, Botany Division, 170 Whitney Ave, P.O. Box 208118, New Haven, Connecticut, 06520-8118, USA
2 - Yale University, Department of Ecology and Evolutionary Biology, P.O. Box 208105, New Haven, Connecticut, 06520-8105, USA

Keywords:
Viburnum
Mexico
GIS
geographical distribution
Guatemala
distribution models.

Presentation Type: Poster
Session: 32-115
Location: Special Event Center (Cliff Lodge)
Date: Tuesday, August 3rd, 2004
Time: 12:30 PM
Abstract ID:632