|
Abstract:
|
Invasive species have become a major problem in the US, but our understanding of invasion patterns and key drivers are still limited. Using a powerful tool in predictive biogeography, Boosted Regression Trees (BRTs), we created statistical models which can predict exotic species distribution for the Eastern United States. BRTs build on binary decision trees and combine them to create a linear combination of many trees. This leads to a more accurate model of invasion prediction and allows us to better identify key underlying variables that drive the observed patterns. Our goal was to create a model with many trees and low deviance that could accurately predict invasive plant species richness patterns for the Eastern United States. Ultimately, we were able to map these predictions onto a map using R. This map, in conjunction with our model, will help us better understand drivers of invasion by quantifying the relative contribution of each variable. Additionally, the results from our studies can then be used by policy makers and practitioners to manage invasions of species with more proactive measures and preventative actions. This work is supported by NSF grant DMS #1246818.
|
