Invasive weeds in Glacier National Park: Species distribution modeling with MAXENT and CART

Adams, Ashley

Glacier National Park is threatened by the encroachment of numerous noxious non-native invasive plant species. Spotted Knapweed (Centaurea maculosa), Leafy Spurge (Euphorbia esula), Oxeye Daisy (Chrysanthemum leucan), Canada Thistle (Cirsium arvense), Bull Thistle (Cirsium vulgare), Orange Hawkweed (Hieracium aurantiacum), Meadow Hawkweed (Hieracium pretense), Yellow Toadflax (Linaria vulgaris), and St. Johnswort (Hypericum perforatum) are the among most virulent invasive species found within the park and are currently considered the greatest invasive threat. Prevention, early detection, and immediate action against invasive weed species in their initial establishment phases are paramount in reducing this threat. In order to prioritize monitoring and preventative management, this study built and compared Classification and Regression Tree (CART) and Maximum Entropy models to determine potential invasive species distribution for these nine highest risk invasive species and all of Glacier National Parks’ invasive species combined. It was found that the Maximum Entropy model using MAXENT software had the highest species distribution prediction accuracy, and that potential invasive species distribution covered between 0% and 10% of Glacier National Park, depending on the individual species and the model used. For all invasive species combined, the CART model predicted 13% of the area in Glacier National Park is potential habitat for all invasive species, while MAXENT predicted 6%, with an overlap agreement between the two models of 5%. The environmental variables most influential in invasive species distribution were proximity to roads and trails, and elevation, respectively. These finding can be used by Glacier National Park as a guideline for monitoring and prevention management actions.

Invasive weeds in Glacier National Park: Species distribution modeling with MAXENT and CART

Collections