Document Type

Honors Paper

Advisor

Chad Jones

Publication Date

2013

Abstract

Invasive species present a growing risk to native ecosystems. It would be impossible to eliminate them, but by learning where they are more likely to occur it may be possible to concentrate resources on removal and management in the most at risk areas. A number of site characteristics may influence invasive species presence and invasion trends. In order to determine what factors affect species presence and historical invasion trends, I utilized data from decennial surveys conducted in the Bolleswood Natural Area of the Connecticut College Arboretum in New London, Connecticut. In 1952, 4 transects were established across the area, composed of contiguous 3.05 by 3.05 meter plots. Every ten years, presence, height, DBH, and cover of all species in each plot were recorded. Other environmental variables in the plots were recorded as well. I used logistic regression models to determine the factors that explained the distribution of invasive species in 2012. I also determined if the factors that influenced invasion changed over time. Models were formulated to analyze how overall presence of invasives in 2012 and new occurrences in each sample year related to site history in regard to disturbance, soil characteristics, topography, light availability, dispersal, and species richness. For presence data in 2012, topography, light availability, dispersal, and species richness. For presence data in 2012, plots on historically agricultural land with moister and less well-drained soils and higher soil pH were more likely to host invasive species. For historical invasion trends, plots on historically agricultural land with greater total richness and proximately to the fence during initial invasion and later proximity to adult plants were the best indicators. However, age class appears to play a role beyond absolute forest age, as invasion was consistently found inthe youngest, historically agricultural, plots even after 60 years of forest development. These conclusions suggest that there are factors that can be utilized to predict invasion, and confirm the benefit of long term studies within a large area, such as the one at hand.

COinS
 

The views expressed in this paper are solely those of the author.