Date of Award

Spring 3-23-2016

Level of Access Assigned by Author

Open-Access Thesis

Degree Name

Master of Science (MS)

Department

Ecology and Environmental Sciences

Advisor

Christopher Cronan

Second Committee Member

Robert Lilieholm

Third Committee Member

Cynthia Loftin

Additional Committee Members

Hamish Greig

Abstract

Catchment urbanization has deleterious effects on freshwater resources and aquatic communities in small stream ecosystems. In the State of Maine, many streams have been negatively affected by urbanization and are in need of management and restoration. Impervious cover (IC), i.e., any surface that impedes water infiltration into the ground, can serve as a measure of watershed urbanization. Recent studies conducted in Maine have indicated that stream biotic community structure and function begin to decline at impervious cover levels of approximately 1 to 15%. This wide range presents a challenge to regulatory agencies and watershed managers charged with protecting stream quality to avoid costly restoration efforts. In this research, we employed three statistical analyses to identify spatially-explicit watershed characteristics associated with climate, geology, and land use/land cover that affect stream vulnerability to urbanization. First, a Kruskal-Wallis one-way analysis of variance was used to discriminate watershed characteristics associated with macroinvertebrate and algal sample data classified into high and low vulnerability categories. Next, a logistic regression analysis was applied to predict attainment of stream regulatory standards based on macroinvertebrate and algal sample data combined with watershed biophysical parameters. Finally, a Bayesian network was developed to predict stream vulnerability to urbanization using an expert-informed model structure. Results from the three approaches identified a number of watershed parameters that are associated with the vulnerability of streams to impairment from urbanization stress. The Kruskal-Wallis analysis indicated that watersheds with higher amounts of well-draining soils, deeper water tables, and fewer wetlands are less likely to become impaired at a given value of IC. The logistic regression models provided evidence that watersheds with an intact riparian buffer, a shallow aquifer, soils resistant to erosion, few wetlands, and shallower soils are more likely to attain their regulatory standards and are thus less vulnerable to urbanization. The Bayesian network shared a number of similarities with the two statistical analyses in terms of important watershed parameters. Overall, results of the three analyses indicated that stream vulnerability tends to increase with a higher percentage of agriculture and wetlands in the watershed and to decrease with a higher percentage of forested or natural buffers and percent resistant surfaces in the watershed. The ultimate goal of this research was to identify specific streams that are at risk of becoming impaired by future development. This goal was achieved by integrating the results of the three-step vulnerability analysis with earlier work that created spatially-explicit development suitability indices for two major watersheds in Maine. Areas likely to face future degradation were identified as watersheds in the top quartile of vulnerability that coincide with areas highly suitable for development are likely to face future degradation. We highlighted the locations of these “at-risk” streams and provided resource managers and policy makes with a tool that can be used to prioritize and guide the protection of vulnerable streams in the Maine landscape.

Download

Share