Date of Award

January 2015

Degree Type

Open Access Thesis

Degree Name

Master of Science (MS)

Department

Biological Sciences

First Advisor

David R. Brown

Department Affiliation

Biological Sciences

Second Advisor

Stephen C. Richter

Department Affiliation

Biological Sciences

Third Advisor

Amy Braccia

Department Affiliation

Biological Sciences

Abstract

Within the last two centuries, Kentucky has undergone wetland losses exceeding 80 percent (approximately 500,000 hectares). As a response to these losses, the Kentucky Division of Water (KDOW) and Eastern Kentucky University (EKU) developed the Kentucky Wetland Rapid Assessment Method (KY-WRAM) to evaluate the condition of Kentucky's remaining wetlands. The goal of this study was to validate the KY-WRAM for forested riverine wetlands using a vegetation index of biotic integrity (VIBI), bird surveys, and landscape development index (LDI). Specific objectives of this study were to: 1) determine the correlation between bird species richness, VIBI, and LDI with the KY-WRAM in forested riverine wetlands; and 2) determine which combination of vegetation and landscape metrics best explain each of the KY-WRAM metric categories. At twenty five sites throughout the Green, Upper Cumberland, and Kentucky River Basins, a KY-WRAM, VIBI, LDI, and survey for bird species richness was conducted. A linear regression indicated that the KY-WRAM was significantly, positively correlated with the VIBI and bird species richness, while the KY-WRAM showed a negative, marginally significant correlation with the LDI. Model-averaging using model selection and parameter estimates indicated that the top models and predictor variables were (1) percent forested, (2) floristic quality assessment index score and percent adventive, and (3) percent adventive and Carex species richness for Metric 2 (Buffers and Surrounding Land Use); Metric 4 (Habitat Reference Comparison); and, Metric 6 (Vegetation, Interspersion, and Microtopography), respectively. Overall, the method's effectiveness was demonstrated by its ability to be predicted by biological and landscape indices at the method level and biological and landscape variables at the metric level.

Share

COinS