Date of Award

5-31-2016

Document Type

Restricted Access Thesis

Degree Name

Master of Science in Environmental Science and Policy

Department

Department of Environmental Science and Policy

Thesis Advisor

Amy Villamagna

Committee Member

Mark Green

Committee Member

Scott Bailey

Abstract

Habitat degradation by human-driven increases in stream siltation is a global issue. Many species, such as the Roanoke logpech (Percina rex) are listed as federally "endangered" due to critical habitat impairment caused by sediment. Re-establishing vegetative riparian buffers potentially reduces sediment loading and improves benthic species habitat overtime. However funding and time allocated to riparian restoration projects by conservation and resource management agencies is greatly limited. Given the global extent of water impairment status, critical habitat degradation, and growing development, a method of prioritizing riparian restoration sties is needed for efficient resource allocation. In our study, we created a decision support framework to help conservation and management agencies prioritize sites for riparian restoration efforts. Our framework integrates a statistical modeling approach to assess instream silt cover and embeddedness (inhibiting factors to benthic species) in sections of the Roanoke and Nottoway River basins in Virginia and North Carolina. In the first section of our study, we assessed landscape characteristics' contribution to explaining variance of instream silt. We also assessed spatial scale effects on statistical model performance and site ranking. The second section of our study focused on assessing differences in statistical model performance integrating annual soil loss estimates from two soil loss models, the Revised Universal Soil Loss Equation (RUSLE) and the Soil and Water Assessment Tool (SWAT). Our results show that landscape characteristics provide various contributions to explained variance of instream silt cover and embeddedness. Spatial scale analysis results show smaller spatial scales more accurately predicted instream silt and ranked sites by predicted silt conditions. Comparing soil loss and sediment load models, we found RUSLE performs similarly to SWAT in a statistical modeling context, and is much faster to implement, making it a more desirable soil loss model to integrate into our framework. Our decision-support framework provides a first step towards incorporating statistical models into restoration site prioritization for the potential improvement of instream habitat.

Share

COinS