Date of Award

2005

Document Type

Master's Thesis

Degree Name

Master of Science in Environmental Engineering (MS)

College, School or Department Name

Department of Civil and Environmental Engineering

Advisor

Brian D Barkdoll

Abstract

Riparian ecology plays an important part in the filtration of sediments from upland agricultural lands. The focus of this work makes use of multispectral high spatial resolution remote sensing imagery (Quickbird by Digital Globe) and geographic information systems (GIS) to characterize significant riparian attributes in the USDA’s experimental watershed, Goodwin Creek, located in northern Mississippi. Significant riparian filter characteristics include the width of the strip, vegetation properties, soil properties, topography, and upland land use practices. The land use and vegetation classes are extracted from the remotely sensed image with a supervised maximum likelihood classification algorithm. Accuracy assessments resulted in an acceptable overall accuracy of 84 percent. In addition to sensing riparian vegetation characteristics, this work addresses the issue of concentrated flow bypassing a riparian filter. Results indicate that Quickbird multispectral remote sensing and GIS data are capable of determining riparian impact on filtering sediment. Quickbird imagery is a practical solution for land managers to monitor the effectiveness of riparian filtration in an agricultural watershed.

Share

COinS