Thesis Title

Remote Sensing Classification Methods For Monitoring Eastern Red Cedar

Date of Graduation

Spring 2006

Degree

Master of Science in Geospatial Sciences

Department

Geography, Geology and Planning

Committee Chair

L. Monika Moskal

Keywords

ASTER, cedar encroachment, decision tree, accuracy assessment

Subject Categories

Botany | Remote Sensing

Abstract

Eastern red cedar (Juniperus virginiana L.) populations are increasing throughout the Ozarks region. Fire and animal grazing have historically kept cedar populations under control, but today their occurrence has decreased in most areas. Cedar encroachment is taking place as these populations move into new landscapes. This encroachment has closed previously open canopies in savannas and glades, as well as converted grasslands into cedar forests. The result of which is a loss of biodiversity in native plant communities as well as destruction of species habitat. Given such environmental impacts it is critical for sustainable resource management that cedar encroachment be monitored. This study examines the use of ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) satellite imagery in monitoring cedar populations. Different remote sensing image classification methods were evaluated to determine the best approach for identifying cedar locations. An alternative accuracy assessment method was created to better evaluate classification results. The results show that using a decision tree classifier and ASTER imagery, cedar locations can be accurately identified with an overall accuracy of 92%.

Copyright

© Nathaniel J. Huggins

Citation-only

Dissertation/Thesis

Share

COinS