Date of Graduation

8-2016

Document Type

Thesis

Degree Name

Master of Science in Geography (MS)

Degree Level

Graduate

Department

Geosciences

Advisor/Mentor

Tullis, Jason A.

Committee Member

Stahle, David W.

Second Committee Member

Cothren, Jackson D.

Keywords

Social sciences; Earth sciences; Forestry; GIS; Lidar; Machine learning; North carolina; Old growth trees

Abstract

The remnants of ancient baldcypress forests continue to grow across the Southeastern United States. These long lived trees are invaluable for biodiversity along riverine ecosystems, provide habitat to a myriad of animal species, and augment the proxy climate record for North America. While extensive logging of the areas along the Black River in North Carolina has mostly decimated ancient forests of many species including the baldcypress, conservation efforts from The Nature Conservancy and other partners are under way. In order to more efficiently find and study these enduring stands of baldcypress, some of which are estimated to be more than 1,000 years old, LiDAR remote sensing and geospatial analysis techniques can be employed. Promising results have been discovered correlating LiDAR-derived metrics and known stands of old growth baldcypress. A number of percentile height metrics and other composite metrics like canopy cover and density were extracted from LiDAR data collected across North Carolina. Along with the metrics, locations of known stands of old growth were used as training data for a supervised classification with the C5.0 decision tree algorithm. C5.0 was used to condense the patterns found across the training data into a set of rules that could then be applied to other areas within the study site or anywhere else across the LiDAR data. Both existing stands and new areas were selected by the machine learning rulesets indicating that the use of machine learning is valid to identify stands of ancient trees along the Black River. Overall C5.0 accuracies of approximately 98.5% (based on training data) and 88.6% (based on independent test data) were achieved. More than 8 km2 of predicted old growth forests, outside of available in situ reference areas, were also identified within the Black River site.

Download

Share

COinS